Transcript
Okay. And we are at time and at quorum as well. So, really appreciate everyone who's come and joined us for this webinar today. This is Self-Service BI: Cutting Through the Hype run by InterWorks.
We are a full-stack BI consultancy who's been working in the BI space for a few years now, coming up on twenty-six years at this point. But I really want to thank everybody who got our invitation and who joined. As a quick reward for starting that, we're actually going to start out with a poll question just to make sure that we're here and we've got everyone back. If you could come in, the poll should have popped up on your screen right now.
We're just curious to start out. What BI platform are you or your company using right now? We've got one choice. I apologize if we can run this as a multiple choice next time, but we've got Tableau as an option, ThoughtSpot, Sigma, Power BI, Looker or other.
Feel free to throw it in the chat. I'll be monitoring the chat throughout this webinar, and we'll be sure to give you a shout-out as you're going in. We've got Focus from Steven. Thanks for typing that in there.
And please, everyone who has answered again, feel free to go through and do that. Got a nice mix of tools right now. Looks like primarily Tableau and Power BI are our clear top two. So thanks for everyone who answered for that.
It's always nice to see the context there, and we'll use that and that will inform things going forward. You can pop the results out to the screen really quick so you can see our distribution. But just know that as we're going through, we're going to look at different BI tools and how they relate to self-service analytics, because it's something that we at InterWorks have run into time and time again. So our mantra is really doing the best work for the best clients with the best people.
And all that means is we do our best to make your BI journey something that is useful for you. We want to work with the best people. We want to work with people who are kind, who are smart, who are trying to solve the problem, but also who are realists at heart. It is nice to meet people where they are and go through and have a good time working together to try and solve data problems.
We don't have to be isolated. We don't have to be overly professional, but we can all have a great time working together. And that as a strategy has served us well, both in the US where I'm based. Apologies.
I'll do this multiple times, but for my outrageous American accent, it is what I was born with, so I apologize for that. But across Australia and APAC as well, we have a great list of clients here who we work with regularly, appreciate their help and appreciate the chance that we have to help them with their data journey.
Along the way, when we look at this group of clients that we work with, we serve those clients by helping bring in partners who are our technology partners that we work with to provide a best-in-class experience. So we'll talk about different partners that are on here, but very excited about all elements of the modern data stack, whether we're looking at a cloud data warehouse like Snowflake, whether we're looking at ELT tools like Matillion or Fivetran, got our BI tools working with Tableau and ThoughtSpot. We have our data science tools with Dataiku, and then we have our infrastructure tools and our platforms that we work with AWS, Dell, Microsoft, and we even work with Google as well. But the point is, we meet people where they are with the partner that makes sense to help them be successful with their context and their content.
Now, little bit about me. I mentioned that I'm from the US. I'm based outside of DC, and I live there with my wife and our three little girls. We got three girls, three separate hair colors.
Not sure how we pulled that off, but we've got Mabel there on the left. I have a blonde Wonder Woman. We got Joan, our oldest in the center, and then Ruth, who loves all things unicorns, both in balloons and face paint. So that's the three of them.
They're asleep in the other room, thankfully, so shouldn't have them bounce in. But if they do, I'll be sure they say hi. And then there's my wife and I together. We enjoy traveling.
Haven't made it back to Sydney. My wife lived in Canberra for a few years, and because of that, she's a bit burnt out on the trans-Pacific flight. So working to sell her on taking the rest of us back there.
But maybe I'll try and leverage my hobby at the outset to get there, which is triathlon. So I love this mug. It takes exercise and makes it three times harder than it needs to be. But for me, it's a fun challenge as I am getting older despite the baby face.
But I have a good time working with that. It keeps me busy along with the kids and the spare time that I have there, but didn't necessarily come here to learn a lot about me.
We'll definitely dive into what you're here for, which is this concept of self-service BI. So self-service business intelligence has been thrown around as a buzzword for years. It's something that more and more people are asking about. It's something that is published in every single industry publication.
It's seen as the gold standard for how to use interactive data or how to enable your organization to become data-driven. Self-service BI is a key tool to make that happen. Now, unfortunately, with that, there's a lot of hype that comes with that. So everything gets labeled as self-service.
So we found that it's interesting with our clients to start with a discussion of where does self-service even come from? What does it mean? What's its origin? What's its history there? So I'll run through a little bit of that background, both in self-service as a concept, but also how it applies to the industry.
And then we'll tie that back to your day-to-day and your tool selection. As a note, as we're coming in here, I have the chat up on my screen. I also have Q&A available. So if you have anything that you want to comment on, please feel free.
If you have any questions throughout, you're welcome to drop them there. I'll do my best to get to them in the flow. I may have to reserve them towards the end. But either way, we'll make sure to get to your questions while you're here.
So please feel free to utilize those.
So we'll first get started with the history of self-service as a concept. So I don't know if anybody on the call has heard of Piggly Wiggly. It is not just a fun thing to say, but it's one of the origins of self-service from the retail sector. So Piggly Wiggly was its own supermarket based in the Tennessee area where they came to the realization that they were wasting a lot of time and a lot of manpower having clerks behind the counter, taking everyone's order, gathering their food, packaging it up for them, and handing them out.
The real innovation here was coming in and saying, hey, let's actually change from having all of our goods behind the counter to having aisles with food that people could choose from. People could make their own choices, bring everything and put them inside their nice wooden baskets that they started out with, bring them up to the front. They'd ring them up like normal, package them up, and send them on their way.
There were other imitators in the area that quickly picked up on their ideas, but Piggly Wiggly gets credit, at least in the US, for originating the idea of self-service retail.
This quickly spread outside of supermarkets to all different industries, and we ended up with the point where it then came to the BI industry. And now as that has accelerated, basically every aspect of the modern data stack or every modern BI tool will claim to have self-service analytics.
So here we've just got a series of links that will go out. I'm not going to run you through all of these. You can take a screenshot and take this as homework if you'd like. You don't need to.
But the idea here is that we have ThoughtSpot, Fivetran, Matillion, DBT, Snowflake, Tableau, Power BI. Everybody is saying, here is how you do self-service BI or self-service data transformation or self-service analytics as a general term. So it's this thing that everyone is aiming towards. But, again, what does that actually mean?
So we're playing with some definitions here.
Let's first agree on what self-service BI is not.
Coming up swinging here a little bit for those of you who are diehard Excel fans. But self-service BI is not Microsoft Excel.
And I say this as a diehard Excel user and fan. If you want to chat in the Q&A or stay afterwards and we can talk about how index match is better than VLOOKUP, I'm happy to do that. Or if you really want to take it to the next level and play around with XLOOKUP and how we can now return arrays and replace both of those, I'm for it.
What we're talking about here though is self-service BI and how that solves some problems that Excel tends to exacerbate. And again, as an Excel defender here where it has a place that we need to continue to use it, we can cast the net wider. I can target Excel. I can target Google Sheets. I can target Numbers. The point is, let's call it spreadsheets.
They are a necessary part of our day-to-day environment. I don't know of a single business that runs without them, but it is not self-service.
And we'll dig into that a little bit because if we're looking at the rise of self-service or where self-service started, I actually tend to hearken back to this little stone here. If we were in person, I'd have people raise hands and comment on what these are. If you can throw it in the chat, if anybody knows what language this is or where it comes from, if you want, you can look at the link on the bottom to see what museum it's housed in.
But this is from the Ashmolean Museum at Oxford University over in the UK.
And this is one of the earliest examples of writing that we have.
This is ancient Sumerian, and when they found this, it's a tiny little thing, roughly, I don't know, let's call it five centimeters square.
They were really excited, and they found a key and they were able to translate it. And the translation that came through was not some breaking news story. It was not some long-lost love letter between two friends. It, in fact, was a transaction record.
I sold two bushels of wheat. He gave me two caskets of wine.
Not very exciting, but if we're talking self-service analytics zero point zero, a starting point, to me, I can't look at that and not see a spreadsheet. They are very similar. Rows, columns, great way for listing out items, listing out transactions for us to get to a point where we have data that's available and we can work with it. So again, the argument here is not let's replace Excel. The argument here is let's let Excel be good at what Excel is good at and then use BI to augment that and improve on it.
Now it begs the question, if it can be such a good transaction ledger or if it works so well, why can't we just use Excel or Sheets or Numbers?
And there are really three major flaws in any of these tools, whichever one you deploy on, and they all have to do with your data and how it's accessed and how it's used. So I've got them up here on the screen. We can talk through each one individually. Number one, the major flaw of an Excel-based self-service environment is security.
If you are working in .xlsx files or specifically just working off of files, regardless of how many passwords you put on it, regardless of how many different ways you try and lock it down, fundamentally, you're dealing with a file that somebody can pick up, attach to an email and send to somebody else. There is no way around that.
And so it is a massive potential flaw in your environment where data can get out of hand very quickly. You want to have something that has proper access control.
You want to have something that is locked down as opposed to something where everybody's downloading their data, doing their own thing, and then it gets shared out to who knows where.
Closely related to that is the concept of governance.
The worst thing in the world is when two people come into a meeting and the CEO at the time stands up and she says, hey, what were our fill-in-the-blank numbers for last quarter?
Both people raise their hands and say, I've got the number, and then the numbers disagree.
You need to have data integrity where people are saying the same thing and content is governed where you have gate checks to avoid that type of scenario.
You want to make sure that your production environment is your production environment, not a glorified dev environment. You want to make sure that people are working together. So when a number goes out to the company or reaches the CEO's desk in her very limited time, you are able to actually be confident in those and confident the right people are seeing the right numbers.
You have to worry about that in an Excel environment. Again, we're based off of files. You're not able to make sure and enforce that that would happen. And then the last one is scalability. I don't know if anybody else has managed to break the fifty megabyte limit on Excel files. I have.
And it absolutely slows down and destroys the experience. There's a row limit of one point one million for a reason. In these days of big data, in terms of quick streaming data, in terms of tracking interactions that our customers do across all of our different websites, businesses, interaction points, apps, a million rows is table stakes. You need to go significantly further than that.
Then when we deal with those challenges, right, security, governance, or scalability, I don't know about you, but for me, those scream IT.
Right? So if we want to have a successful deployment in terms of self-service analytics, we cannot ignore IT. In fact, they are central to making this happen. Because if I'm looking at having a secure, governed, scalable data environment, that does not happen without IT involvement. There has to be a partnership there between your business users and your IT users.
And if we're looking at the history of where self-service BI comes from, that's actually where things started. So we'd run through a little of the history of the industry, just where collaboration started in terms of BI tools and how they work together.
You could also narrow that scope down from the industry to an individual company. So this could be your company, likely where things started, or you could look at yourself on your data journey and say, where am I along these steps?
So take it on a market level, take it on an individual client level, take it on a project level, whatever works for you.
The point is, when it starts out, we have very promising beginnings where we have business users and we have IT.
IT is there in blue, business users are there in gray, and we have requests coming in.
Business users would like some data. IT says, yep, I got you covered, turns around the request, they have what they need and go on their merry way.
Word spreads when that happens. IT is now servicing two people, able to help out or two different departments, and things are increasing, and they're seeing good momentum.
Now, unfortunately, IT is a limited number of people. They do their best, but when they're not able to get to everything, a backlog starts to creep up, but it's nothing huge. They're able to be successful as they're working together. Self-service BI is growing.
Unfortunately, word keeps spreading, and poor IT ends up being targeted more and more. More requests are coming in. Hey, can I just get that number?
Hey, could you pull this for me really quick? Could I do this, this, this, and this? And as a service organization, IT steps up and says, yes.
We'll do our best.
But the problem is they get overwhelmed with requests rolling in and the backlog starts growing to a point where it's unsustainable.
Now that happens because at their core, when we're looking at data and what we need data to do, there is a natural and inherent tension between IT and business.
So IT, again, secure, governed, scalable data. We want to make sure that what we have gets to where it's supposed to be, is accessible to those who need to have it, and is something that can scale to whatever scale we want.
On the business side of the house, they say, sure, I want it to be secure, but I also want to share it with whoever I want to.
Fundamentally at odds. Governed datasets, governed processes. Awesome.
In terms of making sure the right data shows up in the right location, business usually will come in and say, that's great. I would also like to just copy-paste in this Excel file that I got emailed from a vendor, and I want to then mix those two together and then shove that out to the group really quick because they need this answer by a meeting tomorrow morning, which then leads to the speed element here. Where business needs it fast, we need it yesterday if possible.
And we don't really know what we want to do. So I don't have time to plan out a massive, scalable, massively parallel infrastructure that's able to keep up with our demands because I don't know what they're going to be. Right? IT is making sure that what we know performs well and is secure and stable, whereas business keeps going out and finding new things to make IT's life miserable.
Not necessarily miserable, but there's an inherent tension there. And not just from their day-to-day goals, but also just in terms of number of employees. Right? We can simplify it down to as simple as there are a lot more people outside of IT than there are in. So if you factor in sales, marketing, executive team, finance, accounting, fill-in-the-blank department, you then just have more headcount than you do who are working in IT.
And so you end up in a scenario where IT is getting overwhelmed, where there are more requests than they can handle, and so they bring in help.
This is a natural evolution. As the market grew, IT was saying, this is not enough. We need something more. And so they started to go out and hire, but they hired for a couple specific skill sets.
So I want to introduce these two new players. One right here on the left is our technical developer. This is somebody who is able to go out and build content to help distribute reports to the environments or to the different business users. And we have a technical analyst as well, someone who's generally a little bit more number-savvy who can dive deep and answer the more complex questions. So if a dashboard doesn't answer the question for you, the technical analyst can come back in and say, great, let me pull some of that information for you. We'll get you that deeper dive analysis that's needed.
Unfortunately, even with these two added on, the backlog continues to grow. Again, boil it down to as simple as just volume of users.
Even when IT hires on help, they are still hopelessly outnumbered and hopelessly stuck behind this backlog in this situation where they're saying, yes, I will help you as soon as I can. We need to get it in line. We need to prioritize. We're going to work our Jira board. We're going to work our sprint methodology. We're going to make sure we have points and know what the company's priority is, and the backlog continues to grow.
So we've got a situation where there's an inherent tension between business and IT, which then leads us to need to, again, take a step back from this specific scenario of those two trying to work together and doing their best to map that onto where self-service analytics tries to come in and solve this inherent tension.
Now, I have a confession to make. I did go to business school, which means contractually, I have to put a two-by-two matrix into every presentation that I do.
It's not actually the case, but it sure feels like it sometimes. So I apologize for the two-by-two. We're going to start with a blank. We're going to fill it in. And the goal here is that as we assign different groups to their different quadrants, we're going to figure out where different tools fit. We're going to be able to go through and say, oh, I get it for this user. Here's my expectation of what they should be able to do, what they should be able to analyze, and what we need to do to enable them to be successful with self-service analytics.
So we're actually going to start in the bottom right.
So we're on the bottom half of our screen, which means we're looking at the professionals.
These are people who it is their day job to pull information, to live with the data and make sure that it's where it is successfully.
And then we're on the right side of the two-by-two, which means they are doing object consumption, which means I am going in, I'm pulling data, I'm creating my own datasets. I'm going in and I'm saying, oh, you have a question? Let me build out a novel analysis for you.
This is where IT lives and starts.
From a tool selection, at this point, you're usually looking at SQL as the main, most commonly used data or source of data that's being pulled from. Right? We are seeing a lot of momentum in the market around Python. There's a lot of people, especially from academia, who will prefer R for that. But the point is you have some sort of scripting language that's being used in order to access and interact with your data.
When IT brings in help, we typically see a split across this. So we still have these professionals where their day job is working through data and modifying, creating, sharing, using. But that's your day-to-day. Nine to five, you show up, you pull data, you work with data, you distribute data. And we've got our technical analysts there on the bottom right trying to build out these analyses, trying to use SQL, trying to enable other people to do their own thing.
And then we've got our report creators, our technical developer who is building out reports. Self-service analytics one point zero was very much around burst reporting. We were saying, hey, business users, you're here and I want you to get access to data.
So what we're going to do is we're going to build Excel files that can connect out to SQL, or we're going to build you some Crystal Reports or an SSRS burst or Cognos where you can type in a start and end parameter, hit go, and then download an Excel file.
That's self-service analytics one point zero. The report developer in the bottom left is there trying to predict all of the needs from the end users. And then the main output from that again was spreadsheets.
That's where things started.
But the problem was none of the tools that we have in this portion crossed over to make it over to our citizen developers.
Right. It was not tenable for a citizen developer to say, you know what? Today, I'm going to become an SSRS developer.
Anybody who did that immediately marked themselves as a technical hand and they got sniped by the technical team. It was not something where your average user could log in and say, you know what I'm going to do today?
I'm going to teach myself Python so I can pull my own data.
Now that's changing a little bit. And if you play around a little bit with ChatGPT and other LLMs, the ability to or the barrier to being able to write Python has decreased dramatically. So that's something that we're watching very closely in terms of where things are going. But still, you don't have your average user, your citizen user using these tools.
That's where we lived with self-service analytics one point zero.
If you're an organization just getting started, you may be at this point.
Lots of burst reporting, lots of Excel files that can get sent out, but you're wanting to scale further.
And so it's at this point that we actually need to talk about the elephant in the room. If you know anything about InterWorks, you'll know that we have been a Tableau partner for a very long time, longer than I have been listening to my kids sing We Don't Talk About Bruno.
No joke, in my musical rewind for 2022, my number one song of the year by far was We Don't Talk About Bruno.
And so is this a scenario where that's just stuck in my head and we're going forward? Not necessarily, because there is something that we haven't been talking about so far, which is the fact that on our intro, when we're talking about our partners at InterWorks, we have two competitors up here.
There's even a third if you bring in Microsoft, if you want to talk about the Power suite here. But our main partnerships in the BI space are ThoughtSpot and Tableau. Tableau is actually where we started. That's where we cut our teeth. It's where we made our name in the industry. So it begs the question, why are we talking about ThoughtSpot these days?
Or in all of this, we've talked about self-service for half an hour now, and we haven't mentioned Tableau once.
So we may not talk about Tableau in a ThoughtSpot scenario, but we are going to do a whole presentation where we do bring it up. Right? You have to talk about Tableau in the context of self-service BI two point zero.
To give credit where credit is due for what Tableau did, Tableau created the two point zero market.
Yes, there were other tools that were around. Yes, there were other groups that were doing that. But you don't see any other BI tool that has a twenty-thousand-person conference with a passionate community getting together to say, hey, we are data people. We are excited about what we can do with data and visualization and visual analytics.
Tableau deserves credit for opening up data to the masses in a way that hadn't been done before, and that was done in tandem with this technical developer.
So what Tableau did is it made it possible for your technical developer to make beautiful reports, for consultants to come in and build out a whole suite of reports very quickly that can change the way you look at your business. Right? There's a reason I have a job is because people will ask us to come in and build this content. But what it also did for this business user on the right half of the screen was they're saying, I want data. Tableau made it so you could connect out to anything.
Excel file, Access database if you needed, Oracle, SQL Server, Snowflake, Redshift, BigQuery.
Doesn't matter.
Grab your data, pull it in the extract for your acceleration engine, and then you can go.
You can visualize. You can explore.
So Tableau deserves credit for being the first tool to make that jump where citizen developers began to be able to make their own stuff. So back in the day, self-service analytics one point zero, I wanted a reporting tool that distributed my content, made it so lots of people could see this.
Lots of people could see the insights that we're creating.
Too big of a lift unless you were a professional. Tableau changed that paradigm. So now everybody could become curious. Everybody could come in and start asking questions they hadn't looked at before, and your business users started to be happy. And that helped with the backlog because you were able to go through and participate in self-service analytics two point zero.
Oh, and as a side note, tip from David Noonan here. Hot tip from a new dad, you can now tell Spotify to ignore songs and artists from your recommendation model. I like that.
Having said, we're a YouTube Music house, and we try and hook it up to the speakers so the girls don't have to have screens when they're this young. And so we're actually okay with that, but it does mean that my recommendations get skewed all over the place. So I'll see if we can play around with that. I appreciate the hot take.
But coming back to self-service analytics two point zero, in working through this, really what two point zero is characterized by and seen as a big success was the idea of visual analytics or the dashboard. So leading the charge at the front of the visual analytics movement was this concept of the dashboard. We're not dependent on burst reports getting sent out. We're not looking at PDFs that have just a list of numbers.
We're not looking at an Excel download to give us what we need. We have visual analytics that are showing us what's happening. I can look very quickly and say, hey, 2017, we had a major pop here in July and August in our incidents. What's going on?
Which is very interesting too, because our open problems have declined. Right? What's happening here? The idea of communicating through pictures was groundbreaking. I've been using Tableau since version seven point three about a decade ago, and to see just the growth of visual analytics across the industry is incredible.
But it hasn't been without its challenges, especially in the past couple of years, especially if you paid attention to ThoughtSpot's marketing. So ThoughtSpot came in and said, actually, around 2021, dashboards have run their course. They are dead.
Now, why is that?
The main argument that's made, especially in the past couple of years with ThoughtSpot at the front, is that self-service analytics and visual analytics with our nice dashboards leaves a gap and there's not enough that's there. There's something wrong with self-service analytics.
So I don't actually agree one hundred percent with this marketing ploy. I think it's brilliant as a way to generate buzz. But for me, it always reminds me of a quote from Moneyball. So if anybody's familiar with the author, Michael Lewis wrote a book about American baseball where they used analytics to go through and not game the system, but find a way to find value in players that typical contributions or analyses overlooked.
And there's a wonderful film adaptation of it where Jonah Hill plays a character, Peter Brand, based off a real-life person named Paul DePodesta. And he has this quote that I'm going to play around with a little bit where he says there's an epidemic failure in baseball, but in this case, in self-service analytics, to understand what's really happening. So where I see people saying that dashboards are dead, I actually see an imperfect understanding of where analytics and insights come from. So again, apologies to Michael Lewis and to Jonah Hill. But for me, there's something that we don't understand if we're just saying dashboards are dead. There's something that isn't a successful rollout if we're saying, actually, you know what? Let's just try and do dashboards by themselves.
And that's because there are areas where dashboards are very successful.
Got the four of them right here. I'll talk about each of them in turn, so I'll break slide craft a little bit and read them to you. But first off, it's qualitative judgment. If you need to say, yes, this is good, or no, that is bad, a dashboard does that very well.
Right. Pick any metric. Does it go up? Does it go down? Does it stay flat? I don't know. But if you want to capture that judgment call, a dashboard is very successful at doing that and does that great.
You don't want the decision of is something good or bad to be something that every person has to make and know by themselves. A dashboard accelerates the progress in your organization by doing that for you.
Closely related to that is the idea that you can capture your business logic.
I worked for a company where close rates was a very big deal for us. We sent out twelve million pieces of mail a year, and we looked at close rate on everything from a lead when it came in to an appointment to a different variety of lead sources. There were about fifteen different flavors of close rate. And to try and remember all of those was a royal pain. And I was on the data analytics and strategy team to try and sort through all of that was something that we spent hours on.
To expect everybody in our organization to know that and to understand that is just unrealistic. So dashboards capturing business logic makes it so they don't have to learn that. They can just say, what's my close rate percentage? There it is. Great. You can even have a definition by the side if needed.
Dashboards are also really good at helping you measure against targets. You've agreed upon your definitions. You've agreed upon what is good, what is bad. Let's see the progress against that. Whether it's a bullet chart, whether it is a bar graph, whether it is a fun thermometer that you stylized. Who knows? Point is you can see where you're at against the percentage and easily track that.
All of these can roughly get bucketed into this concept of productionalized analysis.
So some very smart analyst has sat down at her computer somewhere, poured blood, sweat and tears into this analysis, gathered data from disparate sources, figured out a way to automate that, and then boom, we can go through and say, here are the insights that we have found and we want to share those out. And a dashboard does that very well.
Now, dashboards struggle when you want to take that next step. So say you see a dashboard that is spread companywide.
Say I want to go in and find some personalized insights to me specifically. So I run our go-to-market efforts for the east half of the US with occasional moonlighting in EMEA and APAC to help with webinars.
And if I want to go and I want to look and say specifically for me what's happening or for my territory, what is happening, unless we have robust role-level security built into my dashboard and the ability to drill through into metrics that I care about, then the dashboard is only going to take me so far. It's only going to answer maybe my first two questions when I want to really ask fifteen questions of the data.
Which gets to the second point. I have built some truly horrendous dashboards in my day, and majority of those were trying to provide infinite flexibility or drill-down. So someone logs in, they want to be able to choose whatever dashboard they're interested in, whatever column on that dashboard they're interested in, and then ask more questions of it. They find a new insight and they say, oh, that's interesting.
Now hold on. Let me look at the underlying data on that, and let me do this drill path. Trying to predict where all of your users want to go is, on my optimistic days, fun and part of the job. On my pessimistic days, an exercise in futility. It's just, it is very difficult to figure out the infinite flexibility that people want.
You can have some wins that are out there, but a dashboard just can't be able to respond to absolutely everything that's needed.
Dashboards also take time to develop.
There's an inherent backlog that's there when you say, oh, thanks for these insights. I appreciate that. Can you just add this really quick?
When you're first starting out, that's fine. But as soon as more and more changes get layered in and more requests come from different people, all of a sudden you end up with a very brittle dashboard where to just add something with air quotes or inverted commas gets you to the point where very quickly you can't actually turn that around as fast. And so you build up that backlog even more. So where they really struggle, if we're looking at again another generalized principle here, is exploratory analytics.
We want people to be able to ask whatever question they want and interrogate the data however they want.
And a dashboard just doesn't do that as well.
So ThoughtSpot has this up on their blog.
I've got the link down there in the bottom if you want to grab it.
Why dashboards don't deliver on promised business value.
I love this as a lived experience because it's something that I run into a lot where a business user gets a dashboard, says I need information about this that's not on the dashboard.
Analyst says good idea. We don't have that metric.
Pass it off to the data engineer who says, well, actually, we need to go get that.
But we can do that for you. Who then goes back to IT and DBA. They do their thing, making sure it's secure, governed, and scalable, which is great. They roll out a new table. Data engineer slaps into their model. Analyst says, here's your answer that you were asking before. The business user says, oh yeah, I sent you a Slack message on that three weeks ago.
Thanks. I appreciate that. But what about this?
The cycle repeats. Cycles are spun. You start spending a lot of time trying to figure things out. People get frustrated as they're trying to work through this in order to get to the data and the insights they need.
So while this is representative of where dashboards struggle, I still think they have a use case. So with apologies to ThoughtSpot marketing, I think dashboards are incomplete in terms of what is needed. And you can see that in our two-by-two right here. Right?
If we're bringing in Tableau, also I mean, we have Power BI users here on this call. Single one of my major clients in the US has both Tableau and Power BI. They are the two monsters in the industry right now. Everybody uses them.
And so if we're looking at building out that content, we have Tableau and Power BI that make it so citizens can build their own dashboards and their visual analytics.
But we haven't yet solved this group on the top right. Even if we throw in Excel.
Right? So if you're looking at trying to build out the ability to consume your own analytics objects, we're not quite there yet.
So take a step back and check in on our friends from earlier, our technical developer who's building out content, our technical analyst who is pulling data and analyzing different aspects of it.
If you look through here, our technical developer is actually doing pretty well sitting here in the bottom left, working with the center of excellence for the organization, whether it's something that was stood up by them internally or something that we at InterWorks helped them develop, they've been able to enable their citizens. The nontechnical users have been able to go through and start to build out their own content, whether that's a Tableau Creator license where they build it out, connect their own data sources, the Tableau Explorer license where they're building out content on Tableau Server, unpublished data sources, or a Power BI Pro license, and they're able to connect out and do things similar to that Tableau Creator license. The point is you have citizen developers, nontechnical people where it's not their day job able to build out visual analytics and analyses.
Our technical analyst, on the other hand, is still getting inundated with requests, sweating here, saying, hey, I really want to get some new data and get some help.
Which does beg the question, though, is this necessarily an issue? There's always going to be people who want to ask for new things.
There's always going to want to be a population of users who want to be able to ask questions of data but don't have access to it.
Is that the majority of the company? Is that a minority of users?
The thing is when we look across our user base at InterWorks, the problem is this is a massively underserved user base in this top right quadrant.
If we're talking from a Tableau licensing perspective, you can think of your Tableau creators being in the top left, people being able to build content, and your Tableau viewers in the top right. So this is your long tail of technical users, people who want to consume analytics, who want to be able to do more than just change a filter.
But instead, all they're given is a dashboard and it's not able to give them enough. They want to explore and work with their data, but the toolset just isn't meeting them there.
And that's not meeting them there, not because of just licensing or what people are purchasing, but it can be functionality. The learning curve for building out a DAX formula is significant.
The learning curve for learning how to drag and drop the difference between dimensions and measures. There's a reason there are classes around that.
So I got a question here. Where do we see the Tableau Explorer? It's a good question. The Explorer tends to live right on the middle of these two. But if I had to choose between them, I would put it in the top left in the object creation bucket.
So the endpoint of a Tableau Explorer license is still to build visual analytics. You're looking at building out a dashboard, not necessarily just finding an answer.
You want to have a productionalized analysis as opposed to going through and just exploring and finding something there despite what the name says.
And so you have a massive underserved user base, which for us, this is actually where we see ThoughtSpot coming in and fitting in very nicely. So if we're looking at what we'll call self-service analytics three point zero at InterWorks, this is our split in our toolset and an abstraction of our reference architecture. We're not getting into all of the underlying data platforms and things that connect to this, but the idea is if we want to enable our citizen developers to be able to do things, if we want them to make productionalized reports and analyses in dashboards, fantastic.
Let them use Tableau. Let them use Power BI. They're there in the top left.
But if you want people in that viewer category to be able to do more than just change a filter and dig in and answer those questions, that's where ThoughtSpot comes in.
Now we can also layer in DBT as well. I don't, we don't have a lot of time to spend on this, but what DBT does is it gives you a bit of a semantic layer with Git integration on your SQL scripts so that you have better lineage into being able to provide where your data is coming from that can then support these tools, both on the object creation and object consumption side.
So you've got your full mix here where different users can use their appropriate tool to get to that point and that holy grail of self-service analytics where not just your professionals can answer data questions, but your citizen developers can also answer those questions.
We're about forty minutes in right now. I get that this is a lot.
This is, in fact, an extended argument for why you need all these different tools, but it can be really tough to abstract or make the jump from this abstraction to a day-to-day environment.
So if you are looking for a quick example of what this looks like, I'm going to show you. So what we need here is an environment where we have both our object creation and our object consumption living side by side.
So I will drop out of the presentation really quick. You'll probably see a quick screenshot of my wife and kids. And what we have here is a website styled after the InterWorks website. It inherits our brand. It inherits our feel of that fun, techy, but also slightly nerdy where we make jokes about Nintendo sixty-four, where we don't even have one on screen.
But we're going through and we're saying, hey, this is an InterWorks product and experience.
And inside of this, we've got links across the bottom here where I actually have three separate dashboards that are all made available to my end user. So, again, the goal here is surfacing content that's available for people to use as they want.
So I can come in and right here this is an airport performance dashboard. I'll click on that link. We're all inside the same branded experience. Got my title that's coming across here, carried over my branding with me, and then I have a familiar Tableau dashboard that I can interact with.
I've got my tooltips. I've got my different start dates. This dataset is outdated, hence the 2003 start date only going through 2016. Sorry about that.
But the point is we have our familiar Tableau experience.
From there, I can come across the top and I can say, actually, I want to go over and look at my IT spend. As I click on that, it's a familiar website-style navigation, and I have a Power BI dashboard that's being brought up.
So we have different preferences, right? Our airline analysts want to build out content in Tableau. IT wants to analyze their spend based off of Power BI. And both of these can be brought together in the same place. And then if I remember, it was a live board or something. Oh, yes, it was InterBurger, which again, going back to this is a fictional burger place that we have at InterWorks.
But the idea here is I can go in and I can look at this to be able to see this is my interactive experience using ThoughtSpot. So at the first pass, what we have here is all three tools being brought together.
We have gone through and done the work on the API side of the house to make it so that you can easily add content and build this yourself. This is a website builder targeted at your nontechnical user. It was built by a bunch of people who love Tableau to make it easy for a Tableau-style user in that top left quadrant, right? Someone who can go through and build that, and then you have the ability to have this web modern experience with your different BI tools.
But I want to take advantage of the fact that we're in here and we're looking at an InterBurger live board here.
So if I scroll down, this looks like a dashboard.
Right? And we said the dashboards are dead or they're incomplete. So what is different about a live board? Well, it's that flexibility that we're saying you can go and you can look. So if I'm coming in and I say, hey, this afternoon snack is doing great.
Comp sales, so comparison against the time period.
Afternoon snack is really doing well. I want to drill down here.
I'm going to go in and I'm going to say, let's actually break this out by business date.
And I can see my trend. I've just got it on a specific month. Let's go in and look at this by week.
So I can right-click and drill in, and we only have a specific week that's in there. I guess we're looking at just Monday for today, so we're only getting the one specific value. But let's drill in further.
Let's see if we have hourly data in this. Okay, we're only dealing with a single day there, so maybe not the best example to work with there.
But the point is I can drill in on a specific day and try and find out what's interesting to me. I can try and figure out what is interesting about this data point, maybe rather than by daypart. So part of the day, maybe there's a trend by my destination.
So as I'm drilling in, is there anything that's driving it? Holy cow, yeah, I've got a one thousand percent increase on my third party. That's an insight that's interesting.
Let me right-click and drill down again. Let's maybe break this out by region. Is there something geographically in the West? Something huge is happening.
So let's explore this some more. I've got my different daypart that's happening, my different destination. I can now add in a filter. Let's just go in and grab that West region.
So if I'm adding region as a filter, I guess I got region leader name right here. I can actually just keep and drill down on this one specifically. Let's only include that West region across the top.
I can then if I wanted to filter anything else, could work with that. I could add in my own columns if I wanted to as well. So let's add in district as an attribute. I'm now seeing it broken across by my different columns that are available.
With this, I can choose to replace that district column with something else in my dataset. You can see different elements that are getting suggested as we work through this. I can build out my own comparisons. Right, here's a suggestion for what we should look at, afternoon versus all. We've got some that are popping up, so we're only in Nebraska, right? We're dealing with a single day of data, and so we're not really seeing a ton.
But the point is, I am now twelve questions deep based off of a live board.
So I started here.
And I was able to drill through and look at all different elements and questions on this. I can then come in and say, okay, by district, here's that Kansas, East Texas, New Jersey. Any of these, I can go in and have that same exploratory experience trying to add or replace or do something else. I can even add in a measure. Right?
As a developer who cut my teeth in the top left quadrant of our two-by-two, the ability to do whatever I wanted to ask, whatever question I wanted as a viewer, for me really is where things clicked.
We've got a public-facing demo. This is my own environment, but if you go to restaurant.curator.com, a mixed platform demo, we've got this spelled out even more explicitly.
So here's a Tableau dashboard executive overview. You can see our comp sales. We can go through and see what's needed. Of course, we get an error as we're looking at it. But if I want this four point three percent comp sales percentage, let me do that comp sales right here. And then we're going to do year-to-date.
It's going to go in. It's going to find that specific value, and I'm going to get that same four point three.
Let's go in and let's do some of the analysis we were doing on the dashboard. Again, a search bar drilling in by region to find different things. So my West region there again, let's look at South Central this time. Drill down.
We'll do that same destination question we were interested in. These are datasets that are the same feeding both my Tableau dashboard and my ThoughtSpot live board.
On the Tableau dashboard, we've extracted it and aggregated it to get better performance similar to importing it into Power BI to make sure that it's performing as you need.
But you get to this point on live board where we can actually drill down if we needed to down to the individual transaction ID.
And so I don't know if I have that. Yeah. So we got just counts or percentages of that. But the point is anything that's available in our data model, we can actually go through and see that specific value in our dataset. So this is public and available for you to go look at. It's something that is very exciting for seeing how these can report off of the same data and really lead to what we're talking about with that self-service three point zero, where we're enabling lots of different users to do that.
Okay. We're about fifty minutes into a presentation now. And I understand people are going to ask, how was the webinar?
And you'll be like, man, this fast-talking American went off for a long time about this.
So if I want to explain this concept in five minutes, for me, my favorite analogy is actually Yogurt Berry.
Now Pinkfrog is their affiliate or at least their competitor in the US, but Yogurt Berry and frozen yogurt actually is a surprisingly insightful analysis and analogy when you're trying to explain self-service analytics. You walk into a shop, you pick up your bowl, and you want to say, I'm going to go in, I'm going to get this specific flavor.
I'll go over to the chocolate flavor. I'll pull the handle. Chocolate is what comes out. I'm not surprised by the data that I'm getting serviced from.
I'm not surprised when I go over to the pomegranate section and I pull that lever. It's exactly what I asked for. This is the same concept as things like your published data sources or a verified data model that you've worked with IT on. There's no surprises on what's coming out. IT has done all the work behind the scenes to make sure you have your environment that's ready to go. You can then combine that as you will, make a delicious flavor combo, and then go in and pick from different buckets that spice it up, add some toppings.
So I've gone in. I pulled my sales fact table.
I then want to combine that with some marketing data and combine that with some HR data to see if there is something correlated between marketing response rates and how we react to that based off of my sales manager. Tip in a scoop of M&M's, tip in a scoop of crushed-up Heath bars, and you have yourself a nice treat that you're able to do that.
You're never surprised by what's out there, but you're also not complaining that Yogurt Berry is saying, hey, sorry, you're not allowed to bring in your own toppings. We're at a scenario where there has been time and effort put in to make sure that our model is correct ahead of time, that we're working with governed, scalable data.
Because if I want to go and create my own sundae, I could go to the store, I could buy all of my own ingredients and toppings, and I can mix them together and make something.
And that will work just fine for me.
But if I want to share it out with twenty of my friends with full flexibility on what they want, I'm going to take them to a Yogurt Berry.
I'm not going to take them to my kitchen to try and make that happen.
So I don't know if that lands as well for you as it does for me, but a couple of links here at the bottom if you want to go and look deeper into a larger write-up on that. But when we're talking meeting people where they are with self-service BI three point zero, this for us, this two-by-two, the split between professional and citizen, object creation and consumption leading to the appropriate tool for the appropriate user helps make that happen. And the analogy you can play with, if you'd like, is that frozen yogurt stand. So I really just want to say thank you for joining us.
Thank you for giving up an hour of your morning to listen to me ramble on about this. We have dedicated resources in APAC, so you don't have to worry about, again, the strong American accent on that. You're welcome to scan the QR code as you'd like, and I'll throw up a quick poll on here as well just for us to close out. But then while we're waiting here for the last five minutes, please feel free to drop any questions in chat or in the Q&A.
Again, thank you so much for joining us. Really appreciate it.