Executive Summary
Marketing often feels the data team is a bottleneck. The data team, in turn, finds marketing data chaotic. This makes it hard to get clear answers on campaign performance.
This friction means you might be making decisions with data you can't fully trust, which can lead to wasted ad spend. It also makes it harder to have confident budget conversations with the CFO.
The issue isn't usually the people, it's the process. A good fix is to introduce a semantic layer. It acts as a translator between the fast-moving marketing team and the business's need for reliable, consistent numbers.
It's a common story. The marketing team feels the data team is a bottleneck, and the data team thinks marketing's data is a bit of a mess. In my experience, both have a point. This isn't down to a clash of personalities. It's a structural problem I see in most of the scale-ups I work with. Marketing has to move quickly, while the data team is responsible for building something stable and trustworthy. The two approaches can often pull in opposite directions, which tends to slow things down for everyone.
Why it's hard to get fast, accurate campaign data
A typical situation looks like this. You need to know the ROAS of a campaign that went live two days ago. The Facebook Ads dashboard shows a promising number, but when you ask the data team to confirm it, they say it will take three days. The reason? A small typo in the UTM tags, a campaign name that doesn't follow the usual convention, and the fact that conversion data from the platform will never quite match the backend transactional data.
From your perspective, this can feel obstructive. From theirs, it's just necessary diligence. Even with a modern data stack, you might find you're just moving the same problems around more quickly. If the underlying process is a bit broken, automating it just creates unreliable data at a faster rate. This back-and-forth slowly wears away at the most important thing you have: Data Trust. Without it, every report becomes just another opinion.
How inconsistent data affects the data team and the business
What the data team might not say out loud is that when campaign data is inconsistent, it can do more than just create a query for them. A campaign named in an unexpected way can sometimes cause their data transformation models, the ones they spend weeks building, to fail. This puts them in a difficult position with their boss, and with the CFO, who tends to be the final judge of what numbers are correct.
I've been in plenty of meetings where a marketing lead presents a 4.5x ROAS from a platform dashboard, only for the CFO to show a completely different figure. The conversation then shifts to 'whose number is right?', and it's hard to make any real progress. The issue isn't the dashboard itself. It's that the business logic hasn't been written down and agreed upon. There isn't a Single Source of Truth for the most important metrics.
Using a semantic layer as a translation tool
In my experience, the best way to improve the relationship is to look at the structure of the data, not the people. The answer isn't usually more meetings, or even another tool. It's often about introducing a Semantic Layer. You can think of it as a central translation dictionary for the business. It takes the raw data from your marketing platforms, which can sometimes be a bit chaotic, and applies a consistent set of business rules to it.
For example:
This isn't about creating heavy, bureaucratic Data Governance. It’s about light, automated checks that provide guardrails, not gates. This becomes the foundation for a Marketing Attribution model that everyone, particularly the finance team, can trust.
Agreeing on standards
Putting this in place isn't just a technical job. It's as much about people and process. It means the marketing team needs to agree on a standard set of naming conventions and do their best to stick to them. It also means the data team needs to focus on building the infrastructure that makes marketing's life easier, rather than just pointing out problems.
There is a trade-off here. You might give up a little creative freedom in how you name campaigns, but in return you gain significant long-term credibility. It might feel like it's slowing you down for the first few weeks. But it's the sort of foundational work that lets you move much faster for the next few years, without the constant debates about whose data is right.
The outcome: Less friction, faster progress
Once a system like this is in place, the dynamic usually changes quite a bit. You can launch a campaign and be confident that the data will be captured, cleaned, and presented correctly without anyone needing to step in manually. The data team is no longer seen as a bottleneck. Instead, they become the people who built and maintain the system that helps you succeed. You can walk into any meeting, present your numbers, and know with a high degree of confidence that they are the same numbers the CFO is looking at. That’s really the end of the data debates, and the start of being able to make good decisions, quickly.