Jun 1, 2021
- 4:05 - Chris's gateway drug into Microsoft BI was MDX, Training challenges and saving souls, and the predecessor to Analysis Services
- 17:00 - Chris's OLAP experience, DAX vs MDX, Chris and Rob reflect on a decade old nerd fight
- 34:40 - Dave Grohl and Ratatouille, why we need certain functions and OLAP, and the power of PR,
- 50:00 - Tableau to Power BI migration, self-service BI, PBI's feedback system, and Excel adopters
Episode Transcript:
Rob Collie (00:00:00):
Welcome, friends. Today's guest is Chris Webb. There really aren't
many people who can claim as many years of experience with the
evolving Microsoft data platform as Chris can. I myself have been
around this particular campfire for almost 20 years and Chris
predates me by quite a bit. He was one of the OG Microsoft BI
professionals and celebrities, so he has seen a lot. He's
practically a historian of all of this stuff while at the same time
being very much highly active in this space today.
Rob Collie (00:00:36):
Chris and I first met each other professionally sort of like in the
formative years of 2010, that era, when Power Pivot and the tabular
Analysis Services model was first coming to market. And from the
get-go in the early days, we were an opposition to each other. We
had some pretty fierce email arguments like big tough guys. But
that was all a long time ago and things have changed so much, and
really in hindsight, we were both right. We were both keying in on
opposite sides of something really important and we're in a very,
very friendly place with one another these days.
Rob Collie (00:01:12):
And it's really nice to get together and talk with an old adversary
who now very much is completely an ally. It was long overdue and
he's just an amazing conversationalist. We had a really good time.
I hope you enjoy it as well, so let's get into it.
Announcer (00:01:29):
Ladies and gentlemen, may I have your attention, please?
Announcer (00:01:34):
This is the Raw Data By P3 Adaptive Podcast with your host Rob
Collie and your cohost Thomas LaRrock. 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 (00:01:58):
Welcome to the show, Chris Webb.
Chris Webb (00:01:59):
Hi, there.
Rob Collie (00:02:00):
How are you today?
Chris Webb (00:02:01):
It's five o'clock on a Friday afternoon in the UK, so this is the
last thing that I'll be doing before the weekend. But I can't think
of a better way to finish the week.
Rob Collie (00:02:10):
Well, I couldn't think of a better way to finish the week than like
being in a British pub about this time. There's just something
about the way y'all do beer. I think you figured it out. I prefer
the English pub experience. I just feel like the combination of
taste and drinkability and temperature and all of that, I think
you've really got it nailed down over there.
Chris Webb (00:02:28):
So you're not going to get at me for warm beer or anything? Are
you?
Rob Collie (00:02:31):
No, I actually-
Chris Webb (00:02:32):
This is usually what people say about British beer.
Rob Collie (00:02:35):
No, I have really no complaints at all about the British pub
experience. As far as I'm concerned, for me, I find it optimal. I
wish I could get exactly that sort of experience over here. Instead
of you've got to drink the Triple APA, it's whatever. Anyway.
Chris Webb (00:02:49):
It's going to be a bit like that over here as well, to be honest,
so-
Rob Collie (00:02:51):
Is it really?
Chris Webb (00:02:52):
All the old pubs are closing down and there's craft breweries
popping up, but times change, times change. The kids, they want the
Triple APAs.
Rob Collie (00:03:00):
They want the Triple APAs. Well, not everything is progress.
Chris Webb (00:03:05):
Yes. Well, nevermind. Even if we don't have the beer, we can have
the kind of rambling half drunk conversation that doesn't really go
anywhere and doesn't really mean anything.
Rob Collie (00:03:13):
That's right. Maybe that's what I'm really confusing it all for.
Maybe that's the thing I like.
Rob Collie (00:03:18):
You're a number of things you're on the CAT team at Microsoft-
Chris Webb (00:03:21):
Mm-hmm (affirmative).
Rob Collie (00:03:22):
... and you're also Cross Join.
Chris Webb (00:03:25):
Well, I was, but I still am as far as my blog goes. I didn't sell
my domain or anything. I've still got that.
Rob Collie (00:03:31):
You've had a long time ringside seat for the evolution and
development of the Microsoft BI platform. People think of me as
being a long time insider, but I think your involvement in this
whole story predates mine. When did you first discover the
Microsoft BI stack?
Chris Webb (00:03:52):
Would it be been '98, 1998?
Rob Collie (00:03:57):
Way before me, maybe like four years, but that's a long time back
then.
Chris Webb (00:04:01):
I was an early adopter, I think. Definitely.
Rob Collie (00:04:04):
What was the gateway drug there? How did you stumble into it?
Chris Webb (00:04:07):
I was working at a company they put me on a project to evaluate
this brand new OLAP Server coming from Microsoft, and that was it.
I was the first and only person in the world to fall in love with
MDX. People are rude about it nowadays, people are openly rude
about it. Even Microsoft people are openly rude about it. But it
was and is a beautiful language and a beautiful product and I loved
it from the very beginning. Even though it's been a while since
I've written an MDX, I think it's one of those things that I'll
carry to my grave.
Chris Webb (00:04:41):
It's my dad's 80th birthday today, and when I am 80, if you were to
open up an antique copy of Windows 7 and pop open SQL Management
Studio, even after like 35 more years, I would be able to write an
MDX query, I'm sure.
Rob Collie (00:04:56):
That's funny, that's obviously exactly the opposite of my
experience. Every six months, I would forget that I hated MDX and
I'd ask someone at Microsoft to teach it to me. And I'd be like,
"Look, all I want to do is write an if. Just want to write an if."
And we'd be the 15, 30 minutes in to hierarchies and all that and I
go, "Oh, right. I forgot. Yeah, this sucks. I don't want to do
this."
Chris Webb (00:05:17):
Well, that was always the problem. When I was doing training, I did
a lot of training back in the day when I was self-employed. And I
remember I did this training course once in Germany and there was
somebody on the course, it was an MDX training course. And there
was somebody on the course and I recognized them, you recognize
people from different places. And I said to this guy, "You've been
on this course before." And he said, "Yes, this is the third time
I've been on your MDX training course."
Chris Webb (00:05:39):
And he said, "Well, the first time I did it, I thought it was a
really great course, but then I got back to the office and I didn't
use it. And then about a year later I had some training budgets, so
I thought I'd come on to your course again. So I came on the course
and it was great, but I didn't use it. And now I'm in a new company
and they've got some training budget and I forgotten MDX, so I
thought I'd come on your course again." So I think that's a vote of
confidence in my training, even if.
Rob Collie (00:06:04):
Well, MDX training, which you're just talking about, it was one of
the original subscription models. I have got a really good monthly
recurring revenue of returning students. There's like a six month
wavelength, they come back.
Chris Webb (00:06:17):
That's the ideal thing. Why look for new customers when you can
sell the same thing to your existing customers over and over
again?
Rob Collie (00:06:23):
In fairness, I had the same experience one time with a whole team
that I trained on Power Pivot back before Power BI existed. And it
changed the way I thought about training and it changed the way
that I messaged it going forward. This group of Excel analysts who
are completely overworked and behind, their manager brought me in
to train them, and I did. And then like a year later, they call me
back and I went back and trained the exact same people on the exact
same thing.
Rob Collie (00:06:50):
It was worth a post-mortem, like I sort of interviewed like, "What
happened?" And they got the training, and the next day after the
training, they went back to their jobs, which had been on pause
during the training, a couple of days of training. And so stuff had
just piled up, they told themselves, "We'll dig out first. We'll go
and do, we use the old tools, the regular Excel VLOOKUP stuff.
We'll do that, we'll dig out and then we'll do the new stuff."
Well, no, no, you're not. The freshest day is that first day.
Rob Collie (00:07:18):
If you don't pick to start using it that day, it's not like the
following day is going to be easier. It's going to be incrementally
harder. And so I started telling people that to a certain extent,
people listen, some of them don't, but I would tell them sort of
like... Once I had them in three quarters of the way through the
first day, I would tell them, "Now, listen, you didn't know this
when you signed up, but you are not allowed to use VLOOKUP or
regular pivot tables ever again.
Rob Collie (00:07:45):
And I'm sorry if didn't understand that you were signing that right
away when you took this class, but at least for one real business
thing, tomorrow, the first day after training, whatever, you have
to be applying it from the get-go. If you don't, you're going to be
a repeat customer."
Chris Webb (00:08:04):
You're ruining your pipeline there. But yeah, it's interesting, the
whole self-service BI thing is meant to be transformational. It's
meant to be change your life, change the way you work, but it's
like all these things, it's like a lot of people might hear the
message, but they won't necessarily change their ways. I think the
only people who use tools like Power BI, the best are the ones who
find it themselves rather than the ones who were taught it.
Rob Collie (00:08:28):
Oh, yeah. We have this concept in training at our company. We have
a distinction between volunteers and hostages. If you volunteered
essentially to be in the training class, that's a different thing
than your manager saying, "We're going to do training for the next
three days and you have to be there."
Chris Webb (00:08:46):
You could always tell the people who will have to be there in a
training course.
Rob Collie (00:08:49):
That's right.
Chris Webb (00:08:50):
They might be initially friendly for the first hour and then just
go off and read their emails. You can try your best to entertain
them and get their attention and show them something cool. What I
think of something cool is not necessarily what somebody is forced
to go on a training course is something cool, but they will just...
It's incredibly difficult to engage somebody who's like that.
Rob Collie (00:09:10):
Well, that's another change that I made, and then ultimately we
made as a company in terms of how we train, is that if we know that
there's a proportion of hostages in the class and there always are.
Even if it's like a public training that someone voluntarily signed
up for, you know someone got dragged along. There's a volunteer and
a hostage. From the beginning we say, "Look, we get it, training
sucks. Sitting in a classroom sucks. We don't want to be sitting in
your chair either. But here's the thing you actually should pay
attention to this one. This one is going to change your life."
Rob Collie (00:09:43):
We really throw down the gauntlet. We sell out to get their
attention and say, "Look, if this is the one training you pay
attention to in your life, it will be the right one." Even that
crew that I talked about, that I trained twice, a really funny
story, like you talk about, you can tell when you're training. You
can tell the person who's just not there, not paying any attention.
This one woman in that first class had just been reassigned to the
analyst team. She'd had a different job, they just sort of like
parked her on the analyst and they didn't know what to do with
her.
Rob Collie (00:10:13):
And she did the equivalent of checking her social media, you could
just tell. You could almost see her pressing the Facebook Like
button or whatever. She was in the back of the room, she had no
interest at all in being there. But a year later when I went back,
that same person, she had then been exposed to a year in the
trenches of being the VLOOKUP and Pivot Analyst and now she was
totally motivated. She was in the front row, she was asking the
best questions.
Rob Collie (00:10:43):
Six months later, she left that company and took a full-time BI
position at another company. Everything was different for her, and
it was just like it's motivation. She hadn't glimpsed how awful it
could be.
Chris Webb (00:10:55):
I think you saved another soul there.
Rob Collie (00:10:58):
Yeah, I did.
Chris Webb (00:10:59):
Congratulations.
Rob Collie (00:11:00):
I did. I suppose there's like a count of souls saved and souls
corrupted, I don't know. I hope to be above the x-axis-
Chris Webb (00:11:07):
Oh yeah.
Rob Collie (00:11:08):
... in the final analysis.
Chris Webb (00:11:10):
And especially one more soul saved for the Microsoft party rather
than potentially, if you hadn't been there, they might've gone to
Tableau or something, perish the thought
Rob Collie (00:11:19):
I know. Have you seen the Bootleg Hitler video about Tableau?
Chris Webb (00:11:25):
Oh, I think I have, but I've forgotten it. Go on, tell me
again.
Rob Collie (00:11:28):
You've seen the Hitler memes, right? Where he's looking at the
maps-
Chris Webb (00:11:30):
Yeah, yeah, yeah.
Rob Collie (00:11:31):
... right?
Chris Webb (00:11:31):
Yes, from downfall.
Rob Collie (00:11:32):
Yeah, from downfall. Exactly. They're pointing at the map and
talking about retail sales are up here, here and here. And then he
goes, "That's nothing. Tell me about seasonally corrected profit
margin." They're like, "Ah." Anyway, I made that, but didn't attach
it to brand because Hitler, but yeah, I agree. It could have gone
different way. I actually think though that it was one of those
situations where she ended up taking like a business objects job.
So she was out of the frying pan of Excel.
Chris Webb (00:12:07):
Well, she'll come back to Power BI, it'll just give her an even
greater appreciation.
Rob Collie (00:12:11):
That's right. She's got to go experience all of the hard ways.
Thomas LaRock (00:12:14):
I wanted to ask something, because Chris, your experience predates
mine. You said 1998. How'd you say it, there-
Chris Webb (00:12:22):
OLAP Services. So it would have been in the beta period for SQL
Server 7. So it would have been '98, '99. I think it was late
'98.
Thomas LaRock (00:12:29):
Okay. I was expecting, you were going to say Analysis Services.
Chris Webb (00:12:33):
No, no. It wasn't called Analysis Services until... I think Rob
will particularly appreciate this. I think the first thing that I
got annoyed about publicly about something to do with Microsoft BI
was somebody renaming OLAP Services, which was a good descriptive
name to Analysis Services. And they did it, they did it because
they added data mining functionality. And I thought, "Nobody's
going to use this. Why don't we stick to the..." It was typical of
me, completely typical of me.
Chris Webb (00:13:00):
I didn't have a platform to be rude about it back then, so I got
away with it. It was the first instance of my cheese being moved,
so I didn't like that.
Thomas LaRock (00:13:08):
So OLAP was with SQL 7, and then it got renamed to Analysis
Services when they folded in data mining. Okay. I came in really
after SQL 2000 roughly. So Analysis Services have... it's always
just been there. When I heard MDX, I'm like, "Is that way you
really mean, Analysis Services?"
Chris Webb (00:13:27):
Yeah. People do tend to use... In fact, I saw one of my colleagues
very politely smack down somebody recently saying... When they used
MDX to describe Analysis Services multidimensional and it does kind
of irritate me. I'm very good on using the right terms. But yes MDX
is the language, Analysis Services multidimensional is the server
products.
Thomas LaRock (00:13:49):
Yeah, for me, that was just something that we installed very
alongside with SQL Server. Back in the day when you just installed
everything, because it was just easier. No need to go back.
Rob Collie (00:13:58):
Installation ironically is my first place where I crossed paths
with Analysis Services. In '98 when you were using, learning MDX
and falling in love-
Chris Webb (00:14:09):
For the first time.
Rob Collie (00:14:11):
... a lifelong romance, I was the program manager on the first
version of Windows Installer, the MSI technology. And I don't
remember why, but there was something that we had to install, I
think in office, it was this Plato thing. I remember it being a
real pain in the ass. There was something about it that was
controversial in terms of its install, and so I actually had to get
involved and figure out how to get all of their self registration
ripped out or something like that. It was probably something
along.
Rob Collie (00:14:41):
It was always self-registration, that was the evil of the era. Do
you have any idea, Chris why Office 2000 would have needed to
install anything related to Plato?
Chris Webb (00:14:49):
Oh, Plato is the code name, Aristotle was the client that didn't
get bored and it would have been because even back then, Excel 2000
could connect to OLAP Services. It was PivotTable still being able
to connect to Analysis Services. In fact, earlier this week, I've
got a new user group presentation, I take around the user groups
showing, connecting PivotTables to Power BI, nothing changes.
Skills I learned 20 years ago still being recycled. Now, we can do
PivotTables in the browser in Excel online back to Power BI, but
it's the same old stuff.
Chris Webb (00:15:23):
It's PivotTables connected back to some kind of OLAP technology,
and people still get wowed by it. Which is great.
Rob Collie (00:15:30):
Keeping that MDX alive, right?
Chris Webb (00:15:31):
Oh yeah. Still MDX in the background.
Rob Collie (00:15:35):
The majority of MDX usage today is overwhelmingly Excel, querying,
whatever-
Chris Webb (00:15:40):
Yes.
Rob Collie (00:15:40):
Even a Tableau model, right?
Chris Webb (00:15:42):
Mm-hmm (affirmative).
Rob Collie (00:15:43):
But I didn't know that in 2000, Excel already had lit up with that
first connectivity. I just knew that I had to install
something.
Chris Webb (00:15:50):
Yeah, Excel 2000 was the first version, I think that... I have a
feeling that maybe even Excel 97 could have done it, but no, maybe
it was Excel 2000, but it's a long time ago.
Rob Collie (00:16:01):
It had to be. I'm struggling even to imagine that 2000 did, but of
course, I was installing it, there had to be some Excel. Even as
late as 1999, at the end of every Office release for a long time,
those big long two and a half, three year monolithic releases,
there would be sort of like this almost like open job fair at the
end where like people could like shuffle internally from team to
team. And I remember talking to a manager one time, the words came
out of my mouth. This is going to make you laugh, I think.
Rob Collie (00:16:31):
I said, "So for instance, I would never go to work on something
like Excel, for instance." I was very derisive about it, I would
never go to work on something like Excel. And he was like, "What is
wrong with you, youngster? You idiot. You need to go work on
something like Excel." And I'm like, "I know better than you." And
a couple of years later after a failed product and a reorg, I've
found myself in Excel and I was like, "Oh, this is great."
Chris Webb (00:16:56):
I do feel like Excel has got its mojo back. They're back doing cool
stuff nowadays. Apart from the cool stuff they always did with
people like you, but they've got like lambda functions and things
like that. There's interesting, exciting things happening in
Excel.
Rob Collie (00:17:11):
Yeah. It's an edgy, sort of cutting edge tool again.
Chris Webb (00:17:15):
Yeah. That nice dark green color scheme again, after the fairly
washed out years. It's got some back.
Rob Collie (00:17:20):
Yeah. The past early years, we try not to think too much about the
washed out pallet. So that's crazy that Excel was already doing
that, but I have this Plato thing seared into my brain for some
reason. I had no idea that it was like a mirror. Didn't get to meet
him there. So what were you doing at that company before they said
to you, "You need to go kick the tires on this OLAP Services
product"?
Chris Webb (00:17:44):
No much because it was like my first IT job and I'd spent maybe six
months or something learning to be a VB6 programmer-
Rob Collie (00:17:53):
Oh, oh.
Chris Webb (00:17:53):
... and doing a bit of testing and things like that. And it was
literally... As a company, it was good because they've got a long
history of using every single OLAP tool ever. Had used
MicroStrategy, they'd built their own kind of OLAP type engine, so
they knew what they wanted. They knew what they wanted this stuff
to do. And they had a whole bunch of really quite complex
requirements and calculations that needed to be done, not the basic
stuff, good complex things.
Chris Webb (00:18:21):
And it was the end of a long period of them evaluating a ton of
other products they'd rejected. And then Nigel Pendse, if you
remember Nigel Pendse. That's also a name that will separate the
old people from the kids. But I think Nigel Pendse retired about 15
years ago from being the OLAP guru. He recommended using this new
Microsoft technology that was in beta and it was the end of this
long project. And I guess it was just like, oh, we'll give it to
the graduate trainee to go away and evaluate this stuff, but
actually I really enjoyed it.
Chris Webb (00:18:54):
The things I enjoyed about learning MDX are the things I still
enjoy about DAX and M. It's like taking some kind of calculation
logic and expressing it as a formula that works and always works.
And no matter how you put your rows and columns together, it gives
you the right answer, and I enjoyed that. That's when I really got
into it. And I suppose there was also a factor of feeling like
you're breaking new ground or learning something, because this is a
technology that nobody else knew anything about.
Chris Webb (00:19:24):
And then I realized that there were newsgroups out there where
people would ask questions, and I realized that actually I knew the
answer to some of these questions. So I got really into answering
questions on newsgroups. I suppose, that gives you a bit of feeling
that you know something and you're getting somewhere and you're
helping people. And I suppose that's how you fall in love with
things like this. You feel like it's not just interesting, but it's
something that's good professionally, good career-wise you're
meeting people, you're helping people.
Chris Webb (00:19:55):
As soon as you get active on newsgroups, this is how I first met
Amir and Moshe and all of these people because they were out there
answering questions as well. And yeah, I felt like I was being part
of something bigger, part of the community, and that really spurred
me on.
Rob Collie (00:20:08):
There's something you said in there that actually made me go, "Ooh,
maybe I just missed my window." Here's a confession. This isn't
something that is a strength of mine. I'm going to just be clear
about that. As I was thinking about it, that part of the reason why
I didn't want to learn MDX was because, first, it did seem overly
hard for something simple. That is an objective or more objective
anyway, sort of objection that I had to it. But I think the other
problem was is that on top of that, there were already a lot of
people who knew it.
Rob Collie (00:20:39):
So I was judging myself against them from the get go, and as a
result, like it was easier for me to sort of deny it and distance
myself from it, than go through the hard work of sort of climbing
that hill. Whereas with DAX, I had exactly the opposite. There was
no prior expertise really in the world. It was, like you said, like
this frontier, you were the first like Lewis and Clark to discover
Yosemite or whatever. I think that weakness of mine didn't get
activated in the DAX world. It also, I think DAX has a smoother
on-ramp curve.
Rob Collie (00:21:20):
I talk about MDX as having a learning cliff that you first have to
scale under machine gunfire and everything before you can get your
Hello World. Whereas you can Hello World pretty easily in DAX, but
maybe if MDX were brand new, maybe I would have been a bit more
into it. I'll never know obviously, but.
Chris Webb (00:21:42):
I know what you mean. I will quite freely admit that it is a bit of
an ego trip when you learn something and you kind of first there
with a blog posts and answering the questions. And I don't know if
that makes me a bad person, but I enjoyed that. And I was lucky in
that I was there at the right time that somebody... There were
other good people around. I distinctly remember they were a whole
bunch of other people around on the forums who then
disappeared.
Chris Webb (00:22:08):
I guess they went on other projects, they didn't get to use this
stuff. There was a guy called George Spofford, he was the original
guru with Analysis Services and MDX. He was such a lovely guy. He
sort of lost a couple of years ago. He was there, he wrote the book
on MDX and brought me in to write a chapter on the second edition,
which was really good. But yeah, he was older than me, probably
younger than I am now. But he was also incredibly clever, he came
up with all kinds of amazing stuff that I had no idea about.
Chris Webb (00:22:42):
Just as a little sidetrack on this, he told me this story once even
before OLAP Services. He'd worked on a BI startup and they were
burning through money, but they had gone through most of the
process for getting a patent on the idea of a measure. You know how
software patents are crazy-
Rob Collie (00:23:00):
Yes.
Chris Webb (00:23:01):
... you get a patent on the most obvious, stupid thing. His company
had almost got this patent on the idea of a measure of taking a
column of data and specifying how it aggregates up. He said, they
had just got to the end of their startup and it would have been
another $10,000 to finish getting this patent or they could spend
that $10,000 on trying to get their company going. And he said,
they spent the money on the company and let the patent process
laps. And he said, if they'd got that patent, they would have been
stupidly rich for the rest of their lives because they could have
just collected royalties from every single BI company; from
Microsoft, for everybody forever and ever.
Chris Webb (00:23:42):
Can you think about how fabulously wealthy you would be if you've
got such a fundamental thing like that? He was way ahead with MDX
and he was always the person that I was looking up to and catching
up to. And then he went and took a job with Hyperion. I never quite
understood why, but... And he disappeared from the scene and it was
like, "Wow, it's just me. It's me." Stacia or Varga as is now and
things around, but it was like, "Wow, the field is clear." Not that
I really knew how to exploit it.
Chris Webb (00:24:11):
Again, I think talking about Stacia, I think she had a much better
idea of how to exploit and monetize her position than I ever did
because I was just a green, 20 something kid. It's just being in
the right place at the right time. And I can look back and I can
see other opportunities that maybe I should have taken that I
didn't. DAX is a great example. I was there right at the beginning
with you, but I didn't really invest enough time with it.
Chris Webb (00:24:35):
And if you look at what Marco Russo and Alberto Ferrari have done,
they were there at the same time, but they took that and they ran
with it. They were clever enough to know how to monetize it. And
yeah, look at the great situation they're in now.
Rob Collie (00:24:48):
I made a great and terrible error with Marco and Alberto one time.
They were over here at a conference. Tom, this is your fault.
Thomas LaRock (00:24:56):
Okay.
Rob Collie (00:24:56):
It was one of your conferences.
Thomas LaRock (00:24:57):
All right.
Rob Collie (00:24:57):
It was one of those past bacon conferences that you organized. If
it weren't for that, I wouldn't have gone to dinner with Marco and
Alberto. At the time I was still running relatively unopposed as a
blogger in the DAX, Power Pivot space, and I let slip. Just not
really let slip, I just was friendly, collegial conversation. I let
slip my sort of average daily page views was on the blog at the
time. I don't remember what it was, and I could see Marco's head
just snap back and look at me and goes, "What?"
Rob Collie (00:25:37):
And over the next six months, suddenly the thing that we know of as
the SQL BI web empire, it kicked into gear. It was like... So
funny.
Chris Webb (00:25:50):
It's an awesome empire. Complete respect for the great job they do
with the content, with the way it's presented, with the general
look and feel as well. It's the thing of beauty and they deserve
all the success they get, I think.
Rob Collie (00:26:04):
It reminds me of my friend that used to kick my ass at video games,
sometimes like real-time strategy games. You go and watch him play
in his office when he's killing you over the network, after he's
already annihilated you, Age of Empires or whatever, and he's
finishing off everybody else, you just walk over and watch what
he's doing. He was just always kick back on... His feet up on the
desk, he's got one hand on the mouse, he's not even bothering with
the keyboard. And it's just like effortless.
Rob Collie (00:26:27):
They seem to hustle without ever feeling... You never interact with
them and have this feeling like they're frazzled.
Chris Webb (00:26:34):
Yeah.
Rob Collie (00:26:35):
It's really amazing what their output is able to look like without
ever being like the disheveled mess that you would expect them to
be while doing it.
Rob Collie (00:26:46):
You and I then have something in common, which is that we were for
a time sort of like the running unopposed in the community, far and
away most prolific for a particular Microsoft BI technology. And
then for various reasons, relinquished that role sort of at an
inflection point in the story. I was the DAX upstart that was crazy
running with it, while you're going, "Oh, come on, MDX is awesome.
Why are we doing this?" But then I also stayed for a while anyway,
as a blogger in the Excel world, even as Power BI desktop was
coming along.
Rob Collie (00:27:24):
It doesn't really matter in the end because what really happened
with me is that we turned into a company and being a blogger wasn't
anywhere close to the most important thing for me to be doing. I
mean, still isn't, but we were talking about this. Like the ego
trip of being the sort of like one and only for a while. And then
you in a way, you almost like get your way and the world wakes up
to the value of the thing that you've been so keen on. And now,
"Oh, now I'm not alone anymore."
Chris Webb (00:27:50):
Now, my biggest claim to fame is that I work on the same team as
Guy in a Cube.
Rob Collie (00:27:54):
Right.
Chris Webb (00:27:55):
The interesting thing is my blog stats have never been higher, but
in relative terms, I'm just a knee-jerked like The Beach Boys. I
was big back in the day and I still do good business on my greatest
hits. I still put out some nice new material as well, but I'm not
relevant since. The kids have moved on to video and YouTube is
going to be a thing of the past. It's all streaming as well now.
It's like, as long as I still got my audience.
Rob Collie (00:28:20):
Pretty soon, we are going to be like 3D, hologram, livestreaming,
projecting.
Chris Webb (00:28:25):
Exactly.
Rob Collie (00:28:26):
Probably not going to be my thing.
Chris Webb (00:28:29):
It's nice to feel like I still got an audience. And like I said, I
think my stats are still higher than anything. It's probably
because I'm blogging about something that people actually use
widely rather than Analysis Services, where I think there were only
about a couple of thousand people in the world, whoever used it. So
it was quite feasible for me to have probably met a large
percentage of them at conferences and things. Now, we're getting
towards Excel type numbers, not quite Excel numbers, but the
real...
Chris Webb (00:28:55):
Obviously I can't talk about it when I see the usage numbers for
Power BI and how crazy they are and how crazily quickly they go up.
You just realize that it's like... It's a different scale now.
Rob Collie (00:29:06):
Tom.
Thomas LaRock (00:29:07):
Yes, sir.
Rob Collie (00:29:07):
Let me just tell you that back in 2010, Chris and I were at mortal
enemies.
Thomas LaRock (00:29:12):
Completely understandable.
Rob Collie (00:29:14):
So we found ourselves on opposite sides of something that ended up
being really silly, I think in hindsight. Let's just cut to the end
of it first. It all ends in basically, I'm okay, you're okay. We
were both right kind of situation. But when you don't have the full
picture yet-
Thomas LaRock (00:29:30):
Was this by any chance, a nerd fight at an MVP Summit?
Rob Collie (00:29:33):
It was a nerd fight that played out-
Chris Webb (00:29:35):
The PASS Summit.
Thomas LaRock (00:29:36):
Oh, it's PASS? Okay.
Rob Collie (00:29:37):
It played out over and over and over again, in particular on the
SSAS insider's email distribution lists.
Thomas LaRock (00:29:44):
Right. Got it.
Rob Collie (00:29:45):
The private conversations. Honestly, Chris, I don't even really
remember any single specifics about this. And it wasn't just you
that I was arrayed against, but let me see if I can describe it as
objectively as possible in hindsight. And then you can tell me if I
did describe it objectively. I had struggled with MDX and failed. I
had commissioned projects at Microsoft and hired third party
professional services firms to implement multidimensional Analysis
Services and SSIS projects for me on football, on American football
of all things.
Rob Collie (00:30:17):
So I'd had the experience of being a client of it. I had built the
tools in Excel that were that kind of client of it. I'd sort of
been on every side of the coin there. And then I started using
Power Pivot in 2010 to repeat that project and going, "Oh my God.
This isn't even an equivalent thing. This is better. It's better
when I do it myself." It's not cheaper, faster, it's those things.
It is those things, but it's also, I get better results.
Rob Collie (00:30:46):
So then not jump to the conclusion that everything we know about
the BI industry is going to change. Every last thing.
Chris Webb (00:30:53):
I will freely admit that I was wrong and I didn't really get
self-service BI. I was a typical IT department bigot, I guess. I
was all for self-service when it was people building their own
reports on stuff, on a cube that I already built. But I don't think
I really understood enough about what business analysts were doing
and how they worked to appreciate the appeal of self-service BI.
And I also think, I didn't understand the commercial potential of
it either.
Chris Webb (00:31:25):
Let's face it from a Microsoft point of view. We're not building
beautiful technology for the sake of it. We're building technology
to make money. And I think the success of Power BI in the
self-service space shows what the potential of the self-service
market was. I think I didn't really appreciate it, and I will be
the first person to say, I was completely wrong.
Rob Collie (00:31:47):
Oh, that's-
Chris Webb (00:31:48):
There were a lot of instances like that throughout my career where
I can say, I was wrong or I missed that boat, or I didn't
understand. But yeah, fundamentally that was it.
Rob Collie (00:31:56):
Boy, you were a lot harder on yourself than how I was going to be.
I think that I missed that there was still going to be a place for
enterprise BI. It's like, let's get really nerdy, if you tried to
back in the day, earlier centuries of physics, you try to model
light as a wave and purely do that as like, "Oh, that is almost
perfect," but there's some problems. There's some holes, there's
some places where it doesn't do what we expect it to do. And then
other crew is trying to model it as a particle.
Rob Collie (00:32:28):
It's not a wave, it's a particle, you idiots and they have the same
sorts of problems, the opposite problems. And it's only when they
do the crazy thing of going, "You know what? This thing actually
does both, it behaves as wave and particle. And at different times
it does different things." Until they expanded their minds, they
were never going to be accurately predict anything.
Rob Collie (00:32:52):
I feel like it's one of those things, it's each one of us was
seizing on a truth. And this is actually true of almost any
polarized issue when you really think about it. You see this in
politics, you see it everywhere. It all starts with truth, and
which truth tends to resonate with you most. You cling to that. And
then you end up talking past each other, because everyone's missing
your point. That's how I look back on that era.
Chris Webb (00:33:17):
I think we're still trying to resolve that tension between
self-service and enterprise BI. Because I agree they are both
necessary, and I think maybe I felt like the pendulum had swung too
far towards self-service BI. I think it was probably on your
conversation with Donald Farmer a couple of weeks ago. Hopefully,
it wasn't on some other podcasts, but I'm pretty sure it's only
you're talking to Donald Farber-
Rob Collie (00:33:40):
Absolutely this one.
Chris Webb (00:33:41):
... when you, talking about those days and how paused development
on all corporate BI and everything went into Power Pivot. And I
suppose I can see now that we wouldn't be, we, the Microsoft royal
we, we wouldn't be in the great commercial situation we're in now
without all the investments that took place back then. But it felt
really painful at the time. And especially, I think if you've
invested and built yourself up.
Chris Webb (00:34:06):
And you've had those years of the ego trip and then suddenly it's
like, "Oh, hold on. That great technology that I made my name on
that my whole business was based on is now not being invested in
and there's something else. And there is this guy, Rob Collie
coming along, trying to do his thing."
Rob Collie (00:34:23):
And that guy, Rob, he's so smarmy on those emails.
Chris Webb (00:34:27):
Let's go back to the beach boys, it's like, word punk comes along
or something. The Sex Pistols and my Beach Boys. Yeah, that's going
to hurt.
Rob Collie (00:34:34):
I only need two cords, Chris.
Chris Webb (00:34:36):
Exactly. Exactly. Where are the complex harmonies of MDX?
Rob Collie (00:34:40):
Have you heard Dave Grohl talk about music?
Chris Webb (00:34:42):
No. Go on.
Rob Collie (00:34:44):
For your Beach Boys metaphor, I really think you should just look
up a couple of quick interviews with Dave Grohl about music. He
says things like he hates music reality shows like American Idol or
whatever. He hates it because some of these people get up there and
play their hearts out, make themselves vulnerable. And then this
panel of judges tells them that they suck. That's part of the
entertainment, some of them have to suck.
Rob Collie (00:35:07):
He says, "That's terrible because here's how music really works.
You get together with a bunch of your friends in a garage and you
suck and you're just awful. And then you turn into a nirvana."
Every musician I ever talked to says, "Kurt was a terrible
guitarist." On MTV Unplugged, he's famously making excuses before
he even plays a song that he's going to screw it up. And then
afterwards saying, "Oh, I didn't screw up." He was so insecure
about it.
Rob Collie (00:35:30):
And to this day, Dave Grohl will tell you, he can't even read
music. He's the guitarist and songwriter. He's transitioned from a
drummer to lead guitarist and songwriter and he can't even read
music. He's not trained in music at all. And so I think the punk
metaphor is a really good one. And we talked about this with Lori
Rodriguez about the idea in Ratatouille, the movie, that a cook can
come from anywhere. They don't have to come from a pedigreed
background you'd expect.
Rob Collie (00:36:00):
It turns out though that even though Dave Grohl can't read music,
he's ended up in a different way, in a different language in his
own head learning to do all the things that a high-end musical
professional would need to do. He still has the skill, it just
comes from a different background. We've had this blending of these
two communities, and I think this is gorgeous. I think it is
absolutely amazing that today Marco, Alberto, yourself are writing
in the same language that the Excel analysts also learning to
do.
Rob Collie (00:36:33):
And some of them actually turn out to be like up in that
stratosphere. They get so good at it, they get way better at it
than I am. I find that that lack of an explicit boundary between
those two communities is just amazing. I love it.
Chris Webb (00:36:50):
That's got to a strength. There are just no barriers. If you're
good, you'll get there.
Rob Collie (00:36:55):
Even if you suck, you're really good. Sucking at DAX is in ways
that like I used to do this to myself. I'd look back on what I did
six months ago and go, "Oh, I was so cute back then." I didn't know
what I was doing, and yet that thing that I built was changing the
world in really impactful ways.
Chris Webb (00:37:12):
But I think that's like one of the strengths of the whole... I
don't know whether it's Microsoft as a whole. I definitely felt it
in the SQL Server community. It's just a whole bunch of people who
didn't go to some top college for IT. After I went to university, I
did evening degree in IT, so at least I've got an IT qualification.
But yeah, the number of ex musicians, the number of people from the
armed forces, the number of accountants, and then did something
else.
Chris Webb (00:37:38):
The number of people who've just fallen into it, it's like you
said, you don't have the formal training, but you've got the
experience. And sometimes that experience and the knowledge you've
gained along the way actually makes you better than perhaps
somebody who's just been writing code since the age of six, and has
never, ever done anything else and never, ever wanted to do
anything else.
Thomas LaRock (00:37:59):
If I may. Earlier, Chris, you had said a couple of things when you
were talking about how you were wrong. I just do want you to
understand, you weren't the only one wrong. And you did just touch
upon the SQL Server community, now we'll just call the data
platform community. But there were a lot of the people that did not
understands the shift of what was happening. There's another part
you mentioned where you said you didn't see. When you arrived and
we were having your discussions, you didn't see the commercial
aspect of it.
Thomas LaRock (00:38:29):
And I would tell you that it's because it wasn't evident really to
anyone. I don't think anyone really saw... For me, it was business
objects, crystal reports. This is what we use in our environment
and that was the self-service BI dream. But they were all the same.
They were cumbersome, they're difficult to work with. So you were
right, enterprise BI was still going to be a thing, but what you
didn't see and that Rob somehow stumbled upon was this, I don't
know, 100 million people out there that if they were empowered to
build themselves a Power Pivot.
Thomas LaRock (00:39:06):
And I would look at you, Chris and say, "How many times have I ever
had to pivot a table using T-SQL?" I'm going to say, once for an
exam like 15 years ago. And that's just awful, who would do that?
When I met Rob and I learned Power Pivot and I said, "Oh my God,
why are these people using T-SQL for this stuff? They should just
be using Excel to do a lot of the data manipulation." And these
days, I would tell you, even Python and why you're even in T-SQL, I
don't know.
Thomas LaRock (00:39:33):
But once they got empowered with something like Power Pivot, they
were already familiar with Excel. And you had this merging, you had
that duality that Rob just talked about. All of a sudden, all these
power users said, "Wait a second, bookie what I can do now." And
that was the commercial aspect that apparently somebody knew
existed, otherwise this stuff will never be build. But that's what
took off.
Thomas LaRock (00:39:57):
And a lot of us, and even some of us had just simply believed, but
there were a lot of people that were still just like, "Yeah, it's
not a thing. It's just not a thing, it's not what I'm interested. I
need more nerd knobs inside this database engine because of all
this..." And it's just not a real thing. The fact that I had to
have discussions with people about whether or not the PASS Summit
should have a whole track to business intelligence. Do we really
need that here? Are you kidding me?
Chris Webb (00:40:26):
I remember those years well. The feeling of being like a
second-class citizen or the hate in the room when there was just
too much BI in the keynote.
Thomas LaRock (00:40:34):
Yes. Oh.
Rob Collie (00:40:36):
Like, oh, the thing that the business can really strategically
change it is not just keeping the lights on like basic
infrastructure. The stuff that the C-suite would actually get
excited about, if you could deliver something. Now, we don't want
to add that. You said that I was able to see this coming and most
people didn't. I actually think you had to have had exactly and
specifically my career experience to have seen what I saw. It
wasn't me, it was my path.
Rob Collie (00:41:03):
It comes down to, I'm so lucky that the consultant from Hitachi
Consulting who knew his shit, he was good. I was so lucky that he
was not from the United States and did not know anything about
American Football. Because if he did, the project would have
happened completely differently. But because he didn't know
football, it resembled a regular BI project, where I was the
stakeholder that knew everything about the business domain and had
to somehow educate him on it. And we had to slowly iterate in this
terrible, terrible process.
Rob Collie (00:41:35):
So I had to have been from the Excel group, I had to have had
exposure to Analysis Services from a technical standpoint, from
inside the company. And then I needed to be a customer of it and
get lucky in that way in order for all of these wires to connect. I
don't think there's really anything terribly special about me, it
was just that I had had that path, that allowed me to see
something. And so, of course, I was going to be talking past
everyone.
Thomas LaRock (00:41:59):
And my question to both of you, let's start with Chris. I have a
belief that if you need to manipulate data in some way, DAX is far
more powerful for you than T-SQL would be, right? Is this fair?
Chris Webb (00:42:15):
Yeah.
Thomas LaRock (00:42:16):
Okay.
Chris Webb (00:42:16):
Although, I'm a big Power Query fan, so I'd say, don't forget about
Power Query, as well.
Thomas LaRock (00:42:20):
That's fine. Power Query too. I'm just going to say, when it comes
to the manipulation of data, T-SQL, Transact-SQL to me is a way to
get data more inserted into a database that is to really work with
the data. And that's fine, but my question, and now would be more
to Chris because he's at Microsoft now, and less to you Rob because
you're just Rob now. But-
Rob Collie (00:42:42):
Aargh, I have opinions.
Chris Webb (00:42:43):
You're still special.
Rob Collie (00:42:45):
Pick me, pick me.
Thomas LaRock (00:42:49):
Because I see it from time to time. I see the team in Redmond, they
gave us like windowing functions recently. I see little things that
get added slowly over time like a coffee drip, just some little bit
here and there. But why don't we have a whole suite of functions to
calculate end of week, end the day, end the month, all that stuff?
How many Tuesdays are in a month? That's like a one-line thing in
Excel.
Thomas LaRock (00:43:14):
And you can just simply say, "Done. Now, I got the number, move
on." None of that exists inside of T-SQL. Is there a technical
reason? Is there a business reason? Why can't we get a lot of that
functionality into the database platform as is right now?
Chris Webb (00:43:30):
Well, I suppose this is the kind of perennial product design,
prioritization thing that I've only really come to appreciate in
the last two years since I've been, not involved with it, but at
least I had a ringside seat in how it works out. And I think that's
been the hard thing for me to learn. It's not all about creating a
great product, it's creating a product that will sell. And they're
not always the same thing, sometimes there are marketing
concerns.
Chris Webb (00:43:56):
Those marketing features that annoy the kind of true techie fanboys
like you and me, they serve a purpose, they sell the products. And
if you don't sell a product, we don't have jobs. And there isn't an
infinite resource for developing this stuff, and if you have those
cool marketing features, then that's one less useful feature for
the true believers. And yeah, we definitely see that with Power BI.
There are things that I think get developed that are genuinely
useful.
Chris Webb (00:44:24):
There are things that, at least in my mind would be super useful to
have, but maybe not useful for everybody in every kind of point of
view. But I still passionately believe they should be there in the
product. And there are the things that are there that, well, maybe
people don't always use, but get used for demos or check the box
for Gartner or whoever. And that's not to say that they're useless,
they are really useful. They're useful to make the product viable,
to make people think wow in a demo.
Chris Webb (00:44:52):
To impress the reviewers to and the analysts to make people buy the
product so they can discover the stuff like DAX, which nobody's
going to get really excited about. You do a C-level presentation
about DAX, you'd lose the audience in the first 30 seconds and
people would say, "I'm not interested in this product," even if
that is where the real power of the product goes. So it's a shame,
but there are always these trade offs really.
Rob Collie (00:45:15):
Can I talk now?
Thomas LaRock (00:45:16):
Yes, we still value your opinion.
Chris Webb (00:45:17):
We don't want you to explode with these opinions that are bubbling
up inside you.
Rob Collie (00:45:22):
Part of what you're asking, Tom, comes back to a debate and a
misunderstanding that predates all of this technology, which is why
do we need OLAP? And this is something that we talk about, every
couple of years, Tom, you and I write something about this, which
is T-SQL is transactional. Like Chris is saying, like
prioritization of, not just of product, but even just like of base
direction of a technology. A single technology can't be good at
everything, you got to optimize.
Rob Collie (00:45:49):
And really SQL is fundamentally record-keeping. We used to write
things down on paper, and then SQL came along and that became the
way that we wrote things down. The storage and retrieval, read and
write of records of various sorts is what SQL was made for. And
what I call like the first use of data, the first use of data is
transactional. It's record-keeping, it's making individual
activities happen. Sometimes even across systems like records here,
then trigger the creation of another record over there. And then
next thing you know, I'm being billed by Amazon and my package
arrives. That's the first use of data.
Rob Collie (00:46:30):
Second use of data when you're analyzing it, it turns out SQL is
terrible at that, but the whole world needs to keep records. So
there's jillion people lying around who knows SQL, something like
OLAP comes along and initially it smells like it's the same thing.
But then you get into it and you realize it, it's a completely
different beast. Doesn't speak to the SQL crowd, most of them. And
so we ended up in a situation like when I joined this story in
progress in the early 2000s at Microsoft, reporting services was
the dominant apex predator of Microsoft BI platform.
Rob Collie (00:47:06):
And it's just putting like an HTML transform over the top of a SQL
query is all it is. It has no brain, it doesn't do anything. And
like Analysis Services was the smarter sibling of that product that
no one paid any attention to. It was so funny. Now, we're still to
an extent live a little bit in that world today, Tableau starting
to add that data model, they're starting to, the worm has turned.
But they're a really, really good viewer of SQL views.
Chris Webb (00:47:39):
I had semantic layer, so definitely coming back into fashion. I saw
this article on Medium a couple of months ago talking about
Headless BI and I was thinking all the way through what you're
describing. There was some amazing news platform that somebody had
developed, and what they were describing was OLAP. And OLAP
database, where you could put different client tools on top of some
engine they developed. So it's not a new concept.
Rob Collie (00:48:04):
Hey, Chris, do you know how the PR industry works?
Chris Webb (00:48:08):
Yes, I know, I know, I know.
Thomas LaRock (00:48:10):
It's new to them, though.
Rob Collie (00:48:11):
For those who don't, I've the last few years glimpsed how that all
operates. And it turns out that all the media outlets, they don't
know what to talk about. They need stories to draw clicks to get
advertising revenue and the PR industry is more than happy to feed
them stories. And so for instance, the Harvard Business Review
article from a few years ago that was titled something really
provocative like Finance Departments Abandoning Excel. That was a
PR hit job, absolutely executed, paid for on behalf of Anaplan.
That's what it was.
Rob Collie (00:48:43):
It was an Anaplan ad, Anaplan's PR firm did an amazing job, an
amazing job, creating all this controversy. One of my neighbors
works for them and he's doing very well. That thing worked. And so
this thing, Headless BI, whether it was a PR firm or whether it was
just a really clever marketing department, they're trying to define
a new industry term and trend that just so happens to be about
their company.
Chris Webb (00:49:11):
Yeah, I felt-
Rob Collie (00:49:12):
It's not new.
Chris Webb (00:49:13):
I just get triggered by all these things, regardless. It's like
every time I see somebody talking about the new Excel or the new
thing, I can't help, but get triggered even though I know it's
rubbish.
Rob Collie (00:49:24):
You know who else gets tired of being told about the new Excel?
It's all the people who are good at Excel.
Chris Webb (00:49:28):
Yes.
Rob Collie (00:49:29):
They'd been to that rodeo now. They've been sold that bill of goods
so many times and it's never been that. And it's just the thing
that they use to export back to Excel. After a while, that
marketing message has been used too many times, but I think it's
still effective. It's the thing to shoot at, it's the thing
everyone uses. You've got to talk about it.
Chris Webb (00:49:47):
So long as it still gets a reaction and it still does, then they'll
carry on using it.
Rob Collie (00:49:52):
So something we've run into recently, all this demand, which is a
really good thing for Tableau to Power BI migrations. And if you're
a Tableau shop and you drank that Kool-Aid, that bitter Kool-Aid
and it became pervasive, you got a good infestation going at a
corporate level. And now you're just saying, "Hello, whoa, whoa, we
should really reconsider this and we should switch to Power BI." As
a customer at that point, you have a relatively tremendous
investment in a flat SQL, middle-tier.
Rob Collie (00:50:27):
And Power BI, one of the tenants of Power BI is don't
Franken-Table. You want to keep those Fact Tables separate. It
gives you a degree of flexibility that you wouldn't even imagine.
You won't even know what it feels like until you've seen it. You
can't even describe it to someone how much better it is. And so
what we've run into as a consulting firm somewhat recently, is
customers who have been told that yeah, yeah, yeah Power BI works
great with the Franken-Table.
Rob Collie (00:50:57):
And then our professional advice is, "Well, yeah." And so then now
the customer's confused because they're getting the right technical
advice from us, which is don't Franken-Table. But then, as you
mentioned earlier, you've got to sell something. They do have
jillions of Franken-Tables. Do you need to rebuild your entire
middle-tier in order to switch to Power BI? Have you run into
this?
Chris Webb (00:51:20):
I have.
Rob Collie (00:51:21):
And are we allowed to talk about it?
Chris Webb (00:51:23):
We are allowed to talk. I will talk in a roundabout way, but let me
begin the conversation by saying that one of my jobs on the CAT
team is to work closely with large customers. And the large
customers that I work particularly closely with has a center of
excellence run by the really great guy, and he is somebody who has
thought very deeply on the subject of Tableau migration. And this
large customer was a big, big Tableau customer say, five years ago?
And they are now coming to the end of their migration from Tableau
to Power BI, which they have done pretty successfully.
Chris Webb (00:52:05):
So the things that he's learned along the way are things that we at
Microsoft like to try and reuse, and promote to other customers
because we get asked this question all the time, I mean, all, all
the time. There is a section of the Power BI docs called the
guidance docs, which is owned by the Power BI CAT team, and there
is a section on migration there. A lot of that material is material
that comes from him. But I'll paraphrase what he's learned.
Chris Webb (00:52:32):
Basically, if you think about Tableau migration, there are an awful
lot of companies who come to Microsoft, come to partners and say,
"Hey, we want a bit of this Power BI action. We want to be able to
migrate everything we've got from Tableau to Power BI." Number one
reason is they think it's going to save them a ton of cash, which
it probably will given the relative licensing costs. And at the
moment, everybody wants to save some money. So they think, "Well,
how do you do it?"
Chris Webb (00:52:57):
The obvious first thing to do is go to a consultancy company who is
motivated to come and do this work, to take the obvious analogy of,
to a hammer, every problem looks like a nail, to a consultancy
company, every project looks like a consultancy job. And they will
come in and say, "Let's have a look at all of your Tableau reports
and I will do a mini assessment. And then I will take a look and
see how long it takes to migrate a single report. And then we'll
find the number of Tableau workbooks," or whatever they call
it.
Chris Webb (00:53:28):
"And then we'll multiply that time for a single report, multiply it
by the number of Tableau workbooks you've got. And then we'll give
you a bill." Six months later because it'll only take six months,
it'll all be right and you'll have exactly what you had before with
Power BI at a fraction of the cost that you were paying in Tableau.
And usually what we find is that somebody has come in and slap
Power BI on top of that Franken-Table, as he called it, because of
course, you can get most of the way there.
Chris Webb (00:53:53):
But then the numbers don't properly add up because somebody's
munched together different granularities, and then maybe the
performance isn't great. And the DAX is all suddenly really complex
and horrible because the data isn't modeled properly. And it's
like, "Oh, well this Power BI doesn't really perform very well."
And even if we get to that point, if you think about the sheer
number of reports of any BI product that somebody's got, and even
if you think that it would be... Well, a large company might have
easily five, 6,000 Tableau workbooks or something that need
migrating.
Chris Webb (00:54:28):
And even if you were saying that it took $1,000 of somebody's time,
and can you imagine even doing it that cheaply to migrate a single
Tableau workbook? 5,000 Tableau reports times $1,000, how much that
costs. And think about the number of resources that would be
required to that that no consultancy company could ever do. So
there's a fixed amount of parallelism. The costs become crazy, the
amount of time taken becomes crazy. The technical problems that you
run into because nobody wants to remodel the data or you run into
these issues.
Chris Webb (00:55:01):
And of course, a consultancy company coming and doing this job will
promise, "Yes, we can do it in a fixed amount of time. And yes, you
won't need to do any retraining because we can recreate everything
that you had in Tableau in Power BI. And oh, no, we can't do that,
oh, it's like a disaster, it's completely unusable." It falls flat.
That's what we see the trap a lot of people are falling into. There
is a different way because I've seen it, I've seen a customer
successfully migrate that much.
Chris Webb (00:55:28):
And so the approach they took was different. No consultants here
involved, I'm afraid. Maybe there's an opportunity for you to come
in and-
Rob Collie (00:55:35):
Hey, that's great.
Chris Webb (00:55:35):
... and-
Rob Collie (00:55:37):
I do want to continue with the story, but I just wanted to jump in
for a moment and say, "This is really like the DNA that turned into
our company, is that I knew that industry." I know the business
models you're talking about. I've heard vice presidents at
consulting firms talk openly about like, "All we need to do is get
this client pregnant." They use that word. "We get them hooked. It
doesn't matter what we say up front, once they're in that it can't
get out."
Rob Collie (00:56:04):
It's just so repulsive, the whole thing. In a way, the way that you
just described where that industry operates, we set out to say,
"Look, it can be different than that. The tools have changed now
that it can be different from that." We don't know what that firm
really looks like, but we're going to go find out, we're going to
go build it. I saw a tweet go by recently, I think it's because of
Tom following someone that I now follow someone, and then someone
replied to them and said, "Remember, an umbrella salesman will
never talk to you about building a roof."
Rob Collie (00:56:36):
So a customer or an organization that's able to pull this off
without help, great. That's someone that we don't necessarily need
to be in there helping. There's plenty of people who do need help
and that's where we focus our time. And it's been a rough road
discovering, defining.
Chris Webb (00:56:52):
Yeah. There are other consultancy opportunities involved here. The
approach they've taken and has worked well. Obviously, it's not
something that can happen in six months, it's a several-year
process. But what they did is they, first of all said, "No new
development on Tableau." So everything you had to happen on Power
BI. They were always going to be exceptions, but that was the
general rule. And then they started turning off the Tableau
workbooks reports, whatever.
Chris Webb (00:57:16):
I call them reports because I don't know the great term that nobody
was using. Because let's face it, all BI's got a shelf life and
people build stuff that's useful for awhile and then it just...
People don't use it anymore.
Rob Collie (00:57:28):
Especially, Franken-Table-powered reports. They tend to be built
for a particular question to be answered. They answer that
question, then the need expires, and then that report is like the
SSRS reports. They just proliferate forever, but there's a very,
very, very, very, very long tail of, "No, we don't look at it
anymore."
Chris Webb (00:57:46):
Yeah. You look at the ones that nobody's used for 18 months, delete
those. Wait for the occasional angry emails come in, they can put
that back. Then look at the ones that haven't been used for a year,
delete those. Look at the ones that haven't been used for six
months, delete those. And then at the same time, people are
beginning to use Power BI for doing things. Train people up, get
people excited, get people enthusiastic, and then say, "Well,
listen, we're now going to turn off the ones that nobody's used for
three months, you need to go away and migrate them."
Chris Webb (00:58:16):
And it's basically been a grassroots campaign to get the business
educated so that they're able to support themselves and move what
they need over to Power BI. And rebuild things in a way that makes
sense, that works with Power BI rather than just trying to rebuild
things. Because I think that with any kind of migration from one
tool to another, there is always a tendency and always a temptation
for the business to take the easy, lazy way out and say, "Just give
me what I had before."
Chris Webb (00:58:45):
But if the business is doing that themselves, it gives them...
They're motivated to make the right choice, which is, "Let's look
again at what we want and see if we can reimplement it in a
different way that actually makes sense with a tool that you're
using." And that's the beauty of self-service BI, if you're doing
it yourself, then you can at least express yourself and do what you
need, rather than have to go through that whole tedious process of
explaining to somebody what you want and getting that.
Chris Webb (00:59:11):
And so as a result, after three years of that, they are about to
turn off the last of their Tableau and move everybody over to Power
BI.
Chris Webb (00:59:19):
If you'll, excuse me going off on a little bit of a tangent, there
are two things that I've wanted to say that have been bursting
inside me. First of all, to go back to the whole idea of
self-service BI, an analogy I always used to use when I was doing
training was self-service BI was with typing. 40, 50 years ago, if
you're an important business analyst type person in the company, or
anybody. If you wanted something typed, you would go to the typing
pool and have somebody type it.
Chris Webb (00:59:48):
And there would be somebody there who was really good at typing,
much better than you or I, could type incredibly quickly.
Especially given the technology, they'd type incredibly quickly,
they could spell brilliantly. They could put together a
beautiful-looking professional document and then it would come back
to you and you just sign your name on it, and then it would be
done. And nowadays that doesn't exist, where have the typing
professional's gone?
Chris Webb (01:00:12):
We're typing ourselves badly, putting together badly formatted
documents in Word, just these half semi-literate emails that people
send to each other with emojis and stuff. So why are we doing that?
What happened to the beautiful professionals of typing? Well, it
turns out it's just easier for us to type the stuff ourselves and
get it out the way than actually have to go to the typing pool,
dictate it to somebody. Wait for it to come back, check it as what
we meant to say.
Chris Webb (01:00:37):
Correct the errors and send it back for typing again and so on. And
that's the thing with self-service BI. It might not be done as
elegantly, but at least the interface between your brain and what
you want, there's not another person there. You can just express
yourself directly, and that's so much more powerful. And that's
what you see there with getting the business to take ownership.
Chris Webb (01:00:57):
And maybe the other thing that I wanted to get off my chest, and
again, not to criticize Donald Farmer, but when I was listening to
that podcast you did with him, something he said a little bit. He
said that nobody was using Power BI because they loved it, they
used it because it was the tool that was there. The implication
being that maybe other BI tools like Tableau people use because
they love, and Power BI is coming in because it's cheap and it's
there on everybody's desktop.
Chris Webb (01:01:24):
And I would take issue with that, especially with Tableau
migrations. Tableau is so great at the Kool-Aid of we're so
amazing, we're so beautiful that you can't help, but think, "Is
Tableau a better tool than Power BI? Do the users really want to
use Power BI over Tableau?" And then you see a migration like this
and somebody says to you, "Well, actually we prefer using Power
BI," and it's like, "I'm a Power BI fanboy and sometimes I get a
little bit surprised by that."
Chris Webb (01:01:51):
It's a wonderful, happy moment to feel that somebody is actually
genuinely prefers it. Then you realize that people do prefer it
because it's got different strengths. You're not dealing with some
Franken-Table. You've got that modeling, you've got the power of
DAX to do all of those complex things that people struggled to do
in other tools. You've got Power Query thrown in so you can do your
own ETL instead of having to pay extra for it.
Rob Collie (01:02:13):
I had a similar reaction when Donald said that, but because my
immune system is so advanced in this area, I just turned it into a
thought that I was comfortable with and assumed that that's what he
meant.
Chris Webb (01:02:24):
Sorry, Donald, if you were listening, which I'm sure you are, then
I don't mean it as a criticism. Take it from me, people genuinely
do love Power BI.
Rob Collie (01:02:31):
So I took that sand crystal irritant and I turned it into a pearl.
And again, I don't know if this was actually what he meant, this is
what I turned it into. In my interactions with the Microsoft
product team, it is abundantly clear to me that they're getting an
tremendous amount... And this is correct, this is how they should
be. I'm not criticizing them for this. They are getting a
tremendous amount of their prioritization of what they need to do
from their conversations with top level management at the customers
like CIOs, IT directors, whatever.
Rob Collie (01:03:07):
And not nearly as much as what I would think they should be doing
in terms of appealing from the bottom up. I do think that there's a
little bit... It's not an either or, it's an and, and I think part
of the and has been lost a little bit lately, which is that Power
BI is a better tool. It has a tremendous impact on the lives of the
people that we introduce to it.
Rob Collie (01:03:37):
When I see all of the Excel integration features happening these
days, are ones that CIOs would have asked for and not authors. So
much of my interaction with Microsoft these days, presupposes that
the data models are already there. No one talks about how they come
to be. Again, like you said before, you have to sell things. If you
don't sell things, there's nothing. So, such a loud voice for all
of the complaints.
Rob Collie (01:04:09):
I had kind of a ringside seat when many-to-many relationships
became available in the product. And it was like a knee-jerk
reaction to one big customer saying, "I don't understand why we
can't do this, you guys are stupid. If we can't do this, we're not
going to buy your product." And so next thing you know,
many-to-many relationships were in the product. They were like very
easy to discover, very easy to use, and everyone running around,
stabbing themselves in the eye with that fork.
Rob Collie (01:04:38):
And I feel like the voice of the enthusiast and actually even more
so, the voice of the enthusiast that hasn't found it yet, I wish
that were a little louder.
Chris Webb (01:04:49):
Probably because I'm directly involved in collecting feedback and
bringing it back. There's still an awful lot of customer feedback
that's taken into account and CAT team is directly involved in
collecting feedback. The thing that I have realized is that again,
there are many authentic voices of the user. And let's take data
visualization as a visualization and Tableau type features as a
thing.
Chris Webb (01:05:17):
It's very easy for somebody who is very passionate about
visualization to say, "Oh, Power BI isn't great at visualization
compared to some other tool and Power BI should be spending a lot
more time doing that." And that is one voice of the user. But then
at the same time, when I talk to some users, they say, "Yeah, it's
fine, all I want is a bar chart. I don't want them to think more
complicated than that."
Chris Webb (01:05:43):
And as soon as you get into the really kind of Rolls Royce type
data visualization features, inevitably you will make things more
complicated by making things more powerful. And yeah, that's what
we hear. Sometimes people don't like Tableau because they only want
to do something really simple. And Tableau makes those things maybe
too complicated and Power BI is fairly dumb. Click on a button,
there's a bar chart.
Chris Webb (01:06:08):
Drag the fields into it, and it just works and it gives you a bar
chart is actually better for some of those use cases. There's less
of a mental tax, there's less education involved to get there. I've
realized that there are multiple valid viewpoints, even from the
point of view of the customer. It's difficult to listen to all of
those different sources of feedback and get things, and then
balance that with the need of being able to sell, being able to
integrate with other products that Microsoft have a coherent story
across the power platform, all of those different things.
Rob Collie (01:06:43):
I think where we both definitely agree is that it would be a shame
if Power BI somehow came to be perceived just like the IT-driven
edict, the tool of choice, right?
Chris Webb (01:06:57):
Yeah. The cheap option that somebody's bought because they've done
a corporate-wide deal.
Rob Collie (01:07:01):
Right. Which of course is the reality of how a lot of these deals
happen. This is the weird thing about how the way the world works.
It's easier to sell Power BI at a corporate level on price than it
is to sell it on its merits, which is that it's a fabulously
superior tool. We want it both ways. We love the fact that the
price point makes it easier to sell. You can't go to war negating
your own advantages, you have to lean into this. And at the same
time, you don't want to lose that other thing.
Rob Collie (01:07:30):
And honestly, really, for me, all this comes down to is the
customer, the user at a current customer or at a customer that
hasn't decided to buy in yet. The person who hasn't discovered it.
That's the only thing that I care about that I think could use, and
actually would pay immense dividends for Microsoft, even in their
superior market position these days, is to invest a bit more into
the people that we keep running into over and over and over again,
who are angry that they haven't been told about this. They're out
there and they still outnumber the people who have been told by a
large multiple.
Chris Webb (01:08:08):
It's difficult to know how to reach those people because they've
got day jobs. They're accountants, they're actuaries. They use
Excel in the way that I use my car. I'm not a car fan, I use a car
to get from A to B. I've got like a five, six-year-old car that's a
bit beaten up and I don't really care about it. And various members
of my family keep dropping hints that it's a bit embarrassing and
it's time I bought myself the spiffy new car that I can afford.
Chris Webb (01:08:33):
It's just something that I use from day-to-day, and a lot of people
see Excel like that and it's difficult to know how we could market
them. Maybe we just need to have more big adverts on TV, or at the
airport, or whatever-
Rob Collie (01:08:47):
Yes.
Chris Webb (01:08:48):
... to reach those people.
Rob Collie (01:08:49):
Anything. Anything that acknowledges that the Excel crowd and a
particular segment of the Excel crowd, they're like about a five to
6% slice. That's where your authors are. That's where all of your
future authors are. I think, and rightly so, like when and CIOs are
your customers and you're getting a lot of feedback from them, they
aren't necessarily attuned to that either. When you're talk about
adoption, like driving usage even at a customer who's already
bought Power BI, when you're talking about driving usage up, and
adoption, and all those sorts of things, or what's the monthly
active users? Is that-
Chris Webb (01:09:23):
MAU.
Rob Collie (01:09:23):
MAU.
Chris Webb (01:09:24):
It's MAU.
Rob Collie (01:09:24):
Okay.
Chris Webb (01:09:24):
That's what we talk about.
Rob Collie (01:09:25):
These are the same problems, when I'm talking about surfaces on
your radar. I almost feel like it's like the movie 300. You need
King Leonidas to just pick a handful of software engineers that are
now safe from all other resource allocation. They're earmarked for
onboarding Excel people. It wouldn't take that much, but someone
somewhere has to pony that up, figure out how to either integrate
the Power BI engine into Excel.
Rob Collie (01:09:55):
The equivalent of Power Pivot is on by default as the gateway drug.
PivotTables are the place to introduce people to this. And you know
what? Within 24 hours of introduction to it in a PivotTable
concept, they're embracing Power BI visuals. It doesn't take long.
They are totally cool. You've got to step it. I don't know. Again,
I'm positive that's not what Donald meant, but that's what I turned
it into.
Chris Webb (01:10:20):
Wow. Okay. That's definitely where the big numbers lie, if you can
just tap into the billions of Excel users. That's the kind of state
today I've seen Amir say this in public. He wants to get to Excel
scale with Power BI. I hope that I live to see that happen in my
lifetime because I think it probably is. It's a task that would
take that long. We've made such a good start, it's feasible, I
could believe it. Maybe I'll see it before I die.
Rob Collie (01:10:48):
I think that if it had been taken seriously at that level in like
2010, we'd already be there. You got a couple of things off your
chest. I'm glad that we got to those, those were good. Is there
anything that you were hoping we talk about that we haven't?
Chris Webb (01:11:02):
All right. One more thing I'll get off my chest. Maybe I'll dig it
up, but somebody put a meme on Twitter. I'm not even sure where
it's from. It's one of those famous memes, I think maybe makes more
sense in the U.S., but I think it's those bounty hunters shouting
at each other, that meme.
Rob Collie (01:11:17):
Oh, yes. I think it's the motorcycle guys.
Chris Webb (01:11:19):
Yeah, the motorcycle guys, yeah. Obviously, we didn't get the show
here in the UK, we just get the meme, so a lot of the cultural
context is lost. But it was those guys arguing about, "Can you fix
my dashboard?" And the Power BI guy says, "It's a report." I
remember being at the MVP Summit when Power BI was launched and
somebody going in front of the crowd of MVPs with me and they were
saying, "Oh, and we've got this thing called the dashboards, which
will go alongside your report."
Chris Webb (01:11:47):
And I knew right then that this was the biggest mistake that they
were going to make. And I put my hand up and I said, "No, don't
call it a dashboard. Call it a pinboard, call it something else.
Just don't call it a dashboard because we're going to be cursed
forever with people calling dashboards reports and it's just going
to lead to so much confusion," and it has, it has.
Rob Collie (01:12:08):
Yes. Microsoft can not be trusted with nouns. Seriously. It's like,
you've got to get the nouns away from the people in charge of
naming things. In fact, it's so bad that at one point, even back on
the Excel team, I was telling... From experience, from having been
a noun abuser for a while, telling my new hires on my team that
they had to come to noun court with me if they wanted to put any
new nouns into the product.
Rob Collie (01:12:35):
They could do verbs, they could put verbs into the product without
the high court. But they wanted a noun, they were going to have to
bring it because I was not going to allow that. Again, having been
someone who had done nouns and done them wrong, it took me a while
for me to get over that. I think it's a computer science thing. I
think honestly, people who are drawn to technology are really
enamored of things like the word entity.
Rob Collie (01:13:01):
If I say the word entity and you start getting a little bit
excited, you start getting a little gooey inside-
Chris Webb (01:13:07):
You can see it my eyes down here.
Rob Collie (01:13:08):
... we need to keep you away from naming.
Chris Webb (01:13:10):
Yes, absolutely.
Rob Collie (01:13:13):
And I was an entity guy. I was a computer science, A prime, B
prime, C prime, kind of graph theory is everything. Like, "Nope,
mm-mm (negative). No, it's humans." Learn the hard way.
Chris Webb (01:13:25):
In any of the honesty, if they were building a product that was
going to be used by millions and millions of users. If you're
launching a new brand of soda or something, you'd have focus groups
to look at the labeling, look at a whole bunch of different options
for the names. Make sure it was exactly the right shade of purple
that didn't put people off. Whereas in software, we just put this
stuff out there without any thoughts.
Chris Webb (01:13:47):
Because of course, it's just the name, who cares about it? And if
you are building for the mass, mass market, you've got to be super
careful about these things.
Rob Collie (01:13:56):
Yeah. The same person like me, who'd one minute be adjudicating the
rounding error problem of binary and decimal and some really
obscure corner of the product is at the same time supposed to name
a button? or name an entity? Again, [crosstalk 01:14:13] you and I
have both used the word entity. You used entity to just describe
the organization, I just used entity, we're terrible.
Chris Webb (01:14:20):
I can still remember the controversy about Power Pivot, whether add
a space in it. And to be honest, I can't even remember now whether
it's got a space in between Power and Pivot.
Rob Collie (01:14:29):
It does now because Power View, which we don't care about anymore.
There was a PowerView that was one word that was owned by some sort
of binoculars company. So Power View with a space was A-OK, and so
for symmetry, we had to put a space between Power Pivot. And then
PowerView, no one wants to talk about that dirty secret anymore,
and so now we're stuck with the space in Power Pivot for no
reason.
Chris Webb (01:14:53):
Oh, yeah.
Rob Collie (01:14:54):
Well, Chris, this has been a blast. In a lot of ways, this show is
just a professional vehicle, an excuse to catch up with people, and
have fun conversations like this. I've really enjoyed this one
quite a bit. I'm really happy that you were available.
Chris Webb (01:15:08):
It's a pleasure. Let's do it again sometime.
Rob Collie (01:15:10):
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!