Part 2: The Slow Erosion of Product Intuition

In Part 1, I made the case that Tableau Next feels like a product being pulled towards Salesforce's untapped customer base rather than built for the people who already use Tableau every day. But that raises an obvious follow-up: how does that happen?

Before I get into it, I want to say something clearly. None of what follows is aimed at the people inside Tableau who design, develop, product-manage, build, ship, market, and sell this product every day.

I've seen how hard they work. I've seen the level of care and detail that goes into how they think about features. I’ve personally learnt a ton from them, and I’m certain that without their interactions, I wouldn’t ever be able to cover the product the way I do. I’ve seen how passionately they advocate for the product and how, within the parameters of what they control, they try to make the best of the challenges they face. It's something I've witnessed firsthand as a Tableau Visionary, and it's something I have a huge amount of respect for.

The challenge I'm describing here is about the structural decisions being made above them. Product leadership. Brand positioning. Strategic direction. These are choices that sit well above these teams, and I think most people in the community can sense where the responsibility really lies.

Part of the reason I'm writing this publicly, rather than keeping it to the internal conversations I'd normally have, is because I think the community already feels what I'm trying to put into words. These aren't new frustrations. But articulating them openly gives us something to rally around and hopefully a surface for other perspectives to push against and challenge my thinking.

Right, as I like to say, let’s get stuck in … that feels weird written down!

There's a phrase I've been using a lot lately: product intuition. It's different from product vision. Vision is a roadmap. It's where you say you're going. Intuition is something deeper. It's the instinct for what the product should feel like. What it should say no to. What it should protect. What trade-offs are worth making, and which ones betray the thing that made the product good in the first place? These should be non-negotiable to any brand.

Intuition comes from immersion. From living inside the problem. From feeling the same pain your users feel, day after day, until you develop an instinct for what matters and what doesn't. The team that built Tableau originally had that. They understood what it felt like to wrestle with data, to want to see something and not be able to get there, and they built a tool that solved those problems and eliminated those feelings.

When a company builds a product from scratch, that intuition develops naturally. When a company acquires a product, that intuition doesn't transfer. You get the IP, the code, the brand, the customers, the P&L. But you don't get the instinct. And without it, you default to the next best thing: the numbers, the data!

Mike Swanson wrote a piece recently called Backseat Software that I keep coming back to. He makes the point that once you can measure everything, a strong product vision becomes optional. Not because anyone argues against vision, but because you don't strictly need it anymore. Backing a chart is safer than backing an opinion. If a decision goes wrong, you can point to the data and say "we followed the evidence." From there, you have a session, make a deck called lessons learned, and you go again. Over time, this changes the role of a product team. Judgement gives way to iteration. Taste gives way to performance. The product still evolves, but without a clear sense of direction. Only a sense of momentum.

This is, I think, exactly the dynamic at play with Tableau since the acquisition. The investment hasn't stopped. The features haven't stopped. But the decisions feel like they're being made about Tableau rather than from within it. They're informed by data, by market analysis, by what is usual Salesforce inward logic, I described in Part 1. Everything is defensible. But defensible isn't the same as inspired.

And here's the thing about building intuition from numbers and trends rather than from immersion in the field: you arrive at something that looks right on the surface. The features make sense on a slide. They work right in a Figma mockup or demo, customers say, “yeah, that solves the problem”. The roadmap ticks the boxes. But when you actually sit down and use it, something's not quite right. There's a gap between what the product does and what it feels like it should do. This is exactly what I feel using Tableau Next. I can never explain it because, as I said, everything about it is defensible. But that gap is the absence of intuition. It’s not an intuitive product to use. It's the difference between a product shaped by people who understand the telemetry and diagnostics of being an analyst and a product shaped by people who understand the pain of doing the job.

If you want a concrete example of what this looks like in practice, look at what has been shipping recently. There's been a trend over the past year or so of Tableau delivering features that the community has been requesting for a long, long time. Rounded corners, font support, Colour Palettes, and many more. Long-requested quality-of-life improvements. Things that have sat on the cutting room floor for years. And the messaging around them has been clear: we're listening. We hear you. We're shipping the things you've been asking for. On the surface, that sounds like progress. Now, I’ve been really reluctant to admit or say what I’m about to say out loud because I’ve also been known to champion quality-of-life improvements. But it actually all worries me. Here's why.

If I look back at Tableau's history, some of the best product decisions were those in which they chose not to build and ship popular requests. As much as it frustrates me to say I still don’t have a proper layout system in Tableau, the reality is that when Tableau chose to prioritise other things over an example like this, they were generally making the right call. The features they shipped in place of the popular requests tended to be bigger, more foundational, more forward-looking. They were solving problems we didn't yet fully know we had. That takes intuition.

So when I see the product team going back to the backlog and clearing out the long-standing requests, I don't read that as "we're finally listening." I read it as: the intuition about what the big problems are has gone quiet. The ability to look at the community's feature requests and say "actually, this other thing matters more" requires a deep sense of where the product needs to go. Without that sense, the safest move is to go back to the list and ship what people asked for.

It's not a bad strategy. These are features people genuinely want. But it's a reactive strategy. It's the absence of a proactive one. And in a market that's moving as fast as this one, reactive isn't enough. These aren’t features competitive platforms are marketing either.

In Part 1, I used a Venn diagram to illustrate what Salesforce sees as the opportunity: its own customer base, with Tableau as the vehicle to sell analytics inward. But zoom out for a moment.

Draw a bigger circle around both Salesforce and Tableau. Call it the data market. Now look at how small those two circles are inside it. That's the real picture. The total addressable opportunity in analytics, data infrastructure, and AI-driven insight dwarfs the Salesforce customer base. It's not even close.

Salesforce is using Tableau to look inward. To sell into its own base. To keep data inside its own ecosystem. But here's the thing: I'm not even sure that inward play works on its own terms. Salesforce customers who buy heavily into the CRM don't think of Salesforce as a data brand. It's not the place they go to store, process, and visualise their data. That's just not the relationship they have, and I don't think it's one they can easily buy into. The teams are entirely different. Most organisations aren't structured with their marketing and sales teams sitting within the same function as their data, IT, AI, and innovation teams. These are separate worlds with separate budgets, separate priorities, and separate buying decisions. Trying to bridge that gap by slapping the Tableau brand onto a CRM-integrated analytics experience doesn't change the underlying reality of how these organisations actually work.

So the inward play is limited at best. But the outward opportunity is enormous. Tableau is one of the most recognised names in the entire data industry, not just in the Salesforce corner of it. The opportunity is to point the brand outward and go after the space that's growing fastest: composable analytics, warehouse-native workflows, AI-ready infrastructure, headless capabilities. The tools winning in that space are tiny compared to Tableau's underlying capability. The difference is that they have direction. They know where they're going.

But I'd go even further than that. What if instead of just competing in the enterprise data market, Salesforce took the Tableau brand and pushed it somewhere it's never properly been: into everyday life?

This isn't a new idea for Tableau, by the way. Years ago, they built an app called Vizable. You have to search hard to find much about it now, but the concept was simple and ahead of its time. Everyone who had it (I still do) on their iPad could navigate data sources and understand things intuitively. It made data exploration a natural, tactile experience. Since then, Tableau has never truly committed to making data visualisation accessible to everyone, not just analysts. The aim here is to make Tableau as synonymous with visualising data as Excel is with spreadsheets. On your phone. On your tablet. In the browser. In the real world. Wherever data needs to be understood, Tableau becomes the conduit for understanding it. That's a 20-year vision. That's the kind of ambition the brand is capable of. And that's the kind of direction that would energise the community, attract developers, and give Tableau a purpose that goes far beyond being a line item on the Salesforce invoice.

Tableau has everything it needs to get there. It just needs a vision that matches the size of the opportunity, and leadership with the conviction to pursue it.

The tools that will define the next decade of analytics won't be the ones that followed the evidence most diligently. They'll be the ones where someone had the conviction to make a call, but the data couldn’t justify it yet. In Part 1, I put forward an idea around a headless BI SKU. That was my attempt to do something I wish Salesforce could do more of with Tableau. No usage metric is going to tell you to unbundle your own product. That's an intuition call. It requires someone to look at what Tableau already has, see what it could become, and have the nerve to prioritise that over what the spreadsheet says.

Tableau was built by people who lived the problem, not people who measured it. They knew what it felt like to need to see data clearly and couldn't, and they built something that solved for that feeling before anyone could articulate the demand for it. And here's the irony: Tableau's motto has always been about helping people see and understand data. But the people who made Tableau great didn't get there by seeing and understanding data about their users. They got there by being them. That instinct is the most valuable thing Tableau ever had. And that's what I fear is being lost.

The question is whether anyone at Salesforce is brave enough to make the same calls, or whether the product will continue to be shaped by the numbers, the backlog, and the same inward philosophy. I hope it's the former. The community is ready for it, I’m here for it, but Tableau Next is not it!

Next week, lets finally take on this idea of “Apps”. Sigma’s been pushing it and has almost entirely pivoted to grounding it’s brand on this term so I want to spend a bit of time dissecting what exactly is the difference between an Apps, Dashboards and reporting in the analytics context.