0:00Oh, it's good to be back. Now, one thing
0:08that's been going through my mind for the
0:09last two
0:10years is this problem with Tableau. And the
0:16problem is this, the way Tableau sells the
0:19product is completely detached from the way
0:22people use the product. And that might
0:25sound
0:25like a really bizarre concept and you're
0:27probably thinking, well, what do you mean
0:30by that?
0:30Well, in this video, I want to present that
0:32in two ways. Essentially, present the
0:34problem
0:35and then present the solution. As ever, let
0:37's get started. So we're not going to mess
0:39about.
0:40We're going to go straight to the Tableau's
0:42pricing page. And I'm going to sort of try
0:44and tease this out. And I'm going to use a
0:46couple of knowledges in the process. You
0:48see,
0:49when you buy Tableau, I have to say, most
0:51people are actually removed from this
0:52process.
0:53If you asked every single user of Tableau
0:57how on earth they got access to Tableau,
1:01the
1:01only thing they could probably tell you is
1:03who gave them the license key and who's
1:06responsible
1:06for managing license keys in their
1:08organization. But very few people know the
1:11experience of
1:12buying Tableau. And by buying Tableau, I
1:14mean procurement. I mean sitting down with
1:17Tableau
1:17and negotiating the total mix of creators,
1:20explorers, and Tableau viewers in your
1:24organization.
1:25And as part of that process, what you kind
1:27of have to do is you have to have a really
1:29good idea of what the analysts in your
1:31organization are going to be doing. And
1:34Tableau have sort
1:35of broken that down into these three
1:37buckets. Now, there used to be a time where
1:40these three
1:41buckets didn't exist. There used to be a
1:43time where if you wanted Tableau, you paid
1:46for Tableau
1:46server. At the time, cloud was fairly new,
1:49so Tableau server or Tableau cloud. And you
1:51paid for desktop. And there wasn't really
1:54sort of anything in between. And one of the
1:56downsides of that is that if you just
1:59wanted to use an access report, it was kind
2:01of a
2:01little bit ineffective in terms of cost
2:04because essentially you were paying sort of
2:06full whack
2:07for something that some users weren't just
2:08going to use all the features. And so came
2:11this model. This model which has Tableau
2:13creator, explorer, and viewer. And it's
2:16sort of demarcated
2:17by the expectation of what those users are
2:19going to get when they get access to the
2:21platform.
2:21So a viewer in simple terms is just someone
2:24who really just needs to see reports, needs
2:26to interact with reports, and occasionally
2:29maybe even needs to type into their
2:31keyboard
2:32and get a chart back through features like
2:34our state. That's a Tableau viewer. The
2:37lowest
2:37level but probably the most common user
2:40inside of the Tableau ecosystem. Explorers
2:43are kind
2:43of a halfway house between creators and
2:45viewers. They get given a little bit more
2:48capability
2:48but on the same token, they're also quite
2:51limited. So as soon as they hit the edge,
2:53they're almost definitely going to want
2:54creator access because as soon as you hit
2:56those boundaries
2:57in the Tableau explorer license, they're
2:59quite hard boundaries and it's quite easy
3:01to hit
3:02them if you end up sort of moving through
3:04the product. And so the way Tableau sells
3:07its product, and this is where I think
3:09there's a fundamental problem, is it's dem
3:11arcated
3:11by the way Tableau looks at its product
3:14rather than the way people use the product.
3:17You see,
3:17when you approach the product in this
3:19manner, what ends up happening is you
3:21create arbitrary
3:22boundaries that essentially decide what you
3:26have access to on the platform and how you
3:29basically access them, how you use them,
3:31how they work. And so what that ends up
3:34meaning
3:34is you need a matrix like Tableau have on
3:36their website here that sort of explains
3:39for
3:39each license the activities you can do. So
3:43govern, interact, collaborate and author.
3:46And also prepare. Now it's interesting
3:49because to me, these sort of five items on
3:52the left
3:53hand side, govern, interact, collaborate
3:55and author, prepare as well, are actually
3:58the
3:58core tenants of what makes the product
4:01really, really good. And what ends up
4:04happening is
4:06when features come about, when new features
4:08are sort of drawn up, they get put into
4:10sort
4:11of different elements of this quadrant. I
4:13can almost envisage the meeting inside of
4:15Salesforce now where they're going, right,
4:17we have this amazing feature. Okay, great.
4:19What is the feature? It's a zone visibility
4:21. Great. Who's it going to empower? It's
4:23going
4:23to empower the creator. What is it going to
4:25empower them to do? It's going to allow
4:27them
4:27to build more interactive visualizations
4:30that allow them to do more dynamic sort of
4:32capabilities.
4:33That means they can meet the needs of a
4:35whole range of people. Great. So every
4:37single feature
4:38has to sit in the matrix. And then at the
4:41end of that process, once you've gone
4:42through
4:43this matrix, they then decide which part of
4:45the platform it sits in and therefore then
4:48which price then gets allotted to that
4:50feature. And so as the platform gets bigger
4:53and bigger,
4:54what you end up with is this sort of
4:56massive matrix of not just licenses across
4:58the top,
4:59but now add-ons as well that have little
5:02capabilities built into them. Okay. And
5:06this is where this
5:08detachment from the way the product works
5:11starts to really show. Because as a creator
5:14,
5:14I'm just going to take a creator because I
5:16've been using Tableau for eight years, and
5:17I
5:17think this is the one that I can speak to
5:19the most, but maybe as a viewer and
5:21explorer,
5:22you can put something in the comments that
5:24sort of really highlights this. As a
5:26creator,
5:27what this means is that although there's
5:29one license, there's one product for me to
5:30go
5:31off and buy on the shelf, it means that
5:33there are four products that I really have
5:36to sort
5:37of battle with, Tableau Desktop, Tableau
5:39Prep Builder, Tableau Server or Tableau
5:42Cloud.
5:42And the fourth product in my head is
5:44basically these other management add-ons,
5:46these other
5:47capabilities. Because as a creator, you're
5:49going to be expected to have to manage one
5:51of these three things when the organization
5:53gets larger. So in your head, you kind of
5:57have to hop around the experience. Note
5:59that I didn't say the features, I said
6:01experience.
6:02And experience is a very different thing
6:05because experience is not about, for
6:07example, I bought
6:09this notebook. Okay. So there's so many
6:11notebooks in the world. I bought this
6:13notebook, it's
6:13called the Sidekick, it's from a podcast
6:16that I follow called Cortex, and they have
6:18a notebook
6:19called this Sidekick. It's really, really
6:22nice. What makes this nice isn't the fact
6:25that it's a notebook. What makes this nice
6:27is the experience of using it. If I just
6:29show
6:29you one of the pages, it's got a really
6:31well thought out layout. You probably can't
6:33see
6:34it so well. It's got a really well thought
6:36out layout. The cover folds over and it has
6:39a really nice sort of shiny emboss. This is
6:43some notes I'm making for the 23.1 release.
6:47The experience is what makes this product.
6:49And so that's what I'm really sort of
6:50trying
6:50to narrow in on. And what that really means
6:53is if I just, if you just sort of close
6:56your
6:56eyes as I say this and you just imagine
6:58this journey. Let's say I'm an analyst and
7:01I'm
7:01empowered with a Tableau Creator license.
7:04And I've been asked to go and solve a data
7:06problem in an organization. The experience
7:10today looks something like this. First of
7:12all, I'll connect to my data in whatever
7:15data source. Let's say we're connecting to
7:17Snowflake.
7:17The second thing I'll do is I'll then put
7:19into Tableau Desktop most likely just to
7:21get
7:22a look at the data and see how it works. I
7:24'll then realize that I need to do some data
7:27prep.
7:27So what I'll do is I'll then go back into
7:28Tableau Prep, open another product, open
7:31Tableau
7:31Prep, start cleaning my data and making it
7:33work great except for the end of the Table
7:35au
7:35Prep experience, I can't create a data
7:38model. So now I have to publish that data
7:41source
7:42somewhere. In some instances, I can't use a
7:46published data source to build a data model
7:49.
7:49So what I'll then have to do is publish
7:52that output back out to a database so that
7:54I can
7:54then use it in Tableau Desktop again. So
7:57now we're back in Tableau Desktop to build
7:59a data
7:59model that is going to help answer the
8:01analytical questions that I need. And at
8:03this point,
8:04I'm still working with the connection of
8:05the data and the data prep, the very
8:07beginning
8:07of the journey. And then once that's done,
8:10I can actually start building visual
8:12izations,
8:12making them look great and publish them up
8:15to Tableau Server or Tableau Cloud where I
8:17'm
8:18essentially sharing those things. People
8:21will then look at them, give me feedback
8:22and they'll
8:22make subtle suggestions and I'll need to
8:24sort of traverse my way back through each
8:26and every
8:27one of those products. Meanwhile, if this
8:29is my actual laptop, I have Tableau Prep,
8:32I have Tableau Desktop, I have Tableau
8:34Server open as many times as I've published
8:36the particular
8:37tool. And I also have probably my database
8:40open because I need to have all these
8:42things
8:42going. And in essence, what I'm actually
8:45doing is managing Windows and managing an
8:48experience
8:48across multiple products. It's not really
8:51cohesive, yet I am a Tableau creator. I'm
8:53empowered with all these amazing
8:55capabilities, but they're scattered across
8:57so many different
8:58interfaces for doing different things and
9:01they don't sort of come together cohesively
9:03,
9:03if that makes sense. Yes, they all connect
9:05to Tableau Server and Tableau Cloud, but in
9:07terms of my workflow, the thing that
9:09actually makes me work better and faster, I
9:12have to
9:12traverse all of these different interfaces.
9:14And so that to me is the fundamental
9:16problem.
9:17And I know this exists in the viewer
9:18license and it exists in the explorer
9:20license as well.
9:21I'm purely talking from a creator license.
9:23So what is the solution? Well, for this, I
9:27take inspiration from something I do to
9:30make you guys' videos. One of the things I
9:32really
9:32appreciate about the software I use is that
9:35pretty much all of them have had the
9:37benefit
9:38of years to understand the specific needs
9:41of their users. And so one tool I use to
9:44edit
9:45videos for this channel is called DaVinci
9:48Resolve. Now, some of you might have heard
9:50of other tools such as Camtasia, which
9:52essentially does the same thing. You might
9:54have heard
9:55of Adobe Premiere Pro. They're all the same
9:58. The fundamental thing all these tools
10:00understand
10:01is that when it comes to making a video,
10:03there are different steps in the process.
10:05And so
10:05if I take DaVinci Resolve, which is what I
10:07use for editing videos, I haven't got an
10:09actual
10:09project here. I've just loaded it empty so
10:11you can see the point I'm making. The first
10:14step, and what I really want to sort of
10:16call out is what's happening here at the
10:17bottom
10:18of the window. I'll sort of make sure I
10:20highlight it on screen. At the bottom of
10:22the window,
10:23the process that I go through is broken
10:25down into different steps. Each icon
10:28represents
10:29a different step in the process. So the
10:31first step is called media because that's
10:33where
10:33I bring in all my recordings. As soon as I
10:35finish recording this video, I will get the
10:37files on my computer and I'll drag them
10:39into this interface. And once I've dragged
10:41them
10:41here, I can see them, I can mark them, I
10:43can label them, I can tag them. The next
10:46part
10:46of the interface is cutting them. So once I
10:48have all my clips in a folder, I can
10:50essentially
10:51cut them up into sort of short video
10:52segments that I'm going to use and I can
10:54bring them
10:55around, put them together, kind of get an
10:57arrangement and an order that works until
11:00I'm happy. That's sort of called the first
11:02cut in video editing. The second thing is
11:04now the actual edit. The edit is where you
11:06add the finesse, the titles, the labels,
11:09the
11:09things that sort of make the video a little
11:11bit more interesting. And you also add
11:13transitions.
11:14You cut out certain things that might be
11:16distracting. You might add zooms and pops.
11:18All of this
11:19stuff gets done here in the edit. That's
11:21where you're adding to the creation. And
11:24then after
11:24that, you might want to do some special
11:26effects. And you can see here as I go
11:28through each
11:29of these steps, the interface is
11:31fundamentally changing to enable that one
11:33activity because
11:34it understands that it's guiding me through
11:38a specific flow. And so if I sort of close
11:41this and we just go back to talking about
11:43Tableau as a creator, it seems that the
11:46version
11:47of Tableau, the perfect version of Tableau
11:50for a creator is the version where you don
11:52't
11:53have to have five things open to get your
11:56work done. I was about to say job to get
11:58your
11:58work done. And so what does that look like?
12:02Well, just bear with me here because what
12:04I'm going to do is I'm probably going to
12:06put like a graphic up on screen as I talk
12:07to just
12:08sort of narrate this. In my head, the first
12:11thing any analyst will do is connect to
12:14their
12:15data source. That's just the number one
12:17thing. Go find your data, bring it in,
12:18connect to
12:19it, make it work. The second thing you'll
12:22do is you'll want to start to do data
12:24visualization.
12:26And so in that sort of visual step, you're
12:28probably going to miss the really crucial
12:31step of data preparation and data modeling
12:33because you'll go straight to the
12:35visualization
12:36step. And when you get to the visualization
12:38step, you might then at that point realize
12:40you need to go back to do some data prep or
12:42you need to go back to do some data
12:44modeling.
12:45And that prep and modeling, I separate them
12:47as distinct steps because in my mind, they
12:50can kind of happen in a different order and
12:53in a different way. When I talk about data
12:56modeling, I talk about the relationships
12:58between the different tables, the way that
13:00Tableau
13:01allows you to do that. I think Power BI
13:02also lets you do the same thing. When I
13:04talk about
13:04data prep, I talk about explicitly shaping
13:07the data. So making sure the data is in the
13:10right shape or orientation for you to do
13:13what you need to do. So prep is about row
13:15level
13:16sort of adjustments and calculations and
13:19finesse. Data modeling is about sort of
13:21slightly higher
13:22level thinking about how the different data
13:24sets come together to enable analytical
13:26answers
13:27to be done. And so if I draw this line out
13:30for Tableau, what it actually looks like is
13:33you have your connection to data, you have
13:35your data prep, you have your data modeling
13:38,
13:38you have your visualization as the next
13:41step. And the visualization is different
13:43from what
13:44I would then call layout because dash
13:45boarding is actually a layout exercise. It's
13:48actually
13:48a design exercise. And so when you're
13:50working in a sheet building a chart, you
13:52need very
13:53different tools to when you're building a
13:55visualization. Actually, you get a little
13:57bit of that in Tableau. If you want to
13:58create a dashboard, you have to go into a
14:00different
14:00interface to do that. So that is already an
14:03understood concept inside of Tableau. And
14:05once you've done your dashboard, you want
14:07to then be able to publish it up, share it
14:09with people. Okay, so these are I think six
14:12steps. And the ideal version of Tableau, if
14:15they were really building the product to
14:18match the way they sell the product, is
14:20that they
14:21would offer the Tableau Creator experience
14:23in one application. One application that
14:26lets
14:26you do all six things in one place. So you
14:29don't have to switch tabs, you don't have
14:31to switch applications. And more
14:34importantly, when you save your work, what
14:36you're actually
14:37saving is the entire workflow as one piece
14:40of work. And the benefit of that is all the
14:43metadata of how that analytical question
14:46gets answered goes up in one file. And it
14:49enables
14:50other people to come in, look at different
14:53parts of that flow, and create tangents. In
14:57development, they call these forking
15:00projects in GitHub. I've not sworn, my int
15:03onation is
15:04maybe not good enough. But when you're
15:06developing as a developer, it's called for
15:08king, because
15:08what you're essentially saying is that I'd
15:10love to take this project and create a tang
15:12ent
15:12that goes off in a slightly different
15:14direction that suits my needs. And it would
15:17really also
15:17start to open up amazing capabilities for
15:20collaboration. Because what you can say is
15:22that, hey, I don't want the visualization,
15:24I don't want what the person has shared up
15:26to the server. But I love all the modeling
15:30they've done, I love all the data prep they
15:32've
15:32done, and I'd love to use that for my own
15:35creation. And because everything is one
15:37file,
15:37it's just much, much easier to share. And
15:40the other really important thing is it
15:42actually
15:43just leads to much, much more enjoyable
15:45experience. And I think it maybe frees up
15:48each of these
15:49steps to be more of what they need to be,
15:53if that makes sense. So if you look at
15:56Tableau
15:56Prep, as a creator, it's really hard to
15:58understand why that's a separate product.
16:01Why is that
16:01a separate piece of software if as a
16:03creator have access to both of them? An
16:05explorer doesn't
16:06have access to a Tableau Prep, they can't
16:08build with Tableau Prep. The only thing
16:10they
16:10can do is start visualizations with
16:12existing data sources. So why on earth is
16:15that a separate
16:16product? It makes absolutely no sense. It
16:18should just sit inside of Tableau Desktop.
16:20And it might be because you don't want to
16:21sort of encumber the whole product. And it
16:23's
16:23probably really good technical reasons why.
16:25But if you were just sort of, you know,
16:27really
16:27meeting the promise of picking the product
16:30off the shelf, and understanding what the
16:32product is, the one product will be called
16:34Tableau Creator. And the Tableau Creator
16:36allows
16:36analysts to build analytical workflows from
16:40start to finish. That's data prep,
16:42connecting
16:43to data, modeling the data, visualizing it,
16:46laying it out, and then sharing with people
16:49in the organization. That is the creator
16:52promise. The explorer promise starts after
16:56the data
16:57modeling step. So because explorers can't
16:59create their own data sources, they can
17:00only
17:00use certain parts of data sources that
17:03already exist, they would essentially be
17:07coming in
17:07halfway through this sort of flow. So they
17:09would only be really getting access to the
17:11latter half of that. So the visualization,
17:15the sharing, strictly speaking, the
17:18visualization,
17:19the layout, and the sharing. They wouldn't
17:22be able to connect and prep, essentially.
17:24And then the viewer would just get access
17:26to the very end part. Just be the ability
17:29to look at things, and then ask data. And
17:31then everything else can kind of slot
17:33around
17:33it. But to me, that would mean there'd only
17:36really be three products. The creator
17:37product,
17:38the explorer product, and the viewer
17:41product. And that would be one cohesive
17:43experience
17:44across the entire platform. I'm editing
17:46this video, and I realize I made a slight
17:48mistake
17:49here. What I meant was there'd only be one
17:52product, three experiences. It's just got
17:54those completely switched around. So there
17:56'd be one product, it would be called Table
17:58au,
17:58one application, you download it, and
18:01whatever license you have opens up
18:03different parts
18:05of this experience. As I've just laid out
18:08the timeline, different licenses open up
18:11different
18:11parts of that timeline. And all the other
18:14features around add-ons and so on and so
18:16forth
18:16just slot into those particular pillars and
18:19open up different capabilities. But the
18:22product
18:23vision is super clear. What the product
18:25does is super clear. And also it
18:27communicates very
18:28easily how you get access to do more,
18:31depending on the license you have. It's
18:33just all in
18:34one place. It's just much much easier to
18:37understand. And so that's my take. That's
18:40my view on this particular topic. Now,
18:42there are massive flaws with this idea. And
18:45it's
18:46something that I've sort of been mulling
18:47around in my head. And I have talked to a
18:49few people in the community about it.
18:51Sometimes sort of off-book, sometimes off
18:54the record
18:55as well. And it's super interesting to see
18:57sort of some of the thoughts that come out
18:59of it. But this is an idea. I'm not
19:01pitching this as the end all and be all.
19:05There's
19:05tons of issues with this idea. I've thought
19:07of many. I'd love you to sort of highlight
19:08what you think they are. Maybe that's a
19:10separate video. But to me, that vision of
19:16how the product is sold and how the product
19:19is built is starting to get so fragmented
19:21that I think it's getting in the way of
19:24really delivering the experience that users
19:26deserve. Not necessarily the features.
19:28Anyway, I've waffled for way too long. I'll
19:31stop
19:31the video here and I'll catch you in the
19:33next video. I think 23.1 is about to drop.
19:36Probably the day after I'm recording this,
19:3823.1 will drop. And I'll get punished for
19:40not having done any prep for 23.1. But as
19:57ever, I'll catch you in the next one.
Tableau | Tool strategy | Data visualisation