I talked to a lot of people and I feel like there’s been a number of waves of analytics.
So said Ryan Atay, CEO of Tableau, last week prior to the data analytics firm’s annual gathering in San Diego. He expanded on his point:
The first wave is really all about in the past where you had to say, ‘OK, I need a report, I need a dashboard, I need an insight’. I have to make a phone call. I have to walk down the hall. I have to send an email and I send it to someone. That person maybe knows how to code. They have to do some things, access a data warehouse. They generate a report and that report may be useful, but it may not be useful. It actually may be something that I have another question on and I cannot click on it because it’s static. So then I make a phone call. I walk down the hall again, I send an email and that process as you can imagine, it comes perhaps a little bit frustrating for the end user.
The second wave was when the Self Service analytics revolution started to happen. Users wanted to click, they wanted to ask questions of their data. They wanted to drag and drop. They wanted to do all these things that ultimately Tableau helped to lift and start this whole revolution. And they wanted to help everyone see and understand their data.
The third wave, the major focus of last week’s conference, is about about bringing data and AI for everyone, to, as Atay put it, “everyday users like myself”. Tableau CMO Elizabeth Maxon picked up this thesis, arguing:
There is a huge demand in AI. Think about tools like ChatGPT that we all have been using for the last year-and-a-half. I use ChatGPT every day, whether it’s helping me generate my workouts, the gym, or taking-whatever ingredients I have left over in my fridge to come up with a recipe to feed my three kids.
Now as we start to adopt this type of technology in our personal life, we then expect it in our workplace. We surveyed over 10,000 IT leaders and 81% of them feel that urgency to implement AI across your entire organization. But there’s a snag – 59% of those leaders also admit to lacking a data strategy. You have to have a data strategy in order to be successful with AI. But it’s more than having a data strategy.
You also need to have a data culture. Now a data culture is more than just having data as a buzzword. This is where you’re fostering an environment from the bottom up and the top down where everyone across the organization they understand the value of data. They know how to interpret it and then they take action on it.
In practice
That’s the theory and the objective. But how is all this panning out in practice? As diginomica always maintains, the best proof point for any tech is testimony from the customer frontline and there were plenty of examples on show at Tableau Conference of how to build that kind of data culture to which Maxon alluded.
Pharma giant Merck KGaA has been a Tableau user since 2019. Walid Mehanna, Chief Data & AI Officer, observed that the federated nature of the organization means that the data and AI strategy is equally federated:
Our strategy is to be as close as possible to our patients, to our customers, to our partners on the ground. And that means that everything we do needs to be efficient, it needs to be scalable, and it needs to be very close to the customer…Nothing is black and white. It’s a lot of shades of grey and not every organization or not any organization can be fully in wave three or one or two. It will always have pockets of excellence that are further along than the others.
The trick here is to look for champions and early adopters, advised Mehanna:
You look for the ones who are hungry for data, hungry for insights. Those are your best allies to get started. And then you take it from there and you get people excited about it and you start to build a family.
Merck’s approach to analytics is heavily user-centric, added Mehanna, and focused on meeting user needs:
Dashboards are great for an overview, they are great as a control center, but executives are busy. What they want is, what do they need to do and ‘What needs my attention now. What do I need to do next? What do I need to look for? ‘And that’s exactly what our policy is.
That means a variant on ‘eating your own dog food’ for Mehanna’s team:
We have to drink our own champagne. Call it whatever you want, but if we aren’t data-driven, we cannot authentically tell the organization to become data-driven. This is why I want every launch, every app, every [bit of] data that we can get from our own data, managed exactly like the same.
Banking on analytics
Kayne Johnson, Director of Data and Analytics at Arvest Bank, recently listed in Forbes 2024 World’s Best Banks for the sixth consecutive year, reckoned that his organization has been fortunate in how it has adopted Tableau:
We started with a lot of people who really did want to use the data. The problem was there was a lot of it. There’s a ton of data. It all needed to be viewed in the exact same place at the same time, and also when [users] needed it and also had to be accurate. What we really wanted in our goal was really to provide that data and those insights in the most accessible way as we could and most effectively possible.
The reason for choosing Tableau was to take a whole new approach, he recalled. The way to kickstart that was to have “one big win” to make the case:
We actually identified one of our reports that we had out there that was extremely valuable, extremely high use, but it was also tremendously manual to put together. We went back and looked at the calculations. There were 113 different calculations that needed to be updated in this report every single month. That’s insane. And for us, it was also 113 different ways it could go wrong.
While this use case was identified as being suitable to show value and demonstrate the direction the data strategy should to go in, there was one big problem, as Johnson explained:
We knew we couldn’t just turn this one off and then turn on another one. We actually gathered several different users of this report across the bank and gathered up a workgroup. With their insights and the feedback they provided, we identified all the different KPIs that we knew we couldn’t live without as we moved forward into this new world.
That kind of feedback was to prove very useful as the bank moved ahead, he added:
As we developed, we kept them along with us the whole way. When we started showcasing the report, we looked at how to build that really strong feedback loop. Part of that was really to incorporate things in the dashboard, things like customized buttons that whenever they would click on the button, it would generate an auto-draft email that was directed to us, where they could just fill up the feedback.
We had hovers throughout the report that would walk them through some of the visuals. We also suggested when you’re in the report, when you’re looking at it, be sure to put those comments in there – tag us! We get that email as soon as you tag us and while it’s fresh in your mind, we can address it.
Keeping people engaged was important, he noted:
We didn’t want anyone leaving anything on the table. It could potentially be valuable. One of the things that actually got us feedback was someone suggested just to add branch names to the report. It kind of surprised us that we hadn’t thought about this before. We had always referenced these branches as their number and we never really stopped to think that nobody else in the bank did that. That one little piece of feedback, that honestly was a five minute update in Tableau, really did open up the accessibility for the report and really got everyone on board. We’re all speaking the same language. We’re looking at the data the same way.
Having demonstrated value in that initial use case, there was now the issue of how to scale. Johnson said:
We actually had another team at the bank that put together a user group [where] we gathered up all the different Tableau users in the bank, and we would talk through problems, challenges, insights. We also would periodically bring in leaders of the bank and they would actually pull up their screen and they would walk through how they were using our reports. That was fascinating from both the standpoint of understanding what they found to be valuable and also what they didn’t use at all.
As to whether the AI-enabled third wave is going to have impact, Johnson concluded:
I think we’re really excited and I hope I can speak for my team and saying these advances in AI are really going to help us shorten that distance between the question and the answer.
