The overwhelming data dilemma: three numbers that matter

The overwhelming data dilemma is the state where paid media teams have more metrics than decisions. The way out is to anchor on three numbers: true-profit ROAS per campaign, blended MER across all marketing, and creative health per ad. Everything else becomes a diagnostic you consult when one of those three moves, not a report you read every morning.

Why does more data make decisions harder?

Meta and Google each expose hundreds of metrics. Analytics tools add more, agencies add their own decks, and finance adds the P&L. Every number is defensible on its own. Together they create noise, and noise has a cost: meetings that review dashboards instead of making choices, and teams that pick whichever metric supports the decision they already wanted.

"The overwhelming data dilemma" is one of the moments that surfaces again and again in SOMIN's listening on marketers working with AI and analytics. The underlying tension is not a lack of tools. It is fragmented analytics and old measurement frameworks that were never designed around profit. More data on top of that only deepens the fog.

Which three numbers actually matter?

1. True-profit ROAS by campaign

Gross profit on kept orders divided by spend, after returns, fees, shipping and discounts. It tells you which campaigns pay their way. The method is in our true-profit ROAS guide.

2. Blended MER

Total revenue divided by total marketing spend. It ignores attribution entirely, which is its strength. If campaign-level numbers all improve while MER falls, something is being double counted.

3. Creative health per ad

A simple status for each active creative: healthy, watch or rest, based on click decay, frequency and profit drift. It tells you where next week's work is. See our creative fatigue guide for the signals.

What should you stop reading every day?

• Impressions and reach, unless you are running awareness with a defined goal.
• Platform ROAS on its own, without true-profit context.
• CTR across the whole account; read it per creative, when health changes.
• Daily swings in anything; weekly is the right rhythm for most budgets.

These are not useless. They are diagnostics. You consult them when a headline number moves and you need to know why.

A one-page weekly report template

1. Headline: MER this week versus the four-week average.
2. Campaigns: true-profit ROAS for each, ranked, with any that crossed below your floor highlighted.
3. Creative: count of healthy, watch and rest creatives; the replacements ready.
4. Decisions: budget moves made, creatives rested, tests launched.
5. Questions: one thing the data cannot yet explain.

If the report does not end with decisions, it is a dashboard, not a report.

Where does AI help, and where does it add noise?

AI is good at summarising large volumes of data and spotting patterns. It is also good at producing confident paragraphs about patterns that do not matter. The guardrail is the same as for humans: insist that every AI conclusion can be checked against the data behind it. That principle runs through SOMIN's reporting tools, where conclusions trace back to the evidence. For a wider look at how reporting tools compare, SOMIN's comparison of reporting tools is a fair starting point, and the Analytic Partners case study shows how measurement specialists combine audience signals with analytics.

The question of how much AI should do, and how much humans should own, came up directly in one of the Marketing Mondays sessions at Block71, the panel on AI's role in creative, media and brand marketing. The consensus among practitioners in the room was familiar: AI to compress the data, humans to decide.

Definitions

• Headline metric: a number reviewed every week that directly drives decisions.
• Diagnostic metric: a number consulted only when a headline metric moves.
• MER: marketing efficiency ratio, total revenue over total marketing spend.

Why do teams resist cutting metrics?

Partly habit, partly fear. A long dashboard feels thorough, and nobody wants to be the person who removed the metric that would have caught the problem. The answer is not to delete diagnostics but to demote them. Keep every metric available, and agree which three get reviewed every week. When one of the three moves, open the diagnostics and trace the cause. This gives the team both focus and reassurance: nothing is hidden, but not everything is shouted.

It also helps to agree thresholds in advance. Decide what movement in MER triggers an investigation, which true-profit floor a campaign must hold, and how many weeks of decline turn a creative from watch to rest. Pre-agreed thresholds remove the daily debate about whether a number is worrying. The rules decide, and people spend their energy on the response rather than the argument.

Finally, fix the inputs before adding outputs. Most data overload is caused by unreconciled sources telling slightly different stories. One reconciled set of orders, returns and costs does more for clarity than any new chart.

A dashboard that cannot tell you what to do on Monday is decoration.

How FMedia reduces the noise

FMedia is built around the three numbers. Its home view shows MER, true-profit ROAS per campaign and creative health, with diagnostics one click away rather than on the front page. Each week it produces a short list of decisions, not another dashboard. Every recommendation links to the data behind it, so you can challenge it, and nothing changes in your accounts without approval.

Less to read, more to decide. That is the only fix for data overload we have seen work.

Frequently asked questions

What is MER in marketing?

Marketing efficiency ratio: total revenue divided by total marketing spend. It ignores attribution, which makes it a good check on platform-reported numbers.

How often should I review paid media data?

Weekly for decisions, with automated alerts for anything urgent. Daily reviews of most metrics tend to cause overreaction to noise.

Should AI write my ad reports?

AI can summarise well, but every conclusion should trace to the data behind it, and decisions should stay with a human owner.