7 Data Strategy Mistakes Marketers Must Avoid in 2026

Marketing teams have never been better equipped to work with data. Automation is everywhere. Tooling is more sophisticated than ever. And yet, many organisations still struggle to turn data into decisions they can actually stand behind.

The stakes are already clear to marketing leaders. Recent research has found that 30% of CMOs say improving data quality is the single biggest lever they can pull to improve marketing performance. That focus reflects a growing recognition that better decisions don’t come from more tools or more data, but from data that can actually be trusted and used with confidence.

To understand where data strategies most often break down, we asked marketing leaders, data practitioners, and RevOps experts from across Season Three of The Undiscovered Metric one simple question: what mistakes do marketers keep making when building a data strategy?

Here are the most common data strategy pitfalls that surfaced again and again, and what to do instead. Read on for highlights or watch the compilation video below.

1. Trying to do everything at once

One of the fastest ways to stall a data strategy is trying to solve every problem in one go.

Jonathan Sweeney, Senior Director of Data at Gain Theory, has seen teams aim for speed, scale, and sophistication all at once, “trying to get the speed, the capability, everything being goal encompassed all in one go,” instead of focusing on clear initial goals and the accuracy of the data they already have. 

The issue is that this approach delays insight while teams wait for a perfect setup that never quite arrives. As he puts it, “trying to do it all in one go will mean that you'll be waiting for a very long time for that absolutely perfect insight.”

“The idea of having everything perfect the first time, I would say, is a way to ensure you get no results.”

Jonathan Sweeney, Senior Director of Data at Gain Theory

His advice is pragmatic. Early projects won’t deliver flawless insight, but the first one will give you direction.

What to do instead:

Start with a specific question, build insights iteratively. Some results are better than no results. Progress comes from answering clear questions, not from waiting for a perfect model.

2. Treating marketing data as a marketing-only problem

Another common mistake is designing a data strategy entirely within the marketing function.

Laura Erdem argues that this approach breaks down almost immediately. “I would argue that marketers should not be the only one setting up a data strategy for marketing,” she explains, “a data strategy for marketing has to go across departments.” When marketing works in isolation, critical context is lost. 

That marketing-centric lens shows up later in the funnel. Diana Gonzalez points out that teams often celebrate early-stage performance without understanding downstream impact. “A great number of leads this quarter is great, but how is that making an impact for the company?” Without visibility beyond marketing metrics, Diana argues false positives are easy to mistake for progress.

What to do instead:

Build data strategies around the full customer journey, involving sales, product, and revenue teams from the start.

3. Over-customizing too early in the data lifecycle

Customization is often framed as a sign of sophistication. But Vincent Spruyt, Global Chief Product Officer at KINESSO (IPG Mediabrands) warns that spinning up bespoke logic can quietly undermine a data strategy by making workflows rigid and brittle. This is especially true in agencies and enterprise brands where the standardization across data workflows allows marketers to compare metrics across regions, brands and channels easily .

“In the marketing and agency world it’s a bit insane in the way we customize for every single client. Over customization is not to the benefit of a client,” says Vincent. He explains how the disconnect happens early in the data strategy: “The client thinks they want customization, but what they're actually asking for is personalized outcomes.” 

“They don't want us to reforge the hammer each time.”

Vincent Spruyt, Global Chief Product Officer at KINESSO (IPG Mediabrands)

When marketers interpret that as a need to rebuild everything from scratch, Vincent explains, “it hinders that foundational layer of standardization to really leverage AI. You can only do that if you have kind of a standard way of working.”

What to do instead:

Standardize your data foundations and workflows. Use them to deliver tailored outcomes, not bespoke plumbing.

4. Misreading ROI and attribution signals

Several guests highlighted how easy it is to misinterpret performance metrics once they appear in a dashboard.

Sarah Mansfield encourages marketers to question impressive-looking results rather than celebrating them. “I've seen some big ROI numbers here,” she says, and the first response should be to ask, “is it too good to be true?” In her experience, “quite often I say, if it is a really big ROI number, it is too big to be true.”

Attribution models play a major role here. David von Hilchen points to last-click ROAS as a common culprit, describing it as “overvaluing channels that are able to put an ad impression close to a conversion event.” Without considering other touchpoints, teams risk drawing the wrong conclusions about what’s actually driving performance.

What to do instead:

Pressure-test ROI, look beyond last-click models, and always ask what else might have influenced the outcome.

5. Neglecting shared definitions

At the foundation of many broken data strategies is a surprisingly basic issue: teams don’t agree on what their metrics mean, or why they’re collecting them.

Amar Vyas sees strategies unravel because “there's no common definition.” Even simple questions like “what's an impression?” or “what's an engagement?” cause incorrect values, making it impossible to align reporting with business goals or make confident decisions.

What to do instead:

Agree on clear, shared definitions before scaling reporting or analysis. Consistency in meaning is what allows insight to travel across teams and channels.

6. Collecting too much data without a clear purpose

One of the persistent myths in marketing data is that more data automatically leads to better insight. In practice, unfiltered data collection often slows teams down, making it harder to validate accuracy or identify which signals actually matter.

Jonathan Sweeney sees this when teams prioritize volume over intent. “Going out and sourcing every possible data source you can get can leave you in a situation where you have sort of noise and you have volume that you just can't practically deal with,” he explains.

That lack of intent is often rooted in unclear objectives. As Madalina Teodorescu explains, “If you don't set your objectives top bottom, then your objectives will not speak the business objectives. And at some point a disconnect will be created.” When goals aren’t clearly defined, teams default to collecting more data instead of the right data.

What to do instead:

Define the question first, then work backwards to the minimum data required to answer it. Purpose-driven data reduces noise and speeds up insight.

7. Using data without market context 

Marketing data strategies often fail because teams rush to execution and optimization before they fully understand the market they’re operating in or the role their brand plays within it.

Simme Volkers sees this as a sequencing problem. “It's the order in which campaigns are being executed. The marketers put the priorities on executing the campaigns, but I think it's really important to first think about your data and your business case, and invest a lot of time into research and validation. This way you will get to know the market upfront and correct your strategy along the way based on real time data instead of just looking at the data at the end of the process.”

That lack of market grounding becomes even more damaging when data is used to guide strategic decisions. Barry Labov shared an example of a company that relied on competitive data to inform pricing. “Based on their data, they decided they would lower their price because there were too many competitors out there.” The problem wasn’t the analysis, but what it ignored. “The only issue was they had an amazing breakthrough technology that they included in it.” By optimising based on surface-level signals, “they actually hid the uniqueness.”

What to do instead:

Use data first to understand the market and validate strategic choices, then optimize execution. Research and differentiation should guide how data is interpreted, not be casualties of optimization.

Final thoughts

Heading into 2026, automation and AI will continue to accelerate decision-making. Subsequently, the consequences of weak data strategy become harder to ignore. 

The marketers getting the most value from data share a common approach. They are deliberate about what they measure, clear on how metrics are defined, and thoughtful about the order in which work gets done. They use data to understand the market before optimizing activity and make space for insight to develop over time.