The Tableau Conference 2019 review - we're taking on data prep, AI-for-BI, and we're talking Salesforce (finally)

Profile picture for user jreed By Jon Reed November 15, 2019
Summary:
This year's Tableau Conference shook up my thinking and changed my view of the company. Surprisingly, it had little to do with Salesforce - though I got an earful on that too. Here's my top conference takeaways - and two burning themes I pursued throughout the show.

Tableau Iron Viz event
Attendee gives views on Iron Viz dashboards

Heading into Tableau Conference 2019, it looked like the Salesforce acquisition PR muzzles would be on tight. But lo and behold, last Tuesday, November 5, the biggest barrier to forward conversations, the UK CMA review, was lifted.

That meant Tableau and Salesforce executives could begin formal conversations with each other on their hopes, dreams - and roadmaps.

But with Tableau Conference 2019 this week and the Dreamforce insanity, err, celebration next week, the bulk of those conversations were a mad scramble to cross-pollinate the two shows with executives.

That much was accomplished, with Tableau on deck for an official presence as next week's Dreamforce. And, this week, we had some Salesforce execs on site. Salesforce's President and CPO Bret Taylor was on hand to answer questions from analysts:

Alas, there hasn't been enough time for the execs to offer much more than a few soothing soundbites, but I did score a few of those.

Salesforce aside, this year's Tableau Conference was the classic vibe, but with significant new twists. What's the classic vibe, you ask? That's when 17,000+ customers and partners, aka "data geeks," lose their collective minds:

There are two potent themes I was pre-occupied with this year:

1. As companies outperform others with pervasive analytics, it puts pressure on those who don't. It also fuels a mission on Tableau's part to help companies execute better on their data strategies. After all, if you're using the same BI tool (Tableau) and not getting the results of the top performers, don't you want to know why? Tableau sure does.

Last year, in my state of Tableau 2018 piece, I pressed Tableau CPO Francois Ajenstat on why, with all this hyperbolic emphasis on being data-driven, we aren't seeing more quantifiable results from high performers. That's changing.

During an analyst presentation, Tableau's Mark Jewett cited a McKinsey study on customer analytics that found ompanies with intensive use of customer analytics are 23 times more likely to outperform their competitors in new customer acquisition, and nine times more likely to surpass them in customer loyalty. 

That led me to other McKinsey analytics studies that went beyond customer analytics to look at the impact of pervasive analytics. In McKinsey's 2018 analytics comes of age report, referring to high performing companies (the fastest growing companies in this study), McKinsey wrote: 

Respondents at high performers also see a top-line benefit: they are three times more likely than others to say their [data] monetization efforts contribute more than 20 percent to company revenues.

A third McKinsey analytics study on "breakaway analytics" (yes, they like to do analytics studies), from May 2018, beefed up these findings with a bold thesis:

A handful of the world’s companies have cracked the code on embedding analytics into every layer of their organizations.

At the Tableau analyst session, Jewett said that a range of studies have reinforced this message:

They all essentially point to the same truth. When you look at things like, not just revenue growth, but profitability, Net Promoter Score (NPS), new customer adds - there is a very strong correlation between companies that are doing better with their data, and companies that are performing better with those metrics.

However, only 8 percent of the 1,000 companies McKinsey surveyed in their breakaway analytics piece had achieved these "analytics at scale" results. For McKinsey, the good news of documented results also points to reams of companies that are being left behind. They might have analytics, but it's not pervasive. It's not in the cultural fiber. Why?

Tableau acknowledged this data performance gap, and their dedication to helping customers with it. Most of Tableau's forward strategy hits on this in some way, but I'll call attention to these:

  • Helping companies to build a deeper/more passionate data culture with better tooling, ease of querying for more users (as in the Ask Data solution, better data prep support, and "AI in your BI" capabilities to make the data science push viable).
  • Documenting and sharing the characteristics of high-performing Tableau customers. Tableau has identified repeatable approaches across the most engaged/successful Tableau accounts - so customers don't need to wing it when it comes to spreading a modern BI message.

2. Companies have made progress towards a single source of truth on transactional data, but: AI/ML and external data raises the bar. You can't make better decisions (or have proper AI) without that data. Tableau (and its partners) are in push mode to help customers with this still-formidable analytics obstacle.

This simple marketecture slide from Tableau speaks volumes:

Tableau marketecture

Even a few years ago, this type of slide wouldn't have made much sense. Now, you could write an article on each item on this slide. Examples:

  • Customer enthusiasm for embedding Tableau is strong (1/3 to 1/2 of customers are asking Tableau about it, as per our briefing).
  • Cloud: Tableau is on a multi-cloud push, with a third of Tableau's total deployments on the public cloud.
  • 15,000 customers are on Tableau Online, a "fully SaaS offering."

That multi-cloud push makes for interesting co-optition variations:

But it's the top areas in the slide: self-service analytics, AI-powered, and self-service data management that show what the "new Tableau" is aiming for. Tableau hasn't said this to me, but here's how I look at it: The rise of Tableau was about eradicating the BI bottleneck of the IT department, bogged down with legacy BI solutions.

I believe this new phase of Tableau is about eliminating the data analyst bottleneck, and empowering business users to ask their own questions and even help to prep their own data. Data analysts might resent being referred to as a bottleneck, especially after all they've done to liberate BI from the IT quagmire. But truth is truth. However: armed with modern BI tools, and an attitude of empowering business users, analysts too can transform.

Tableau customer Healthfirst captured that in their conference presentation, saying that "my BI role becomes more like a management consultant and data evangelist." Meanwhile, with next-gen tools like Ask Data, which enables more users to ask questions of data than ever before using NLP, Tableau is clearly expanding its sights well beyond their core analysts, admins, and power users - with no intention of leaving them behind.

Tableau Conference 2019 - three monster takeaways

1. Tableau gets serious about data management:

A range of announcements and product enhancements support. A good place to start is this October 2019 data management press release, which introduces Tableau Catalog.

The obvious/awkward question is: how do Tableau's data management partners feel about this? Were there any swear words uttered behind the scenes? Not that I can tell. I tried to get the Alteryx team on the ground at Tableau Conference to air our their concerns, especially given they have their own Catalog product, but nope - no dirt here. Alteryx sees only the benefits of Tableau legitimizing the data preparation/quality space, and plenty of work for all. Similar to this exchange:

My response:

2. AI is getting real - with the help of Tableau's partners. Last year, Tableau presented a well-articulated view of AI on the keynote stage. The question is: what's next? This year, I saw more customers getting on with it, making real progress embedding AI in their processes - no, not comprehensively, but not as pilots either. Healthfirst shared this slide with more than twenty AI models they have in production now (apologies for the angled view):

Healthfirst AI models

Healthfirst accomplished this with their partner DataRobot. DataRobot is one of the partners that are enabling Tableau customers to move ahead with data integration, modeling, and machine learning projects to consume in Tableau (they call it "automated machine learning"). I also spoke with Qubole, another Tableau partner doing an impressive job of helping customers manage and analyze diverse data sources at exabyte scale.

I plan to do a whole piece on Healthfirst. It was their team that talked about "AI for BI," and the notion that putting your AI in your BI is an effective way to move forward. After the presentation, DataRobot's Adam Weinstein said something that stuck with me: 

When you look at a lot of the industries where we've been fortunate enough to be successful in, like health care or financial services, they're heavily regulated. If you can't explain your AI models, you're going to get crucified by a regulator. One of the things we've really tried to bake into the product is this ability to explain what a model is doing at the lowest level. So whether that model is a linear regression or a neural network, you have to be able to explain how that decision is getting arrived at. To make sure there's no bias, or [other ways] your model is going afoul.

And that's why an expert AI partner that knows your industry matters. But Tableau isn't outsourcing AI to partners - they have their own investments also. Look no further than AI-powered Explain Data: "get explanations for the "why" behind a data point with a single click."

3. Salesforce looms - and so does Dreamforce.  As I mentioned, it wasn't easy to go beyond feel-good acquisition sound bites. I typically have little patience for feel-good-isms, but in this case, I sympathize. Getting grilled by media and analysts when you haven't had adequate time to get specific with your acquisition partner is not a situation either side can enjoy. However, I did gather some nuggets, most of which I will save for a Tableau-Dreamforce preview piece Monday.

There was some interesting debate amongst analysts on just how much overlap there is between Einstein, Salesforce's AI/analytics platform, and Tableau. One analyst claimed 80 percent overlap, but I personally don't see that. If anything, Salesforce's AI investments are likely to fuel Tableau's efforts. Einstein is focused on customer analytics; Tableau has broader aims. In our one-on-one session, Ajenstat didn't try to sugarcoat the overlap aspect:

There is some overlap, and it's okay. But what you have to step back is to think: there's lots of use cases that are out there. What they've done really well with Einstein Analytics is solved a really, really nice use case around embedded analytics in CRM. And, for that, they do really well. They've been very focused on CRM, obviously, for obvious reasons. But they don't do the other. And, this is what we do really, really well.

And that folks, is as close as I got to detail on the acquisition; roadmap discussions were out of the question. In my Dreamforce Tableau piece, I'll tee up a few of my own views, but I'm not as grouchy about this deal as I am most enterprise acquisitions. Stay tuned...

Updated November 15, 5pm eastern time, with updated stats/links from McKinsey and clarifications on the stats Tableau cited on McKinsey's analytics studies at the conference.

 

Image credit - Photo of Tableau Conference 2019 Iron Viz event by Jon Reed. Mediocre slide photos by Jon Reed.

Disclosure - Tableau paid the bulk of my travel expenses to attend Tableau Conference 2019. Salesforce is a diginomica premier partner.