The Measurement Blind Spots

Most mobile teams don't fail because they lack data - they fail because they trust the wrong parts of it, or don't really know what they're looking at.

After working with advertisers, DSPs, and internal growth teams, I've noticed a pattern - measurement blind spots rarely look like errors in your data. Most of the time it's just a "good enough" dashboard or an "industry standard" setup coupled with decisions on attribution logic no one remembers making.

There are probably 20 blind spots I could list, but here are the first 4 that came to mind while writing this post.

Blind Spot #1: Treating Attribution as “Truth” Instead of a Model

Attribution is not reality. It's a model of reality, with assumptions you baked into it when you set up your measurement.

When teams forget that, things fall apart:

• Dashboards turn into meaningless scoreboards
• Discrepancies become political between teams and partner integrations
• "Which number is correct?" replaces "What decision should we make?"

Different partners, windows, and engagement types can all be "right" and still tell different stories.

The danger isn't disagreement - it's not knowing which story your organisation is currently acting on.

Blind Spot #2: Assuming “Standard MMP Setup” = “Correct Setup”

Most apps running today go with a default MMP configuration:

• Default attribution windows & engagement priorities (7 days for click, 24h for view)
• Simple and non-strategic event mappings
• The most basic MMP SDK implementation their dev team agreed to do
• A lack of comprehension regarding how the attribution logic functions

None of this is wrong by definition. But none of it is neutral.

MMPs don't know your internal business logic, and they rarely surface the full range of customisation options unless you actively look for them.

Every attribution rule is a business decision in disguise - and if you didn't explicitly choose it, you're still living with its consequences.

What could go wrong? A few examples:

• UA optimises for signals that don't align with revenue timing
• Retargeting claims credit the business doesn't believe in
• Product teams question marketing data but can't point to why
• You are sending wrong signals back to your ad networks, downgrading performance

The real issue: No shared understanding of why the setup looks the way it does.

Blind Spot #3: No Measurement Ownership

One of the most common blind spots I see: each team optimises their slice of the funnel perfectly.

• UA optimises CPI or early ROAS
• Retargeting optimises short-term lifts or dormant users
• BI optimises report consistency and data health
• Product works on alignment of business logic with internal tools

Each team will use different attribution links, access data asynchronously from other teams, or simply activate or deactivate something without notification.

Individually, each decision makes sense. Collectively, they often cancel each other out or create a mess in the data.

Measurement becomes fragmented - not because people are careless, but because no one owns the full measurement logic end-to-end.

Blind Spot #4: Confusing Ongoing Activity With Understanding

Dashboards update, numbers move, weekly reports are sent - and yet, when something breaks, no one knows:

• Which assumption failed
• Which rule is responsible
• Who should own the investigation

A healthy measurement setup isn't just observable - it's explainable.

If your team can't articulate:

• Why a campaign is credited
• Why an event is prioritised
• Why a discrepancy is acceptable

Then you don't have clarity - you only have data on dashboards running pointlessly.

A Final Thought

Most teams don't need more tools, more dashboards, or more data.

They need a shared understanding of:

• Which assumptions their measurement is built on
• Where those assumptions help
• Where they quietly limit better decisions

Looking under the hood doesn't mean something is broken - it usually means the team has matured enough to question what it's been taking for granted.

If this post raised more "we should probably talk about that" moments than clear answers - that's not a problem. That's usually the starting point.