Jul 27, 2021
- 4:00 - Brian's lofty title is Head of Product at Excel, The importance and magic of Excel, and people's a-ha moments with Excel
- 20:25 - The difficulty of not seeing your projects' impact on the world and how the heck does Bluetooth fit into the story?!, Rob and Brian reminisce with some funny conference stories
- 32:00 - The XML file format and some very neat XML tricks that everyone should know about
- 51:25 - The birth of the Excel Web App and Rob can't believe some of the things that Brian's team has done with Excel
- 1:05:00 - How to onboard the Excel, VLOOKUP, and Pivot crowd into data modeling and Power BI, and the future of Excel most certainly includes the Lambda function (maybe!)
Episode Transcript:
Rob Collie (00:00:00):
Hello, friends. Today's guest, Brian Jones, head of product for
this thing you might've heard of called Microsoft Excel. Brian and
I go back a long way. We were both youngsters at Microsoft at the
same time, and we both worked on some early features of Office
apps, and we're friends. Really, really have sincerely warm
feelings about this guy, as you often do with people that you
essentially grew up with. And that's what we did. When Brian and I
first worked together, he was working on Word and I was working on
Excel. But even though Brian was on Word at the time, he was
already working on what we would today call citizen developer type
of functionality in the Word application. So even though we were
essentially on different sides of the aisle within the Office
organization, we were already finding ourselves able to connect
over this affinity for the citizen developer.
Rob Collie (00:00:55):
Now we have some laughs during this conversation about how in
hindsight, the things he and I were working on at the time didn't
turn out to be as significant as we thought they were in the
moment. But those experiences were very valuable in shaping both of
us for the initiatives that came later.
Rob Collie (00:01:11):
Like almost everyone at Microsoft, Brian has moved around a bit.
He's worked on file formats for the entire Office suite, which
ended up enabling Power Pivot version one to actually function the
way that it should. He's worked on Office-wide extensibility and
programmability, back to that citizen developer thing again. And in
that light, it's only natural that Excel's gravity reeled him in.
And in that light, it's only natural that someone like that,
someone like Brian, found his way to Excel, and it really is a
match made in heaven. And if you permit me the Excel joke, that
turned out to be a great match.
Rob Collie (00:01:50):
We took the obligatory and entertaining, I hope, walk down memory
lane. We spent a lot more time than I expected talking about file
format. And the reason why is that file formats are actually a
fascinating topic when you really get into it. Lot of history
there, a lot of very interesting history and challenges we walked
through. And of course, we do get around to talking about Excel,
its current state, where it's headed, and also the amazing
revelation for me that monthly releases actually mean a longer
attention span for a product and how we ended up getting
functionality now as a result of the monthly release cycle that
would have never fit into the old multi-year release cycle. We were
super grateful to have him on the show. And as usual, we learned
things. I learned things. I have a different view of the world
after having this conversation than I did before it, which is a
huge gift. And I hope that you get the same sort of thing out of
it. So let's get into it.
Announcer (00:02:56):
Ladies and gentlemen, may I have your attention please?
Announcer (00:03:03):
This is the Raw Data by P3 Adaptive podcast, with your host Rob
Collie and your cohost Thomas Larock. Find out what the experts at
P3 Adaptive can do for your business. Just go to p3adaptive.com.
Raw Data by P3 Adaptive is data with the human element.
Rob Collie (00:03:26):
Welcome to the show. Brian Jones, how are you, sir?
Brian Jones (00:03:30):
I am fantastic. Thank you for having me, Rob. I'm excited.
Rob Collie (00:03:33):
So let's start here today. Well, you and I go way back, but today,
what's your job title and what are your responsibilities?
Brian Jones (00:03:42):
So today, my job is I'm the head of product for the Excel team. So
I lead the team of product managers that are tasked with or given
the honor of deciding the future of Excel, where we go with Excel,
what are the set of things that we go and build
Rob Collie (00:03:59):
Head of product. That's a title that we didn't have back when I was
still at Microsoft. We did at one point have something called a
product unit manager. Is it similar to that? How does that
relate?
Brian Jones (00:04:11):
That's a good question. So we're continuing to evolve the way that
we use titles internally. So internally, we have titles that still
for most folks externally don't make any sense, like program
manager, group program manager, program manager manager, director
of program manager. And so for externally, whenever I'm on LinkedIn
or if I do PR interviews, things like that, I use the term head of
product. Internally, we don't have the term head of product.
Rob Collie (00:04:37):
Okay. All right. So that's a translation for us.
Brian Jones (00:04:40):
Yes, exactly. Trying to translate the Microsoft internal org chart
to something that makes more sense to folks.
Rob Collie (00:04:49):
Yeah. So things like, if we use the word orthogonal, what we're
really saying is that's not relevant.
Brian Jones (00:04:53):
Exactly.
Rob Collie (00:04:54):
That kind of decoder ring.
Brian Jones (00:04:57):
I didn't realize orthogonal [inaudible 00:04:59] until you said it
and I'm like, " Oh yeah, no. Of course, that is completely a
ridiculous term to use."
Rob Collie (00:05:03):
Or I don't know if they still do this, but an old joke that Dave
[Gayner 00:05:07] and I used to have, it was all his joke at the
time. It was big bet. Do we still talk about big bet? We're going
to place a big bet.
Brian Jones (00:05:14):
Yep. Big bet or big rocks. Big rocks. You know the-
Rob Collie (00:05:17):
Big rocks. Whoa.
Brian Jones (00:05:18):
Yeah. It's kind of an analogy. You've got a jar and you want to
fill it with the big rocks first, and then you let the sand fill in
the rest of the space. So what are the big rocks?
Rob Collie (00:05:26):
Okay. Yeah. But big bet was one that we used to always make fun
of.
Brian Jones (00:05:31):
Especially when there'd be, "Here are the big bets," and there's 20
of them.
Rob Collie (00:05:34):
Yeah. The joke I think we used to make was we would call something
a big bet when we really didn't have any good reason for doing what
we were doing. Anyway, all right. So you're head of product for
Excel. That is a pretty heady job. That's pretty awesome.
Brian Jones (00:05:52):
It's a pretty fun job. Absolutely.
Rob Collie (00:05:54):
I mean, you're not lacking for eyeballs in that business, are you?
We're all friends here. We're all on the same side of this story. I
mean, it is the lingua franca of business, Excel. It is the
business programming tool. People don't necessarily think of it as
programming, but formulas are a programming language. To be head of
product for the platform, you could call it an application, but
really it's probably more accurate to call it a platform that is, I
think, is the single most critical platform to business in the
world. That's pretty amazing.
Brian Jones (00:06:30):
Absolutely. And that's usually the way that we talk about it
internally. It depends on who your audience is externally when
you're talking about it. But yeah, Excel is a programming language.
I remember even before, back when I was on the Word team, but I
would go and meet with PJ, who ran program manager for Office all
up. And he'd always referred to Excel more as an IDE. And that
didn't totally resonate with me at the time because to me, Excel
was just a list app, an app for just tracking things. I didn't
totally understand what he meant by that, but I'd nod cause he was
super important and smart. And it wasn't really until I started
working on the team that I was like, "Oh, I totally understand all
these things that PJ used to reference."
Rob Collie (00:07:06):
This one of the things I had been dying to ask you is when you and
I first met, I was working on the Excel team, but still had...
Gosh, this was year 2000 maybe, maybe 2001. And even though I was
nominally part of the Excel team at that point, I still didn't
really know Excel, and you were working on Word. So the thing we
both had in common at that point is that we didn't know Excel. So I
wanted to get your perspective. I know that you've done some things
other than Word, but we were already sort of teasing this. So let's
just get into it. What's it like to come from "outside" Excel and
how's that transition? How do you view Excel differently today
versus what you did before? We already started talking about that.
The list keeper. That's very common way for people to view it.
Brian Jones (00:07:53):
When I first started, yeah, I was on Word, although I was working
on more kind of end user developer type of pieces of Word. That's
how you and I first interacted because we were talking about XML.
The first feature I owned was a feature called easy data binding to
Excel. And the whole idea was when you could easily bring content
from Excel into Word, but then create a link back so that the
content in Word would stay live. And a lot of this stuff that I did
while I was on Word was all about trying to make Word a little bit
more of a structured tool so that people could actually program
against it because Word is completely unstructured. It's just
free-flowing text. So trying to write a solution against that is
almost impossible because you can't predict anything. So we did a
lot of work to add structure, whereas Excel out of the gate has all
that structure. So it's just much easier to go and program.
Brian Jones (00:08:39):
If I had gone straight from Word to Excel, it would have been a
little bit more of a shock, but I actually had about eight years in
between where I was running our extensibility team. So a lot of the
work we would do was revving the add-in model and extensibility for
Excel. So I got some exposure there. When we did all of the file
format stuff and the whole file format campaign, That was a couple
of years where I was working really closely with a bunch of folks
in Excel, like Dan [Badigan 00:09:06] and folks like that. So I had
a bit of exposure, but I'll tell you when I first joined, I had a
similar job, but it was for the Access team and we were building up
some new tech.
Brian Jones (00:09:17):
Some of it still is there today. Office Forms came out of some of
the investments that we were doing in Access. But when I showed up
into Excel, I was very much in that mode of, "Why don't the Excel
folks, get it? Everything should be a table with column headings."
And like, "That's the model. And why do they stick with this grid?
Clearly word of it is eventually going to go away from the printed
page as the key medium. Excel's got to go away from the grid. And
they've got to understand that this should just be all tables that
can be related." And thankfully, I was responsible when I joined
and didn't try and act like I knew everything. So I took some time
to go and learn.
Brian Jones (00:09:52):
And it didn't take me long. We have some crazy financial modeling
experts on the team and stuff like that, where I'd say it was maybe
six months in that it clicked for me where I understood those two
key pieces. The grid and formulas are really the soul and the IP of
Excel. The fact that you can lay out information really easily on a
grid, you have formulas that are your logic, and you can do this
step-by-step set of processes where each cell is almost like
another little debug point for you. [Cal captain sub 00:10:20]
second, and it's the easiest way to go and learn logic and how to
build logic.
Brian Jones (00:10:25):
I didn't get any of that at that time, but you pick it up pretty
quickly when you start to look at all the solutions that people are
building. And now, obviously, I've been on the team now for five
years, so I'm super sold around it. But I'd say it took me a little
while and I'm still learning. It takes a while to learn the whole
thing.
Rob Collie (00:10:41):
Yeah. It's funny. Like you said, Word's completely unstructured.
You're looking in from the outside and you're like, "Well, Excel is
completely structured." Then you get close to it. You're like, "Oh
no. And it's not, really."
Brian Jones (00:10:52):
No. Not at all.
Rob Collie (00:10:53):
I mean, it's got the cells. Rows and columns. You can't avoid
those. But within that landscape, is it kind of deliberately wild
west? You can do whatever you need to. You're right. Okay. So
tables, yes. Tables are still very important. But you've got these
parameters and assumptions and inputs. And what do you do with
those? I mean, they're not make a table for those.
Brian Jones (00:11:19):
Yep. Absolutely. I think that the thing that I started to get
really quickly was the beauty of that. Like you said, it's
unstructured. You have nice reference points. So if you're trying
to build logic, formulas, you can reference things. But there's no
rule about whether or not things go horizontally, vertically,
diagonally, whatever. You can take whatever's in your mind that
you're trying to make a decision around and use that flexible grid
to lay it out. It's like a mind map. If you think about the beauty,
the flexibility of a mind map, that's what the grid is. You can go
and lay out all the information however it makes the most sense to
you.
Brian Jones (00:11:53):
Really, that's what makes Excel still so relevant today. If you
think about the way business is evolving, people are getting more
and more data, change is just more constant, business processes are
changing all the time. So there are certain processes where people
can say, "This thing is always going to work the same way." And so
you can go and get a vertical railed solution. That's why we use
the term rail. That's kind of like if I always know I'm going to
take this cargo from LA to San Francisco, I can go and build some
rails, and I got a train, it'll always go there and do the same
thing. But if business is constantly changing, those rails are
quickly going to break and you're going to have to go off the
rails. Excel is more like a car than a train. You can go anywhere
with it. And so as the business processes change, the people who
are using Excel are the same people who are the ones changing those
business processes. Those are the business folks. And so they can
go and evolve and adapt it and they don't have to go and find
another ISV to go and build them another solution based on that new
process that's probably going to change again in six months.
Thomas Larock (00:12:52):
So Brian's been in charge for five years of Excel, and he's sitting
there telling us how there's still more to learn. And two weeks
ago, we all got renewed as MVPs. And so I was on the MVP website,
and I'm going through all the DLs I can join because that's all a
manual process these days. I'm like, "Oh, there's the Excel MVP DL.
I don't know why I haven't joined this yet." So I click. I'm
immediately flooded with 100 emails a day. 100 emails a day. Now, I
don't believe I am a novice when it comes to Excel. I don't. I know
I'm not on you all's level at all when it comes to it. You build
and work and live the product. But I know my way around enough that
I can explain things to others when they say, "I'm trying to do
this thing." "Oh, I think it's possible."
Thomas Larock (00:13:40):
But I read these passionate MVPs that you have and the stuff that
they highlight, and it's not complex stuff. It's like, "Hey, this
title bar seems to be wider in this." And I'm like, I might not
even notice this stuff. And I see these features that aren't a
complex feature, but I'm like, "I didn't even know that was there.
I didn't even know you could do that. Oh, you can do that too."
There's so much. And like you said, it's a programming language.
It's an IDE. It's all these things. As [Sinopski 00:14:10] said,
"It's the killer app for Windows." To have the head of product say
that, there's just so much. He really means it. There is a lot to
it. And it is something that is malleable and usable by hundreds of
millions of people a day.
Brian Jones (00:14:25):
Yeah.
Rob Collie (00:14:26):
My old joke is, if you want to know how good someone is at Excel,
just ask them, "How good are you at Excel?" And then take their
answer and invert it.
Brian Jones (00:14:37):
That's absolutely true.
Rob Collie (00:14:38):
If someone says, "Yeah, I'm really good at it," You know they don't
have any clue because they haven't glimpsed the depth of that
particular mine shaft. And once someone has been to the show, they
know better than to oversell their knowledge because they know they
can't know everything.
Rob Collie (00:14:54):
You say you're good at Excel. And then the very next question is
one that you're not going to be able to answer. So you got to be
careful. [inaudible 00:15:00] person views Excel as Word with a
grid. And that's not obviously what it is, but that's the
oversimplification for... I don't know... maybe 80% of
humanity.
Brian Jones (00:15:10):
Yeah. And the thing is, there's a lot more that we're doing in the
app now to try and make it, one, more approachable, because there's
a set of folks that just find it really intimidating, for sure. You
open it up and it's this huge, dense grid. Like, "Hey, where do I
start? What should I go and do? I've never even heard of this thing
before." In the past, a lot of stuff that we would do, we never
really thought about those first steps of using the app because we
were always like, "Well, everybody knows our app. We're going to go
and do the things for everybody that knows our app." And I think
we're doing a better job now trying to think, "Well, there's a
bunch of people who don't know about our app. Let's go and figure
out what the experience should be like for them."
Brian Jones (00:15:43):
But we've done a lot with AI where we're trying to get a little bit
better about... We look at your data. Recommend things to you. So
we'll say, "If you've got a table of data, hey, here's a pivot
table." You may not have even heard of the pivot table before. So
really more like, "Hey, here's a summary of your data." You want to
go and insert that.
Brian Jones (00:16:00):
In fact, those tests are always fun because then we get to work
with people who've really haven't ever used a pivot table. So it's
always fun to hear the words that they use to describe what a pivot
table is. It's like, "Oh wow, you grouped my data for me." Or stuff
like that like, "Wow. That's a nice name for it too." So we're
trying to do more of that to expose people to really those
higher-end things. But those things where for those of us that use
it, once you discover that stuff, you're even more hooked on the
product. You're like, man, that first experience of somebody built
a pivot table for you and you realize, "Oh my God, I didn't know I
could do this with my data. Look how much easier it is for me to
see what's going on," and trying to get more people to experience
that kind of magical moment.
Thomas Larock (00:16:39):
Now imagine being me and only knowing pivot through T-SQL and that
magical day when you meet Rob and he's like, "You just pivot table
[inaudible 00:16:49]." And you're like, "How many hours have I
wasted? Why didn't someone tell me?"
Brian Jones (00:16:56):
Yeah. We get that a lot when we'll go and show stuff. Oftentimes,
the reaction is more frustration. "I can't believe I didn't know
about this for the past five years."
Rob Collie (00:17:05):
We get that all the time now with Power Pivot and Power Query and
Power BI in general. The target audience for that stuff hasn't been
really effectively addressed by Microsoft marketing. But even back,
just regular pivot tables, such a powerful tool, and so poorly
named. You weren't around on the Excel team, Brian, when I waged a
six-month campaign to try to rename pivot table to summary
table.
Brian Jones (00:17:31):
Oh really?
Rob Collie (00:17:31):
Yeah.
Brian Jones (00:17:31):
How long ago was that?
Rob Collie (00:17:33):
Oh, well, it was a long time ago. I mean [crosstalk 00:17:35]-
Brian Jones (00:17:36):
Pivot tables had already been out for quite a while.
Rob Collie (00:17:37):
Oh God. Yeah. I mean, they were long established. They were in the
product. I didn't even know what they did. Believe it or not, I
worked on the Excel team for probably about a year before I
actually figured what pivot tables could do. People would just
throw it around all the time on the team like, "Well, once you have
the data, then you can chart it. You can pivot it," blah, blah,
blah, blah, blah. And so I would fit in-
Brian Jones (00:17:58):
You would nod?
Rob Collie (00:17:59):
I would fit in... I would also author sentences like that, that had
the word pivot in it. It was a pretty safe thing to do. There was
no downside to it. But believe it or not, the time that I
discovered what pivot tables are for... you'll love this... I was
trying to figure out how to skill balance the four different
fantasy football leagues that I had organized within the Excel and
Access team. I wanted to spread it out. Levels of experience. I've
got this table of data with the person's name and their level of
experience and my tentative league assignment. And just this light
bulb went on. I'm like, "Oh my God, I bet this is what pivot tables
are for." Total expertise by league. Like, "Oh, look at that. It's
totally it." That was a big change for me. That was during the
release, Brian, where you and I were working together.
Brian Jones (00:18:54):
I think I played on one of those fantasy football leagues.
Rob Collie (00:18:56):
You might have.
Brian Jones (00:18:57):
I was one of the people with zero experience. I remember going into
the draft not knowing... I knew football, but I didn't know
anything about fantasy football.
Rob Collie (00:19:03):
That's right. We did loop you in. So let's do that way back machine
for a moment. That release when you and I met was the first release
on Excel. I was the lead at that point. It was my first time being
a lead. It was the first time I was in charge of a feature set, and
it was really my baby, this XML thing we were doing. And the reason
for that was because no one was paying any attention. That was this
weird release. For a whole release, Office went and tried to do
cloud services without having any idea what that really was going
to mean. And so we stripped all of the applications down to
skeleton crews. And this is really the only reason why on the Excel
side, some youngster like me was allowed to be a lead and come up
with a feature, because no one cared. No one was paying any
attention. There was no one minding the store.
Rob Collie (00:19:48):
I remember being so wild-eyed enthusiastic about how much this was
going to change the world, this XML import export future. And I
mean, you might as well just take it out. I can't imagine it's
being used hardly at all today. I bet Power View is used more often
than the XML import export feature. You all have done a pretty good
job of hiding it. So kudos. But it was a good thing to cut my teeth
on. I learned a lot of valuable lessons on that release.
Rob Collie (00:20:24):
How do you feel about the XML structure document work that you were
doing in Word at the time? Do you kind of have the same feeling
looking back at it that I do?
Brian Jones (00:20:33):
It was a similar thing. In fact, we did rip it out a couple of
years later. I think that when you and I would talk about it, we
would talk about these scenarios that were super righteous and
great. And then we just start geeking out on tech. And then we
would get way too excited about the tech and we kind of forget
about those initial scenarios. We wouldn't stop and think, "Wait a
minute. These users we're talking about, are they actually going to
go and create XML files?" Because you need one of those to start
with before any of this stuff makes sense. And no, of course,
they're not. But for me, a lot of it started from that. Like I
said, one of my first features was that easy data binding to Excel
feature. And we thought, "Hey, maybe XML would be a good tech for
us to use as a way of having Word and Excel talk to each other,"
because clearly they have different views on what formatting is and
how to present information, but the underlying semantic
information, that could be shared.
Brian Jones (00:21:20):
And so I could have a set of products show up in Excel as a table.
And when they come into Word, they look more like a catalog of
products. That totally makes sense. And we just did a lot of
assumptions that people would make, do all the glue that was really
necessary. And of course, they didn't. So I had the exact same
experience. The other big thing that was different back then for us
was we would plan something, meet with customers for six months,
but then it'd take three years to go and build it. We had no way of
validating that stuff with customers because we couldn't get them
any of the builds. And then even after we shipped it, they weren't
actually going to deploy it for another three-plus years. And so
the reality is from when you had the idea to where you actually can
see that it's actually not working and people aren't using it is
probably about six years. So you've probably moved on to something
else by then.
Brian Jones (00:22:04):
The only way you really as a PM got validation that your feature
was great was whether or not leadership and maybe press got excited
about your thing, but you didn't get a whole lot of signal from
actual customers whether or not the thing was working, which is
obviously completely different now, thank goodness.
Rob Collie (00:22:18):
Yeah. That Is true. It took some of the fun out of being done too,
now looking back at it, like the day of the ship party, when we
were done with the three-year release. "Okay, fine." We'd dunk each
other in fountains and there'd be hijinks and stuff. But the world
did not experience us being done. That was purely just us feeling
done. And then it was like you take a week off maybe, and then the
next week, you're right back to the grind at the very beginning.
You never got the payoff. Even if you built something really good,
by the time the world discovered it and it was actually really
helping people at any significant scale, you're no longer even
working on that product.
Brian Jones (00:22:57):
Yeah. You're doing something completely different.
Rob Collie (00:22:59):
You might be in a different division, both finding out the things
in real time that
Rob Collie (00:23:03):
[inaudible 00:23:00] Both finding out the things, sort of in real
time, that aren't working. That's the obvious advantage, right? But
there's also this other emotional thing. Like you never got the
satisfaction when you actually did succeed.
Brian Jones (00:23:11):
Right. You didn't see it actually get picked up, adopted. Millions
and millions of people using it, which is what the team gets now.
We no longer pick a project and say, "Okay, how many people and how
long is this going to take?" You really just try and figure out
what's critical mass for that project. And then you just let them
run. And you'd be really clear around what are the goals and
outcomes they're trying to drive. And they just keep going until
they actually achieve that. Or we realized that we were wrong,
right? And we say, "Hey, we thought people are going to be excited
about this. It's not even an implementation thing. We were just
wrong. We misread what people really were trying to do. Let's stop.
Let's kind of figure out a way of moving off of that and go and
figure out what the next thing is we should go and do."
Rob Collie (00:23:50):
That era that we're talking about right now. The 2003 release of
Office. I was still very much a computer science graduate and
amateur human. That's exactly backwards, it turns out, if you're
trying to design a tool that's going to be used by humanity.
Brian Jones (00:24:08):
Well, it's what leads you to get really Excited about XML?
Rob Collie (00:24:12):
That's right. Yeah. That's right. Tech used to have such a power in
my life. I'm exactly the opposite now. Every time I hear about some
new tech, I'm like, "Yeah, prove it." I am not going to believe in
this new radical thing until it actually changes the world around
me. I'm not going to be trying to catch that wave. But XML did that
to me. It was almost a threat. If we don't take this seriously,
we're going to get outflanked. It got really egregious.
Rob Collie (00:24:42):
I had a coworker one time in that same release in the middle of one
of my presentations asked me. This guy wasn't particularly, in the
final analysis, looking back, not one of the stronger members of
the team, but he had a lot of sibling rivalry essentially in his
DNA. And he'd asked me in front of his crowded room, "Well, what
are you going to do about Bluetooth?" And, we didn't know what
Bluetooth was yet, right? It was like, unless I had an answer for
what we were going to do about Bluetooth and Excel, right? Then I
was not up on things. You know, the thing that we use to connect
our headphones. At the time, Bluetooth was one of those things that
might just disrupt everything.
Brian Jones (00:25:29):
It was funny. It was at that same time, I was asked to give a
presentation to the Word team about Bluetooth. We were all assigned
things to go and research as part of planning and that was one of
the ones I was asked. And I gave a presentation that was just very
factual. Here's what it is. And I was given really bad feedback
that like, "Hey, I wasn't actually talking about it strategically
and how it was going to affect Word. I was just being very
factual." And I was like, "I don't understand. I don't understand
what success looks like in this task." Right.
Rob Collie (00:25:59):
I remember going, a couple of years later, going into an offsite,
those offsite big, I don't know if you all still do those things,
big offsite, blue sky brainstorming sessions. There was this really
senior development lead that was there with me. And he and I were
kind of buddies. At one point, halfway through the day, he just
leans over to me and says, "Hey, I'm going to the restroom and I'm
not coming back." And I looked at him in horror, almost like "Thou
dost dishonor the offsite!?" And he's like, "Yeah, you know, I've
never really believed that much in this particular phase of the
product cycle. It's never really meant anything to me. It's all
just BS." It was just devastating. I just knew it was right. He
was...
Brian Jones (00:26:46):
But you didn't want to, you didn't want to believe that.
Rob Collie (00:26:52):
I mean, I felt so special. I was invited to the offsite, the big
wigs and everything.
Brian Jones (00:26:57):
They have nice catering too,
Rob Collie (00:26:59):
Yeah and he was totally right to leave.
Brian Jones (00:27:04):
I always remember getting super nervous to present stuff for those.
Once it was actually, it was one of our XML ones where I was trying
to convince, it was my attempt to get us to create an XML file
format, which actually ended up, obviously, happening. But I got an
engineer to go to work with and we had Word through an add-in,
start to write to XML. And it was just a basic XML format. And then
I built all of these... it was like asp.net tools that would go and
then create an HTML version of the Word doc that was editable. And
it also even created, I think it was called WHAP, I don't remember,
like a tech for phones. It was back when you didn't have the rich
feature phones, but these basic ones.
Brian Jones (00:27:41):
And so I created this thing that was almost like a SharePoint site.
So you could take all your Word docs, go through this add-in, and
then you could actually get an HTML view of them to edit it and a
phone view of them to go and edit it.
Brian Jones (00:27:51):
I think it was probably 2002 or 2001, but I was so excited to go
and show that at the offsite because I was like, "Okay, this is
where I make it, man. Everybody's going to be so excited about me."
But I don't know. I think everybody was excited about Bluetooth at
that point or something. Yeah.
Rob Collie (00:28:05):
Oh yeah Bluetooth, WHAP was so 15 minutes ago. So there's a few,
irresistibly funny or interesting things I want to zero in on from
that era before we come back to present, and we're definitely going
to come back to present, for sure.
Rob Collie (00:28:21):
First of all, we went to a conference like some W3C sponsor. I
don't think it was necessarily W3C affiliated, but it was the XML
conference.
Brian Jones (00:28:31):
The one in Baltimore?
Rob Collie (00:28:32):
Yes.
Rob Collie (00:28:33):
Okay. Now two very, very, very memorable things happened at that
conference. I bet you already know one of them. But the other one
was, and we're just going to make this all this anonymous person's
fault. Okay. We're not going to abdicate any responsibility. And
we're just going to talk about our one coworker from Eastern Europe
who brought his wife and they had vodka in their hotel fridge, or
freezer, or something like that. And every day I would wake up and
say, "I am not going to get suckered into that again."
Rob Collie (00:29:12):
And then the next day I would wake up and say the same thing. That
was a tough trip.
Brian Jones (00:29:16):
I definitely remember that.
Rob Collie (00:29:18):
Even on my young, relatively young, body at the time that... Trying
to keep up with that, that was difficult. But the single most
outstanding memory from that conference, and we will also leave
this person anonymous. But there was an executive at Microsoft who
was hotter on XML than either you or I, which is hard to believe,
right. And we ended up with the sponsored after hours session at
this conference. You remember this? You see...
Brian Jones (00:29:45):
I do.
Rob Collie (00:29:46):
You know where we're going. Okay. So this was a 30 minute sponsored
by Dell or something. Right. It was a 30 minute session, at 5:00
PM, at the end of a conference day where everyone's trying to go
back and get to the bars or whatever, right.? But, it's a Microsoft
executive, it's Dell sponsored, we'll show up. And the plan was at
the end of this 30 minute talk given by this executive, he was
going to bring all of us up on the stage to show everyone the team
that had done all of this, right? Great plan. Except it was the
worst presentation in history. I remember it running for two hours.
It was so bad that we started off with 200 people in the room and
at the end of it, and I'm just like an agony the whole time cause
like I'm associated with this, right?
Rob Collie (00:30:31):
At the end of these two hours, or what felt like two hours anyway,
it was easily 90 minutes. There's five people left in this room of
200 and it's not like the presentation is adapted to the fact that
it's a smaller audience. It's just continued to drone on exactly as
if everyone was there, right? And I'm sitting here thinking, "Okay,
he's not going to call us all up on this stage. There's been more
people on the stage than in the audience. If he does this, he's
clearly not going to do that." And then he did and we all had to
parade up there and stand there like the biggest dodos. I've never
been more professionally embarrassed I don't think, than that
moment.
Rob Collie (00:31:14):
And we're all looking at each other as we get up out of our seats
like, "Oh my god."
Brian Jones (00:31:19):
I definitely remember this.
Rob Collie (00:31:22):
I don't see how you could have forgotten.
Brian Jones (00:31:23):
Well, yeah. And the person that we're talking about is actually one
of my favorite people on the planet. I totally... I love this guy.
I view him as like a mentor and everything, but... Which makes me
remember it even more.
Brian Jones (00:31:34):
I think it was just, there was so much excitement. There'd been so
much build up to this and this was like a kind of crescendo right?
Of bringing this stuff. We probably should have had it a little bit
shorter.
Rob Collie (00:31:46):
I mean when it reaches the point where clueless, mid twenties, Rob
Collie is going, "Oh no, this is not the emotional, this not the
move." You don't do it.
Brian Jones (00:31:58):
I'm no longer excited about being called up.
Rob Collie (00:32:04):
So from my perspective, you kind of parlayed that experience of the
XML and all that kind of stuff. I think you did a really fantastic
job of everything you guys did on that product. Again, it was the
relevance that ultimately fell flat for both of us right. I guess
in the end, the excitement with XML wasn't really all that
appreciably different from the excitement about Bluetooth. I mean,
it's everywhere, right? XML is everywhere. Bluetooth is everywhere
and neither one of them really changed things in terms of what
Excel or Word should be doing. It seemed like you played that into
this file format second act. And I think very, very, very
effectively, actually there was a little bit of controversy.
Rob Collie (00:32:43):
Let's set the stage for people. This was the 2007 release of Office
where all the file formats got radically overhauled. This is when
the extra X appeared on the end of all the file names, right?
Brian Jones (00:32:58):
Yeah.
Rob Collie (00:32:58):
There was a controversy internally. Kind of starting with Bill
actually. That we shouldn't make well-documented transparent file
format specs, right. There was this belief that the opaque file
formats of the previous decades was in some sense, some big moat
against competition. And of course, a lot of our competitors
agreed. Tailor out in the public saying, "Yes, this is a barrier to
competition. It's a monopolistic, blah, blah, blah." We, Microsoft
had just gotten its ass kicked in the Anna Truss case. So it was
really interesting. I credit Brian, your crew, with really
advocating this very effectively. That's a difficult ship to turn.
First of all, you got all these teams to buy into all this extra
work, which no one wants to do. But when it's not even clear
whether you have top level executive support, in fact, you might
actually have C-suite antagonism towards an idea. To get it done.
That's a career making achievement. I'm sure you remember all of
that. Right. But what are your reactions to that controversy? Do
you remember being in the midst of that?
Brian Jones (00:34:12):
I do. It was definitely a long running project. It evolved over
quite a number of years. The beginning of it was, in that previous
release, the XML stuff you and I were talking about was more about
what we called "Custom XML". Right? So people would go and create
for themselves. But in that same release, we had Word, we outputted
an XML format that was our definition, which we called "Word ML"
and Excel did a similar thing. Words' we try to make full fidelity.
So you could save any word document in the XML format. Excel's was
kind of a tailored down, it was less about formatting, it was more,
"Hey, here's like..." It's almost like, "Here's a better version of
CSV, right. But we're going to do it as XML." And so we already had
a little bit of that.
Brian Jones (00:34:53):
And the whole reason we were looking at that was, on the Word side,
for instance, a lot of the customer issues that we'd get where
people would have corrupt files, they were corrupt because they
there'd be some add-in that they had running or some third party
app that was reading and writing word files. The files were fairly
brittle and complex. The binary format... The binary format was
written back in the days of floppy disks, right? So the top
priority was how quickly can you write to a floppy disk and read
from a floppy disc, right? It wasn't about, how easy is this for
other people to go and read and write? Not because it was on
purpose, make it hard. It was just the primary bid is let's get
this thing so it's really easy to read and write from floppy,
right?
Brian Jones (00:35:31):
And so in Word, we were like, "Wow, I think that there's a bigger
opportunity here for an ecosystem around Word if we make it easy
for people to read and write Word docs and build solutions around
them." And so then the next release, the Excel team was looking at
doing some big changes around a lot of the limitations, like how
many rows you could have in columns, right. Lengths of like
formulas and things like that. Right. And so there was this thing
where the Excel team was like, "We are going to need to create a
new file format." And on the word side, we thought this XML thing
was great. We want to move to that as our new format.
Brian Jones (00:36:01):
And so everything kind of came together and it was clear. Hey, this
is going to be the release that we are going to go and rev our file
format, which we hadn't done in a while. This is also the release
of the ribbon. So there were two really big major changes in that
product, right? It was the new file format and the ribbon. It's
funny. I still refer to it as the new file format, even though it's
15 years old.
Rob Collie (00:36:23):
Yeah. It's the new file format it's still new, yeah.
Brian Jones (00:36:25):
I still call it that, which is kind of nuts. But I think that the
controversies you were talking about was really more of a... Boy,
this is a really big deal for the product. We had changed file
formats before in the past and not necessarily gotten it right. And
there were a lot of challenges around compatibility and stuff. And
so there was just a lot of worry of let's make sure you all have
your stuff together here, right? Like let's make sure that this
doesn't in any way break, stop people from wanting to upgrade to
the new version. But it went really well. The whole goal of it was
let's get something that we think third parties can go and read and
write, and this is going to help build an ecosystem. And a new
ecosystem run Office. Office already had big ecosystem with VBA and
COMM add-ons and stuff like that, right.? But we won't have this
new ecosystem around our file formats as a thing. That's why we
chose... There's a packaging layer, which is all zip based. So if
people haven't played around with it that XLSX, you can just put a
.zip at the end and double click it. And it's just a zip file. And
you can see a whole bunch of stuff inside of it. Right?
Rob Collie (00:37:23):
Yeah. If you're listening, you haven't done that go right now, run
don't walk, grab an Excel file or a Word file, whatever. Go and
rename the XLSX or BPTX, go ahead and rename it so that it ends in
.zip and then open it up and you'll be blown away.
Thomas Larock (00:37:38):
PowerPoint is my favorite when I have to find some unknown setting
that I need and I can just search through the whole thing. Yep.
Rob Collie (00:37:45):
Or all the images. You want to get all the images out of the
PowerPoint file. It's just a zip file that has a bunch of images in
it. Right.
Brian Jones (00:37:50):
So I also did this for backpack. It's the same thing. You can crack
open the backpack by renaming a zip file...
Thomas Larock (00:37:58):
An actual physical backpack? What are we... what are we talking
about here?
Brian Jones (00:38:03):
Ah yeah.
Rob Collie (00:38:03):
This is the digital acetate that is over the top of the entire
physical world that you aren't aware of.
Thomas Larock (00:38:08):
Digital acetate, that's it? That's it. That's where the podcast
peaks. Right? Those two words. We're all going home now.
Brian Jones (00:38:19):
Yeah. No. A SQL server, there's DAC pack, which is just the, say
database schema. Then there's a backpack which has the data and the
schema combined. But you can, if you rename them . zip, you can
crack them open to see the XML that makes up those forms. So it's
not just office products.
Rob Collie (00:38:37):
We ended up standardizing the entire thing, but that packaging
format, it was called OPC, Open Packaging Convention, or something
like that. It was something that we did in partnership with a
Windows team. It's part of the final ISO standard for our file
format. And then there were a lot of other folks that went and used
that exact same standard. Because it's a really easy way of you
have a zip package. You can have a whole bunch of pieces inside of
it, which are XML. And then there's this convention for how you can
do relationships between the different pieces. So I can have a
slide. That's an XML and it can declare relationships to all the
images that it uses. And that way it's really quick, easy to know,
okay, here's all the content I need to grab if I want to move
pieces of it outside of the file.
Rob Collie (00:39:16):
So the single coolest thing I've ever done with, we'll just call it
your file format Brian. We'll just pretend that it was only you
working on that.
Brian Jones (00:39:23):
Just me yeah, I was pretty busy, but yeah.
Rob Collie (00:39:27):
So the very, very first version of Power Pivot, first of all, your
file format, the new file format made Power Pivot possible. We
needed to go and add this gigantic binary stream of compressed data
and everything, everything about Power Pivot needed to be saved in
the file. At the beginning of the project, everyone was saying,
"Oh, no, we're going to save it as two separate files." And I'm
like, "Are you guys kidding?" The Pivot cache, for instance, is
saved in the same file. You can't throw a multi file solution at
people and expect it to... This was actually like Manhattan
project, just to get that stream saved into the same file. It was
pretty crazy. However, when it was done, there was something really
awesome I wasn't aware of until the very end, which was, first of
all, you could open up a zip file and just tunnel down and you
would find a file in there called item one.data.
Rob Collie (00:40:21):
Okay. That was the Power Pivot blob. That was everything about the
Power Pivot thing. And it was by far the biggest thing in the file,
like it was like 99% of the file size was what was there. However,
as this backup, someone had decided, I had nothing to do with this,
to save all of the instructions. I think it's called XML for
analysis XMLA. All of the instructions that would be required to
rebuild exactly that file, but without any of the actual binary
data in it. So it was a very, very small amount of XML. Okay. So
here's what we would do because there were no good automation, no
interfaces, no APIs. If we needed to add like 500 formulas to a
Power Pivot file, you could go through the UI and write those 500
formulas, type, click, type, click, type, click.
Rob Collie (00:41:08):
Okay. So what we would do, and my first job outside of Microsoft,
is we would go in there and we would edit that XML backup and add
all the formulas we wanted in it. And by the way, I would use Excel
to write these formulas. I would use string concatenation and all
of that kind of stuff to write these things. It was very, very,
very sensitive, one character out of place in the whole thing
fails. So you make those changes. You save the file, reopen it,
nothing happens because it's just the backup. Okay. So then you've
got to go and you've got to create a zero byte item one.data file
on your desktop and you copy it into the zip file and overwrite the
real item one.data, therefore deliberately corrupting the primary
copy. So when you reopen the file it triggers the backup process
and it rehydrates with all of your stuff, it was awesome.
Rob Collie (00:41:57):
And then a couple of releases of Power Pivot later, suddenly that
didn't work anymore and I was really pissed. But it just really
shows you, it opens up so many opportunities that you never would
have expected. And even a hack like that, that's not the kind that
you'd be really looking for, but the fact that something like that
even happens as a result of this is really indicative of what a
success it was.
Brian Jones (00:42:19):
Yeah. I mean, there's a lot of those things where, I love building
platforms, like that's my favorite part of the job. It's all those
things that you see people do that you never would have predicted.
Right? That's just so exciting. PowerPoint had this huge group of
folks that would go and build things like doc assembly stuff,
right. Where they go and automatically build PowerPoint decks on
demand, right? Based on who you're going to go and present to cause
they've just shredded the thing. In fact, when we did the ISO
standardization, it was a 6,000 page doc that we had to go. And we
built and reviewed with a standards body and we did it over about a
year. Which sounds nuts, a 6,000 page doc in about a year. And the
way that we were able to do that is there was never really a 6,000
page doc. There's a database where there's a row for every single
element and attribute in this, in the whole schema, that would then
have the column which is the description, which would just be the
word XML.
Brian Jones (00:43:09):
And so we could, on demand, at any point, generate whatever view or
part of the doc we wanted. So we'd say, "Hey, we're going to go in
now, review everything that has to do with formatting across Word,
Excel and PowerPoint." And so we just click a couple buttons and
the database would spit out a Word doc that was just that part.
Everybody could go and edit it cause we were using the structured
elements we'd added to Word, which is called content controls,
which was the next version of that XML stuff that we had to
deprecate. And then the process, as soon as you'd finish editing
that Word doc, we just submit it back. The process would go back
and shred that Word doc again and put it all back in the database.
And so we really used the file format to bootstrap documenting the
file format.
Rob Collie (00:43:48):
And then when you dump a 6,000 page document on someone, they have
no choice. But to just say, yep, it looks good to us.
Brian Jones (00:43:55):
Well, there was a pretty, incredibly thorough review still. It was
just pretty impressive. The final vote that we had in Geneva, the
process leading up to that, the amount of feedback that we got.
Cause basically the ISO, you can kind of think of it like the UN,
you go and show up and every country has a seat, right? I mean, not
everybody participates, but anybody that wants to can. And so yeah,
we had to respond to thousands of comments around different pieces,
things that people wanted to see changed.
Rob Collie (00:44:22):
Yeah. I can imagine, right. Think about it. You just said at the
final vote in Geneva. That's a heavy moment man.
Thomas Larock (00:44:29):
Yeah. That threw me off for a second. I thought, for sure, you were
talking Switzerland, but now thinking that was just a code
name.
Rob Collie (00:44:38):
No, I think, I think he was actually in Switzerland.
Brian Jones (00:44:40):
In Switzerland.
Rob Collie (00:44:41):
Have you seen the chamber where they do these votes? It looks just
like the Senate from episode one of Star Wars. It's just like that.
It's pretty heavy.
Brian Jones (00:44:51):
The little levitating...
Rob Collie (00:44:53):
The floating lift. Yeah. I think they call that digital acetate. I
think that's what they call that. By the way on the Excel team, the
way I came to look at the new file format and the open architecture
of it, again, this this will show you how quickly I had turned into
the more cynic side of things. Well, okay. We're going to be
changing file formats. And we're doing that for our benefit because
we didn't have enough bits allocated in the 1980s version of the
file format that was saved to floppy disc, as you pointed out,
right. Who could ever imagine having more than 64,000 rows, it's
just inconceivable or 250 columns or whatever, right.? Because we
hadn't allocated that. We'd made an engineering mistake,
essentially, we hadn't future-proofed. So we need to make a file
format change for our benefit, right. To undo one of our mistakes.
And the way I looked at it was, "Ooh, all this open file format
stuff, that'll be like the 'Look, squirrel!'" To distract people
and to sort of justify, while we went and did this other thing,
which, ultimately it actually went pretty well. The transition for
the customers actually wasn't nearly as bad, because we actually
Took it seriously.
Rob Collie (00:46:03):
The transition for the customers actually wasn't nearly as bad
because we actually took it seriously. We didn't cut any corners.
We did all the right things.
Brian Jones (00:46:07):
Well, there were several benefits too. We were talking about all
the kind of ecosystem development benefits, but the fact that the
file was zipped and compressed right, it meant that the thing was
smaller. And that was all of a sudden, it was no longer about
floppy discs. People are sharing files on networks. And so actually
being able to go and have a file that's easier to share, send over
network because it's smaller was a thing.
Brian Jones (00:46:26):
There were a couple of things that we were able to go and
highlight. There's also a pretty nice thing where it was actually
more robust because it was XML, and we split it into multiple
pieces of XML. It meant that even if you had bit rot, you would
only lose one little piece of the file, whereas with old the binary
format, you had some bit rot and the whole thing is impossible to
open up.There are a couple of things that were in user benefits
too, which helped.
Rob Collie (00:46:50):
And ultimately, on the Excel side, the user got a million row
spreadsheet format and the ability to use a hell of a lot more than
like 14 colors that could be used in a single spreadsheet or
something. It was .like a power of two minus two, so many bizarre
things. Like Excel had more colors than that, but you couldn't use
more than a certain subset in a-
Brian Jones (00:47:10):
At a time, yeah.
Rob Collie (00:47:10):
-In a single file. So yeah, there were a lot of benefits. They just
weren't-
Brian Jones (00:47:15):
It's not like it's an explicit choice. It's just that at the time
somebody is implementing something, you're right in a way,
assuming, "Oh, this is fine. This is enough. I'll never have to
worry with this issue."
Rob Collie (00:47:25):
Why waste the whole byte on that? When you can cram four different
settings into a single byte. If you read the old stories about
Gates and Allen programming up at Harvard, they had these vicious
head-to-head competitions to see who could write the compiler or
the section of basic in the fewest bytes possible. This was still
very much hanging over Microsoft, even the vestiges of it were
still kind of hanging over us even when I arrived. But certainly in
the '80s when the Excel file format was being designed for that
rev, it was still very much like, "Why waste all those bits in a
byte?" "Let's cap it at four bits".
Thomas Larock (00:48:05):
In that blog series from Sinofsky, he talks a lot about that at the
early start. And I'm at a point now where he's talking a lot about
the code reuse because the Excel team, the Word team, I guess
PowerPoint, but all these other teams, were all dealing with, say,
text. And they were all doing their own code for how that text
would be displayed and shown. And Bill would be the one being like,
" This is ridiculous". "We should be able to reuse the code between
these products". And to me, that would just be common sense. But
these groups, Microsoft just grew so rapidly so quickly, they were
off on their own, and they have to ship. I ain't got time to wait
around for this, for somebody to build an API, things like that.
I'll just write it myself.
Brian Jones (00:48:50):
It's a general thing that you get as you get larger where the
person in charge that can oversee everything is like, "Well, these
are all my resources", and, "Wow, I don't want three groups all
building the same thing". But then when you get down, there's also
a reality of we're just going to have a very different view on text
and text layout than Excel. And Excel is not going to say, "I want
all of that code that Word uses to lay out all of their content to
be running for every single cell". Right? That's just suboptimal.
And so it's always this fun conversation back and forth around
where do you have shared code and reuse and where do you say it's
okay for this specific app to have this more optimized thing that
might look the same, but in reality, it's not really the same.
Rob Collie (00:49:33):
Brian, do you remember the ... I'm sure you do, but I don't
remember what company they were from. But at one point in this file
format effort, these really high priced consultants showed up and
went around and interviewed us a couple of times. Do you remember
that phase? It was like-
Brian Jones (00:49:51):
Was that towards the end? There was a couple summary stories that
were pulled together just to talk about the overall processes. It
was actually after the standardization.
Rob Collie (00:49:58):
I remember this being at the point in time where it was still kind
of a question. whether we should do it.
Brian Jones (00:50:02):
I don't remember that.
Rob Collie (00:50:04):
The thing I remember really vividly is a statement that Chris
Pratley would make over and over again, this encapsulated it for
me. I came around to seeing it his way, which was the file format
isn't the thing. That's not the moat. The thing that makes Office
unique is the behaviors of the application. It's not the noun of
the file format. It's the verb of what happens in the app. It's
instructed to think that even if you took exactly the Excel team
today, every single person that's already worked on it, and said,
"Hey, you have to go rebuild Excel exactly". There's no way that
version of Excel would be compatible with the one we have now. It
would drift so much.
Rob Collie (00:50:43):
You could even have access to all the same specs. We would even
cheat and say, "Look, you can have access to every single spec ever
written". So? It was clearly someone had thought it was time to
bring in like a McKinsey. They were all well dressed. They were all
attractive. They were all a little too young to be the ones sort of
making these decisions. It was just really weird to have them show
up, three people in your office. Like, "Okay, I'll tell you what's
going on".
Brian Jones (00:51:11):
I can totally imagine. It's funny I don't remember that. There were
several rounds of analysis on how we were doing it, what we're
doing and making sure we were doing it the right way. But yeah,
Chris is spot on. I mean, your point about rebuilding it, that's
essentially what we've been going through for the past five plus
years around our web app. It's a lot of work. Unfortunately, we
can't let it drift. The expectation from everybody is, "Hey, I
learned the Wind 32 version. When I go to the web, I want it to
feel the same. I don't want to feel like I'm now using some
different app."
Rob Collie (00:51:44):
What an amazing, again, like a Manhattan project type of thing,
this notion of rewriting Excel to run on the web and be
compatible.
Brian Jones (00:51:55):
Yeah, with 30 years of innovation.
Rob Collie (00:51:56):
Yeah. That started in the 2007 release. Excel services, the first
release of Excel services was 2007. And this whole thing about
shared code, like what features, what functions of Excel, what
pieces of it were going to be rewritten to be quote unquote "shared
code"? And shared code meant it was actually server safe, which
none of regular desktop Excel written in the early '80s, still
carrying around assembly in certain places, assembly code of all
things, right? Excel was not server safe. It was about as far from
server safe as you could get. And so to rewrite this so ambitious
without breaking anything. Oh my God. What a massive ... This dates
back, gosh, more than 15 years.
Brian Jones (00:52:45):
Yeah. I'd say like the first goals around it were a bit different,
right? It wasn't a web version of Excel. It was like BI scenarios
and how can we have dashboards and Excel playing a role in
dashboards. But yeah, I'd say since I joined, it was probably maybe
a half a year or a year into when I joined, we just made the
decision to shift a huge chunk of our funding to the web app. It
was just clear that we need to make even more rapid progress. If
you go, we have a site where you can go and see all the features
that are rolling out there. It's incredible. And it's just because
of the depth of the product. "Wow that's so many features you've
done. You must be almost done". But then you look at everything
else that's still isn't done yet.
Brian Jones (00:53:23):
Now thankfully, we're getting to the point where we can look at
telemetry and say, "Hey, we've got most people covered." Most
users, when we look at what they do in Windows, they could use the
web app and shouldn't notice a difference. But there still is a set
of things that we're going to keep churning through. So that'll
continue to be a huge, huge investment for us. But yeah, the shared
code strategy, we have an iPhone version, an iPad version an
Android version. We've got Excel across all platforms. And because
of the shared code, when we add new features, the feature crew
that's working on that, they need to have a plan for how they're
going to roll out across all those platforms, clearly levered
shared code. But they also need to think through user experience
and stuff like that too. Clearly a feature on a phone is going to
behave differently than it's going to behave on a desktop.
Rob Collie (00:54:05):
Part of me, just like, kind of wants to just say, "I don't even
believe that you've pulled that off, there's no way". It's kind of
like, I've never looked at the Android version, and until I look at
the Android version, I'm just going to assume it's not real. This
is why it's one of the hardest things imaginable to have a single
code base with all these different user experience, just
fundamental paradigms of difference between these platforms. Like
really? Come on.
Brian Jones (00:54:34):
It was a massively ambitious project. Mac shifted over maybe three
years ago. And that's when, all of a sudden, in addition to a bunch
of just features that people have been asking for that we'd never
been able to get to, the massive one there was we were able to roll
out the co-authoring multiplayer mode for Excel.
Rob Collie (00:54:50):
Multiplayer.
Brian Jones (00:54:52):
That's the term I like for co-authoring. It's more fun.
Rob Collie (00:54:55):
Yeah. It's like MMO for spreadsheets.
Brian Jones (00:54:57):
Yes. We were able to get that for the Mac. I mean, all of our
platforms. One of us can be on an iPad, an iPhone, the web app, and
we'll all see what we're doing in real time, making edits and all
of that stuff. That alone, if you want to talk about massive
projects, 30 years of features and innovation, basically that means
we had to go and teach Excel how to communicate to another version
of Excel and be very specific about, "This is what I did." "Here's
the action I took." And that is massive. There are thousands and
thousands of things you can do in the product. So getting it so
that all of those versions are in sync the entire time, and so
we're all seeing the exact same results of calc and all of that.
That itself was a huge, massive project.
Rob Collie (00:55:37):
Take this as the highest form of praise when I say I don't buy it.
I can't believe it.
Brian Jones (00:55:44):
I hope everybody's okay that we just talked for like an hour on
just like listening to somebody at a high school reunion, I think,
or something. Is this like me talking about how great I played in
that one game? And you're like, "Yeah, that was a great
basket".
Rob Collie (00:55:54):
Yeah. "Man, my jumper was on". the thing that's hard to appreciate,
I think, is that you got to come back to the fact that we're
talking about the tools that everyone in the world uses every day,
that we rely on. And I think being gone from Microsoft for the last
12 years, I'm able to better appreciate that sense of wonder. This
isn't just you and I catching up, I don't think. People enjoy, for
good reason I think, hearing the stories of how these things came
to be. People don't know by default how hard it was to get to a
million rows in the file format. If you're like a robot, you're
like, "I don't care how I got here. I just care what it is", then
you're not listening to this show. We call it data with a human
element. Robots can exit stage left. I think you should feel zero
guilt. This isn't just self-indulgence.
Brian Jones (00:56:55):
Well, on the off chance everybody else ... I've listened to a lot
of Rob's other podcasts, and they're awesome. So if you're bored
with this one, it's okay. Go check out some of the other ones.
They're great.
Rob Collie (00:57:06):
Imposter syndrome rears its head. This guy, Brian, all he does is
run the mission for Excel. I mean, he's, he's not good enough to be
on our show.
Brian Jones (00:57:17):
And talk about his past, I guess, all the time.
Rob Collie (00:57:21):
If you struggle with imposter syndrome, right there should be your
antidote. Right? We just heard ... Trust me, Brian, you're
okay.
Brian Jones (00:57:31):
Thank you, Rob.
Rob Collie (00:57:32):
We just talked about all this gargantuan effort, intelligent, but
also heavily resource and effort intensive. You can't just be smart
about this and then suddenly it's done in a month, you've got to be
smart every single day on every single little microscopic detail
while grinding for years to do these sorts of things. And at the
same time, sort of at the same time, there's also been this
JavaScript API. Holy hell. So file this also "under things that
can't be done", a version of the office object model and API that
can be used for local desktop automation, like VBA has forever, but
also to be used for server automation of the cloud-based web based
apps.
Brian Jones (00:58:22):
Ooh. Yeah.
Rob Collie (00:58:23):
And I haven't had the time or the energy to get into this. I have
not written a single line of code against this new API. But when I
started to hear things like ... Yeah, we had little tiny things
like this even back when I was at Microsoft, there were a couple of
APIs that would work on the server, but they were completely
different from the desktop version. And they were an egregiously
tiny subset, just the least ambitious things possible. And so I
kept using the criteria of, until you give me a range object
equivalent in this fancy blah-blah-blah API that you're working on,
Microsoft, it ain't real. It certainly sounds like y'all have that
now, which is nuts.
Brian Jones (00:59:11):
Yeah. It's funny. This one could actually be a long history
discussion because we started this project actually 10 years ago,
basically around when my daughter was born, where the first step
was really looking more pro devs. But it was the idea of we're
moving cross plat. It was still early days for the web app. We
weren't really working on mobile yet, but it was clear, mobile apps
were going to start to become a thing. And so it was like, hey, VBA
and COM APIS just aren't going to be across plat, so we need to
figure something out here. And I would go and talk to a bunch of
developers. I'd go to build conferences and stuff like that. And
they'd talk about how challenging it was to specialize in office
development because so much of the tech that they would learn was
not applicable in other contexts.
Brian Jones (00:59:54):
And so we start thinking, "Okay, well what are the texts that they
know?" We were looking at things like that .net. And we eventually
started looking at well, there's obviously a ton of people doing
web development, HTML, JavaScript. So what if we just said
extending Office was just like writing a webpage or building a web
app? And I'd say the first round of it, it was a little bit more
like data visualization type of scenarios. We would go and kind of
I-frame in your web app into Excel. You could write some HTML. But
that was the starting point to start to build out an API. And it
just, back to the same thing we were talking about the web app, 30
years of API innovation, it takes a long time to get to where we've
got an API that covers most of the things you can do in the
product.
Brian Jones (01:00:38):
And so we stayed really focused on ProDevs for a long time, because
to do an end user, kind of like the VBA, equivalent where you can
do Mac recording and stuff like that, you've got to cover
everything somebody can do. Because it's kind of weird to say
record macro, and then only 20% of the things you do actually get
recorded. And so we reached that point about a year or so ago. And
so I think it was last spring, we went and released the beta
version of this Office scripts. And then we just GA-ed the spring.
Right now, the main platform we run on for the macro recording
piece is all on the web. But of course, the plan is to go across
plat and desktop and everything too. But it's kind of neat. It's an
example of, there are certain cases now where going out first on
the web where if you went back like four years ago, everything was
Windows first and then eventually it would show up, hopefully, on
web.
Rob Collie (01:01:29):
First of all, I'm crazy impressed that this has happened. I just
tuned out of this whole conversation many years ago because I just
felt like you're never going to get there. And then you just
sneaked up on me. It was like the tortoise and the hair. The other
thing that's really impressive to me is sort of the organizational
discipline and willingness to make those long commitments, the
things that aren't going to necessarily pay off in 6 to 12 months,
or even 18 months. To be on an initiative that does take many years
to deliver, that's first of all, a non-Microsoft mindset. I
typically associate that with Microsoft as an outside observer. And
secondly, you're in this world now where you're releasing on a much
faster tempo. It would seem like it would even reduce the
chances-
Brian Jones (01:02:20):
Yeah, it's going to be harder-
Rob Collie (01:02:21):
Further. That you could keep your eye on a project that honestly,
like you say, it honestly doesn't have any payoff at 90% done.
You've got to get to a hundred before you can reap any of the
benefits. It just seems like such delayed gratification, many,
many, many years of it. Hats off. I was on some sort of usergroup
virtual meeting, and someone from your team, someone from the Excel
team was there. I know them, but I just forget who it was. And I
was laughing about my wheel of inquisition, which is the pie chart
that I spin with VBA, at a random distance to pick who has to
answer the next question or who has to ask the next question on
team meetings. I just said, "Look, this is an example of why you
still need desktop Excel". And your team member interjected and
said, "No, I actually think you can do this in the Office script
API now". And I'm like, "Get out of town. Just shut up". That was
the moment that I kind of started to pay attention.
Brian Jones (01:03:16):
Yeah. It's gotten a lot more real over the past few years. We had a
similar experience with a lot of the MVPS where it was kind of
like, okay, I have a set of things that I use come in VBA for, I
don't see you solving those yet. But now it's gotten a lot more
real. I don't know how much you worked with Tristan, but this is
all ... Tristan Davis runs this team. And they're just doing a
phenomenal job.
Rob Collie (01:03:36):
Tristan wasn't really around when I was there. I left the Excel
team in 2006, late 2006, to go work on fantasy football over
in-
Brian Jones (01:03:45):
Oh yeah, that wasn't just a hobby. You actually went to a team. I
forgot about that.
Rob Collie (01:03:47):
I actually did, yeah, the team that existed for six months and then
evaporated. And I got folded into being, which was terrible. But
then I ended up on Power Pivot. So all is, well that ends well. But
by the way, the last time I saw the vice president who's left
Microsoft, but the VP who was in charge of Excel for many years,
the last time I saw him, I was working on Power Pivot. And he
already announced he was leaving. And he just sort of like, "oh, so
friendly, good to see you", you know, whatever, just in the
hallway. And he's like, "Hey, you know, I still use you, Rob, as a
cautionary tale to people on how to not mess up their career.
Brian Jones (01:04:20):
Wow. Oh my God. That's great.
Rob Collie (01:04:22):
He wasn't expecting it to land is a bad thing. He had no idea. And
I'm like, "Oh, that's great". I'm slapping him on the back. "That
is the best thing ever". He had no idea. Actually, in the end it
was fantastic for my career, that whole path. I wouldn't have it
any other way.
Brian Jones (01:04:39):
But that wasn't the point he was making at the time.
Rob Collie (01:04:41):
No, that Was not the point he was making. He didn't bother to find
out whether it was working out for me. So I just buttered you up. I
just told you how awesome it is that y'all have the discipline and
the patience to set a multi-year goal and deliver on it. I got
another one for you that's in this category, which is, let's get
serious about onboarding the VLOOKUP and Pivot crowd into data
modeling. Let's do that. By the way, observation about Microsoft
not being willing to invest in anything that's not six months out,
that impression of mine is primarily built from watching this
problem for the last decade, decade plus actually. There's never an
answer that's six months away. It's a hard problem. It's just
fundamentally a really hard problem. And it's not one that you can
just set aside like a two person team and have them take care of
it.
Rob Collie (01:05:39):
But if we ... I'm going to put myself on the same side of the table
because I understand, I think, a lot of the trade offs. If we
collectively had been doing this even five years ago, it would have
been over by now. It's like when you're on a diet, "Oh, if I'd only
started six months ago, I'd already be skinny". I am encouraged by
the stepped up cooperation between the Power BI and Excel teams.
And yet I still think there's something massive being missed here,
which is that all of the truly valuable data model builders for
Power BI, all of the future builders, it just rounds to 100%.
They're all Excel if you look up in Pivot people today. There's
this sort of myth that it's an IT thing. It's not, it's not an IT
thing. It's the citizen developer type embedded in the
business.
Rob Collie (01:06:34):
Even just emotionally like we were talking about earlier about the
people who, when they see something, they get frustrated that they
didn't know about it. I mean, this is our life. We're not just a
training company. In fact, we do a lot more project consulting now
than we training. Training is becoming less and less of an
important part of our business. But it's still sort of part of our
DNA. So many times people would just like scream from the back of
the room. Like, "No! What?" I can go to a finance conference, like
an FP&A conference, and show this off and be positive that 90%
of the room plus has never seen any of it. It's life-changing to
them.
Rob Collie (01:07:09):
It's almost like we owe it to them. And everyone that listens to
the show knows that this is my soap box issue. I would be so
inauthentic if I didn't bring it up with you. I don't expect you to
go, "Hey Rob. Yeah, that was it. This conversation was the thing we
needed. Now we're going to go and do it". But I can't omit this
one. So anytime I can lobby to add this particular topic, maybe
that multi-year payoff gene. Oh my God. Let's do it.
Brian Jones (01:07:39):
Yeah. And just for folks that for the listening audience, this
isn't the first time I've heard this from Rob. He's only six months
into my job. And Dave, Dave Gayner, who's my boss, he had had my
job. And then he moved up and became a big wig. He said, "Hey, one
of the things you should probably do is go and just get Rob
Collie's training because you'll kind of see a bunch of customers
and how they use the product, and you'll also learn from him". And
it was neat because your next training was in London, so I got to
go travel to London to go and get the training. So that was kind of
a nice perk too.
Rob Collie (01:08:08):
Yes, it was.
Brian Jones (01:08:10):
Well, and yeah, I learned a ton. Same thing. I was like, "Wow, this
is pretty cool. I didn't know my product could do this", because
I'd been on the job for a few months. In fact, I remember one of
the people that was in the session was sitting next to me, and
you'd introduced me and who I was during the sessions.
Rob Collie (01:08:25):
Of course, we got to put the target on you.
Brian Jones (01:08:26):
Yeah, of course. And so I was making some comments, and the guy was
just ... I felt like he was just really offended that I didn't
know. He's like, "How can you be in charge of this product and not
know this stuff?" And so I wanted to go into the background of
like, "Well, I just joined, and that's how it works". But, of
course, I got super defensive.
Rob Collie (01:08:44):
You're welcome, Brian.
Brian Jones (01:08:45):
But I can't remember ... I think you and I are having beers or
something afterwards. And you're just like, "If you just went to
every airport and put up posters that had a V lookup with a
question mark and then Power Pivot below, that would do it. Right?
It's resonating. I still remember that.
Rob Collie (01:09:02):
That's good. That's long been one of my-
Rob Collie (01:09:03):
Yeah.
Brian Jones (01:09:03):
It's resonating. I still remember that.
Rob Collie (01:09:03):
That's good. That's long been one of my favorite things.
Brian Jones (01:09:05):
There's a lot of stuff that we've done over time to try to... like,
for instance, we've done things to try and make people more aware
of Power BI. People have a Power BI license and they're doing stuff
in Excel, we'll go and say, "Hey, did you know about power BI?",
and try and go and take them off to discover that, because they
might be able to benefit from it. Right? And over the long run, we
are working with the Power BI team, we want to make sure that we've
got better compatibility with the engine and the data models that
are run. Right? So you can, whatever you build in Power BI, you can
bring into Excel and vice versa. And so there's clearly like a
north star we have there around the user not having to think, oh,
am I using Power Pivot or am I using Power BI? And those are two
different decisions they have to make. That just kind of feels
weird. It's basically you're doing the same job.
Brian Jones (01:09:49):
And as you know, Rob, like you working on the product, you can't
just put a big button in the ribbon and all of a sudden everybody's
going to discover it.
Rob Collie (01:09:56):
No. I-
Brian Jones (01:09:56):
Right? Nobody sees it. So that's where the trick is trying to
figure out those points in time where it's like, hey, we think
someone would benefit from this higher level of feature. How do we
go and educate them about that? And that's where we're starting
some of the stuff I talked about with AI. We're doing that, but at
a lower level, like you've got a table of data and we say, hey, did
you know you can summarize this with a Pivot table? And we're going
to iterate on and see how that works. But this would like...
Educating people on Power BI will be a thing that we'll continue to
look to do based on the activities that folks are doing. And some
of the triggers you've thrown out are great ones, right?
Rob Collie (01:10:31):
Talking to someone from Microsoft recently on this show, I made
kind of exactly that recommendation. It Was like, forget it. Even
if it was just an awareness thing, the really hard problem is to
look at what they're doing and go, oh, this particular file that
you're working with right now, we can turn it into a Power Pivot
file or a data model. That's crazy. That is really, really
difficult. But identifying the people who are good candidates to
build data models and therefore would change their lives in the
process, right?
Brian Jones (01:11:05):
Yeah.
Rob Collie (01:11:06):
Everybody wins here. Identifying those people by their behavior,
that shouldn't be hard. And then if it's just an on-ramp, "Hey, you
are one of us," right? You are one of this chosen few, and here's
the URL. It just takes you to like a YouTube video that's five
minutes long, it doesn't try to teach anything, it just shows what
they're missing, God, that would be so profound. And that I
wouldn't care about Power Pivot in Excel. That wouldn't matter so
much.
Brian Jones (01:11:37):
Yeah. It's probably a thing that you haven't seen because you
probably wouldn't have gotten the notification, but there already
is in the product. I can't remember the exact triggers, but we went
and looked at some data to see who are the people that would
benefit from Power BI and what are the behaviors in the app? I
don't remember if it was using VLOOKUPs, or it also had to do with
what kind of things they were sharing because it turned out if they
were sharing then they had more likelihood that Power BI would be
super useful because they could go and publish a thing. But there
was a set of folks that we would actually go and notify. We'd have
a little toast that would come up saying, hey, do you know about
Power BI? Here's what it'll do for you. Click here to go and learn
more. And that was super effective. It ended up being that the
logic that we picked for who we'd share that with the click-through
rate was super high. And the Power BI team saw that not only did
they go through, but then they actually became engaged Power BI
users.
Brian Jones (01:12:21):
But those are the things that Arun's team and my team. So for folks
that don't know, Arun runs the Power BI team. We already look at
that and work on that. And then there's a ton of stuff that we've
announced around silly things that just felt more like bugs, but
actually were a good amount of work, like our web app wouldn't
refresh pivot tables that were connected to Power BI data sets,
right? We went and fixed that. So now if you want to go and publish
your pivot table into Power BI, you can. And so that's another way
that those users might first start with just Excel content that
they built, but they go into Power BI start, to see some of the
benefits that Power BI brings and start to onboard into other
things.
Brian Jones (01:12:57):
We're also looking at how to do a better job if you just, if you
happen to have those cases where Excel is your source of truth for
data, you have a big table of data in Excel. Getting that to be
something you push up into Power BI, they have so much cool
intelligence now, right? Where you go and push that data in Power
BI and they'll automatically generate dashboards for you and stuff
like that. Right? So I'd say view this as it's a long running
thing, it's going to be slow, and so you're not going to see like
everything solved right away, but you should notice every six
months that there's that next thing that we've done and just know
that that's not just us doing little small piecemeal things, there
is a longer north star that we're going towards. But it's going.
It's slow. Some of the stuff that we're doing under the hood is
actually... some of it's a little bit heavier lifting.
Rob Collie (01:13:39):
I'm actually incredibly relieved to hear that. I've been waiting 11
years, so a slow and steady approach is fine with me. It's the
possibility that nothing might be happening that has been really
killing me, but any-
Brian Jones (01:13:54):
We are definitely working on it.
Rob Collie (01:13:56):
And look at the success that you've had on long running
initiatives. We just talked about them. If this gets a similar type
of mindset, the ghost of Rob Collie can finally rest easy.
Brian Jones (01:14:07):
You can run... I mean, he's still going to give us a hard time
about it. It's one that I... I should just talk to you every once
in a while about them. Right? Because it'll be... Every time we
announced one of these, I should just kind of ping you just to see,
do you see it as a continued progress or did that seem like just
kind of a swing and a miss, right?
Rob Collie (01:14:22):
Well, I had no idea, for instance about the toaster thing that you
talked about. No idea. Boy, the pivot table refreshing. That's a
huge deal.
Brian Jones (01:14:30):
Yeah.
Rob Collie (01:14:31):
It doesn't do anything really to address growing the author
audience, but it doesn't matter, that one was just too big, and I
know that it was a hard problem, so.
Brian Jones (01:14:39):
It just clearly... I had another clear sign of, there were some
just gaps where we would... like the two things just didn't work
together properly. Right? And clearly we need the systems to work
together seamlessly.
Rob Collie (01:14:50):
All right. What's next for Excel? What's the overall roadmap other
than taking care of this Power BI authoring problem, which is
clearly top of the stack.
Brian Jones (01:15:00):
All of that stuff I talked about, the multiplayer mode of Excel,
super critical there. We're going to keep innovating, making that
better matter. For us, web is a tier one experience. It's got to be
just as good as Win32 and Mac. So we're going to continue to
innovate really heavily there and invest really heavily there. The
whole area around intelligence, just trying to find ways of making
the application easier to use, easier to learn, more approachable.
If you look at, even in schools, people are trying to teach data
literacy. Excel should be a key tool for that. And so it's really
important that we go and make Excel a thing that teachers feel like
they can use to go and teach kids how to work with data. And so
there's a big opportunity there, but there's a lot of work we need
to do there to just make the application feel more approachable for
that audience.
Rob Collie (01:15:46):
I completely agree. So many times if I had more time and energy,
thought about it so much, like going and volunteering when my kids
were still little. Can you imagine showing conditional formatting
to a room full of like eight year olds?
Brian Jones (01:15:59):
Absolutely.
Rob Collie (01:16:00):
They would just jump out of their seats at the insights that leap
off the page. Like you'd show them a sea of numbers. Who did best?
No, I don't know. No. Okay. Look, voom! Conditional format. And
everyone just knows.
Brian Jones (01:16:15):
Yeah, you should go and look, we put a couple of templates. We did
one back in November. It was a thing we did with NASA where it was
to get kids excited about looking at space data or data on rockets
and things like that, so we built a bunch of templates. We've got
one that we just launched about orcas. It was World Orca Day. And
so we've got these templates that let you go and analyze the
different pods of orcas that exist. It's using the Wolfram data
types, so you can actually have cells that are each orca and you
can click on them to see all the data about them and start to do
analysis. So there's a lot of fun stuff we're doing there. But
probably the really big one, back to your point, big bet. The other
big bet is you were calling it citizen developer, but really
there's people that like to build and create and make stuff,
right?
Brian Jones (01:16:59):
You've got people that are heavy data and want to do an analysis,
but there's another side of Excel, which is really more just like
solution building or modeling or things like that. Right? Like back
to the whole, I talked about the beginning, the soul of the product
is the grid and calc.
Rob Collie (01:17:14):
Yep.
Brian Jones (01:17:15):
And there's a set of investments that, this is a long running
project that we have rolled out in pieces, and I don't talk about
them as being one combined thing, but really they are one combined
vision, which is really upleveling Excel as a programming language.
We're working really closely with researchers in Cambridge, UK who
specialize in functional programming. They're experts in functional
programming. We've been working with them for a while.
Rob Collie (01:17:40):
Is this where we get to mention, namedrop, Simon Peyton Jones?
Brian Jones (01:17:44):
Yes, it is. Simon Peyton Jones, who is just amazing. He's been
heavily involved in helping us develop that. And again, Excel,
worlds must be the language, people don't usually think that. You
think of things like JavaScript, but those are like in the maybe 10
million developers. Excel is literally, there are hundreds of
millions of people who know how to write formulas, right? It's by
far the world's most popular programming language.
Brian Jones (01:18:08):
And there are certain things though that the product was always
lacking. Some of the stuff we announced recently, things like
Lambda and let, those were just basic primitives you'd expect to
have in a programming language, right? The ability to have code
reuse. So if you write a formula that's complex, do you have to
rewrite it every time? And then if you realize you made an error,
do you have to then go back and fix it everywhere? Or you use
Lambda where you can just define it once, give it a name, use that
name everywhere, and if you need to fix it you go to one place and
you go and fix it. Lambda also added a whole bunch of really
powerful stuff around like you can do recursion, it made Excel
Turing complete. In theory, Excel was Turing complete because of
VBA. But with Lambda, it means just the grid and calc and formulas
is Turing complete, take all the VBA and all that stuff out of
it.
Rob Collie (01:18:52):
The thing I really like about that feature is the name, because you
hear it and you know immediately what it does for you. I mean, it
just jumps off the page, right? Like I say, Lambda, and everyone's
like, yeah, totally, I know what I'd do with that. I couldn't
resist.
Brian Jones (01:19:07):
Yeah.
Rob Collie (01:19:08):
It's like, it's not reusable formulas. It's not.
Brian Jones (01:19:11):
No. There's a reason. If there's anything that you should take
away, the things that we do or are doing are really long running
bets. Lambda is... it's a geeky, underlying enabling tech that a
lot of end user-facing experiences will start to show up on top of
over time. And yeah, if you even look at like our blog posts, we
were targeting a specific audience with Lambda that would actually
get it, right? Over time we will add more. The other big one there
is data types. I don't know how much you've looked at data types,
but this will actually resonate with you from back in your XML
days. If you think about all the innovation that's happened in
Excel, tons of innovation around things like pivot table,
conditional formatting you mentioned, all that stuff, formulas then
stuff like XLOOKUP and crazy fast calc, there's been so much
innovation, yet what you can type into Excel is no more interesting
than like a notepad. Right? It is the most basic thing. You could
write a string, a number or a formula. What the heck?
Brian Jones (01:20:10):
It's been around for 30 years and we haven't innovated in what can
go into a cell. That's what data types is. We talk about it with
just things like stocks, but underneath we have plumbed Excel so
that a cell can be a deep, rich value, completely jagged data. If
you look at like the Wolfram data types, it's hierarchical data
that's all stored in that one cell, but it still works with calc.
You can now have a cell value can be an image. You can compare two
images with each other. You can say, does this equal that? You get
true/false. It's kind of mind blowing what data types actually
does.
Rob Collie (01:20:46):
I can do a VLOOKUP or an XLOOKUP or whatever against a table of
images now?
Brian Jones (01:20:50):
Yes.
Rob Collie (01:20:50):
And I won't need like the camera tool.
Brian Jones (01:20:53):
Right. Now, the only way right now to get images is if you have a
data types that have images and then you can pull that out through
a reference. But of course eventually we'll just have where you can
just go and insert your own image, right? But one of the
partnerships that we did with Power BI is you can have Power BI
data come into Excel as a data type. So I could have a set of data
defined up in Power BI, bring that in as a data type, I can look at
that data, I can hover over it, I see a card, it tells me where
that came from. So now if I'm like, oh shoot, is that really what
the product price is? Did somebody just copy paste this? Where did
it come from? I know where it came from my hover over it, I see a
card. It came from that Power BI source, right?
Brian Jones (01:21:31):
So it means Power BI can be your master data manager. Right? You
can say, okay, I've got all my master data here, I bring it into
Excel, and Excel I actually see where that data came from, I can
trust it, and now I can go and do a whole bunch of calc around it.
So the data type stuff, it's just, again, it's kind of scratching
the surface a little bit in terms of the way that we go and talk
about it, but it's some deep, deep plumbing.
Brian Jones (01:21:53):
And then the last piece, which is also about just new types of
values is arrays. Right now I can go and have it where I can have a
formula that returns an array of values, right? So I could have
stock dot history and that one formula will return a huge set of
data that goes into the grid, and then I can write another formula
that references that result and I can have that formula apply to
the entire thing in the array, right? So I can have an array of
data that I can pass through the calc chain and do manipulations on
that array of data. And so I can do that step by step by step
process, manipulating that full array, where in the past, that was
the beauty of Excel, is I do a step by step by step, but it was
always with that single value, like I get a number and I'm going to
go do a few operations on it to get the final result.
Brian Jones (01:22:41):
Now I can have a data type like a person and do multiple operations
on it using the programming language, or I can have an array of
data that I can do multiple steps on it. So it's the beauty of the
grid and calc, but now I'm working against the world's data, I'm
working against all sorts of objects, not just numbers and
strings.
Rob Collie (01:22:59):
So I'm going to give you all the disclaimers. It's one person's
opinion and it's an unsolicited opinion. I know that.
Brian Jones (01:23:05):
Yeah.
Rob Collie (01:23:06):
Right. But what better time to give you that opinion than when
we're recording for the world to hear? So I think that of all the
features you just mentioned that are pushing the envelope, like the
state-of-the-art of Excel forward, I have this instinct that the
one that's going to become most common, sort of like most
ubiquitous, is Lambda. And so I want to project that sooner or
later you're going to rename it, because I think it's probably more
applicable to a wider audience. You know the book, or the concept,
crossing the chasm? I think for early adopters, Lambda is the right
name. It's not the right name for early majority. And I think there
is one for Lambda. It's just so fundamental, right?
Rob Collie (01:23:52):
This is a problem. The lack of formula reuse is a problem that
basically plagues every single formula writer ever. I can imagine
living my whole life without, in a particular niche, without ever
needing to manipulate arrays as if they were a single value. Now,
there's plenty of cases where I can imagine that I would, right?
It's not that I'm saying that it's a bad feature. I think it's an
awesome feature, but man, Lambda, I just see Lambda as just like
incredibly mainstream. I could be completely wrong about that, of
course. But I think I'd have a better chance teaching Lambda to a
broader audience than I would arrays for instance, even the new
dynamic ones. So anyway, end of rant. It's not even really a rant,
it's-
Brian Jones (01:24:37):
That's no rant.
Rob Collie (01:24:37):
As rants go, it's pretty polite.
Thomas Larock (01:24:39):
That wasn't a rant.
Brian Jones (01:24:39):
That's just an opinion. To me, I'm like, hey, tyranny either
or.
Rob Collie (01:24:43):
I know. Friends like, I hear your business book concept and I have
raised you my business book concept.
Brian Jones (01:24:50):
All the time, man.
Rob Collie (01:24:52):
You've read the same best 100 business book's cliff notes that I
have.
Brian Jones (01:24:57):
I haven't thought about this from a crossing the chasm, early
adopter thing. I think of it more like in terms of user persona,
it's more closer to being truly like would identify as a developer.
And a lot of the stuff that we've done here so far is at that
layer. The place that Lambda's exposed as like a name is in the
formula itself you build. But clearly as we go and build a UI on
top of it to make it easier to go and create those, we wouldn't go
and say like have a big button that says, generate my own Lambda.
That's not really the way that it'd be implemented. But then the
data types in the array stuff, I think you'll be surprised over
time. I get that you're not... like it wouldn't be a thing that
you'd say, "Hey, I'm going to go and teach somebody arrays," but I
think that what you'll find is it'll just almost naturally start to
become how you work.
Brian Jones (01:25:42):
It's just like data types is almost one of those things where you
say, I don't have a class right now where I say I'm going to teach
you about cells, let's talk about cells and what goes into a cell.
And so this is really just saying we're just upleveling what can go
into a cell. And so eventually nobody really even thinks about it.
It just will seem natural. It's like, well, of course you could put
that stuff in there.
Rob Collie (01:25:59):
Well, so it's not about... It might've come across this way. It's
not about a lack of belief in these other things. It's more like, I
just think the bar size, the pie is bigger. And the example that I
can't get out of my head, it's something that I wasn't exposed to
until I come outside of Microsoft, all these income statement
forecast spreadsheets, giant spreadsheets that cover acres of
screen real estate of this like projecting forward month, month,
month, month, month with different variables being fed in and
different growth rates and different attrition rates and all of
that, and all of that intermediate calculation, all of that real
estate is just to produce a very small handful of outputs. So it's
like I start with a small number of assumptions and some rules, and
then I expand that, blow it out across this whole spreadsheet that
thankfully can be more than 256 columns wide now, right? And then
all I do is harvest those three values at the end. That's all I
really care about. Or what's my break even month, how long before I
break even? There's just a handful of outputs.
Rob Collie (01:26:59):
And I end up with this gargantuan device and I can't iterate on it.
It's like I changed the inputs and then I want to save off the
outputs. I just want to turn that whole thing into... and then just
turn it into a table on another sheet, hide that machinery behind
the scenes. And I just think that's just so incredibly mainstream,
like everyone has this problem. And so that's why I just can't get
this one off my brain. So it's really an obsession with the Lambda
feature as opposed to a lack of interest in the others. Because
I've already used the others, too.
Brian Jones (01:27:29):
Yeah. The idea around reusable logic and even going further, right,
like Lambda's that I could go and create and then call from outside
of Excel, a Lambda as a web service you could publish, right?
There's a long way that we'll be able to go with this. And so I
think that it's a huge, huge opportunity for us. Right now what
we've released is just kind of the beginning underpinnings of a lot
of innovation that will be coming.
Rob Collie (01:27:58):
I have to say as a long-time observer of all this, former insider
on exactly that product, I think it's really truly exciting. It's
real innovation. It's not keeping up with the Joneses. It's not
like making sure you have parody with your competitors. And it's
also not the old disease of, wouldn't it be cool if? It's not the
old stuff like the things that we were doing with XML, right? It's
grounded in true, actual value for the customer base, for the
people of the world, and is innovative. That's hard to do. It's
doable, it turns out, but there's a lot of discipline to it and I
salute the direction that y'all been headed in. It really is
exciting.
Brian Jones (01:28:38):
Yeah. Thanks. The thing that's been fun about that project is the
combination of, you know them, there's a bunch of people on our
team who their background is like financial modeling. Right? And so
in addition to us researching the customer, we hired the customer.
Right? Which, is great. And then we've combined that with the folks
in, like Simon Peyton Jones in MSR and Andy Gordon, this
combination of computer science experts. Right? And then people who
deeply understand the customer problem. And like you said, it's
really easy to go and see, what are those challenges? Well, one is
when I'm looking at my model, can I trust the data, where'd the
data come from? Right? And so this partnership with Power BI around
data types helps a ton with that. Can I go and write logic that's
reusable, so it's not fragile? Lambda's going to be huge for
that.
Brian Jones (01:29:26):
And then even our collab stuff, we announced that in our web app
now, and it'll come to the desktop, I can now go and write... if I
see a value and I don't know where the hell that came from or why
did it change, I can right click and say, show changes, and it will
actually show you the edit history of that.
Rob Collie (01:29:39):
Wow.
Brian Jones (01:29:40):
So I can see, oh, Rob edited this two days ago, I can now go and
contact Rob and say, "Hey man, what's going on? Why did you change
my formula to a hard-coded value? That's not good practice,
man."
Rob Collie (01:29:49):
For yucks.
Brian Jones (01:29:50):
Yes.
Rob Collie (01:29:51):
I just wanted to see if-
Brian Jones (01:29:52):
Just for fun.
Rob Collie (01:29:52):
... you're paying attention. Yeah.
Brian Jones (01:29:54):
I wanted to get the output I was hoping for.
Rob Collie (01:29:56):
Yep. That's what we do. We cook the books. Well, I can't keep up,
both because you're doing so many things... It used to be a lot
easier to keep up. You're doing too many things for me to keep up,
but also it's weird, Excel has now reached the point where it's not
the app that I spend the most time in on my computer. I'm in Power
BI, I'm in Outlook, I'm in PowerPoint a lot. I'm a mascot. The
people who work at our company are so much better at the tool. So
I'm not even really in Power BI that much. Brian, thank you so
much.
Brian Jones (01:30:29):
Yeah. This was fun, man. It was great catching up.
Rob Collie (01:30:31):
I think people are going to enjoy this, and if they don't, we won't
know.
Brian Jones (01:30:34):
All right.
Thomas Larock (01:30:35):
Yeah, right.
Brian Jones (01:30:38):
It's like how we used to ship software, right?
Rob Collie (01:30:39):
Yeah. That's right. Actually-
Brian Jones (01:30:40):
Throw it out there and move on.
Rob Collie (01:30:42):
Actually, I hate to tell you, we have detailed files. We have an
amazing Power BI report that uses robotic process automation,
courtesy of Ash, to go export the data from the unfriendly portal.
And then we power query it. We'll be able to tell you next week how
you're... how you're trending.
Brian Jones (01:30:59):
Oh great.
Rob Collie (01:31:00):
Are you trending ahead of Ken Paul's or are you behind Ken Paul's.
If you want to know where you rank in the world we'll tell you.
Brian Jones (01:31:07):
Please don't tell me where I'm ranking. Ignorance will be
bliss.
Rob Collie (01:31:12):
Yeah. Well, just seriously, it was a real pleasure. How often do
you get to talk to someone who's in your role and how often do you
get to listen to someone who's in your role? I don't think very
often. So I certainly appreciate it. I think our listeners will as
well.
Brian Jones (01:31:25):
Well, I had fun, man. It was great catching up.
Announcer (01:31:27):
Thanks for listening to the Raw Data by P3 Adaptive podcast. Let
the experts at P3 Adaptive help your business. Just go to
p3adaptive.com. Have a data day.