The metric you choose dictates the decision you make.
Most paid media reports are built to show progress, not truth. The right metric for this month's slide deck isn't always the right metric for the business. Here's how we build reporting that runs a company.
What most agencies report. What we report instead.
The metrics in the left column aren't wrong. They're incomplete. Every one of them becomes meaningful when it's anchored to a business outcome. On their own, they're theater.
| Most agencies report |
|---|
| Impressions |
| Clicks |
| CTR |
| Overall CPL |
| Monthly blended ROAS |
| "Great results this month" |
| Loocro reports instead |
|---|
| Revenue per visitor |
| Cost per qualified opportunity |
| Install-to-trial rate (for apps) |
| CPL by qualification stage |
| ROAS by channel, campaign, and audience |
| Cohort-based CAC payback |
Six principles behind how we build every reporting layer.
- We start with unit economics: what does it cost to acquire a customer, and how long until that cost is recovered?
- We segment before we aggregate. Blended numbers are for context. Segmented numbers are for decisions.
- We close offline loops. If deals close on WhatsApp or by phone, we import that data back into the platform.
- We track cohort churn. Monthly averages don't tell you whether the business is getting better or worse.
- We report on what can be acted on. Every metric in our weekly report has a corresponding decision attached.
- We separate what the data shows from what we recommend. The two are often different — and both matter.
Four businesses that changed what they measured — and changed what they decided.
A fashion store with healthy blended ROAS and stagnant revenue. When we split new vs. returning buyers, new customer acquisition was below the margin floor. Rebuilt around margin-aware targeting, hit 11.3x.
Read the case →Hundreds of leads, almost none converting. Switched the optimization goal from form fills to qualified opportunities. The lead volume dropped by half. The deal volume tripled.
Read the case →Deals were closing on WhatsApp. The ad platform had never seen a closed deal. We imported offline conversions, the algorithm learned what a real buyer looked like, and CPL dropped 70% in 12 months.
Read the case →B2B and B2C audiences were mixed in the same account, producing blended CPLs that meant nothing. We separated campaigns, conversion goals, and reporting — giving both teams numbers they could actually use.
Read the case →The reading behind the method.
Why blended ROAS hides the bleed.
The blended average is a number for the monthly slide. Segmented ROAS is the number that runs the business. Two real cases inside.
Read →How to build a segmented ROAS report in 4 steps.
The setup that separates brand from non-brand, new from returning, and product categories from each other.
Read →Cost per qualified opportunity matters more than CPL.
A cheap lead that doesn't close is the most expensive lead you can buy. Why CPQO is the B2B metric that pays the bills.
Read →Why B2B campaigns break when offline conversions aren't tracked.
If 80% of your deals close on WhatsApp, phone, or in person, your campaigns are optimizing against half the truth.
Read →The questions we get about measurement.
What's wrong with reporting on ROAS and CPL?
Nothing, if they're segmented correctly and anchored to unit economics. The problem is when blended ROAS becomes the primary optimization target. A 6x blended ROAS can coexist with negative new-customer acquisition if repeat buyers are inflating the average. The metric isn't wrong — the scope is.
Our dashboard shows great numbers but revenue growth is flat. What's the disconnect?
Usually one of three things: you're measuring activity rather than outcomes, you have attribution inconsistency across channels, or the dashboard is showing you what happened rather than what caused it.
How do you handle multi-touch attribution?
We use a combination of platform-reported data, UTM-based tracking, and offline conversion import where applicable. We don't rely on any single attribution model. The goal is to triangulate, not to declare one source of truth and ignore the rest.
We have a complex funnel with many touchpoints. Can you still measure incrementality?
Yes. We use holdout tests, time-series analysis around campaign changes, and channel-level revenue per visitor tracking. We don't promise pixel-perfect attribution — but we can tell you which channels are pulling weight and which ones are taking credit.
What reporting cadence do you use with clients?
Weekly performance reviews focused on decisions, not updates. We send a structured report that flags anomalies, explains causes, and recommends actions. The goal is a 15-minute read that produces a clear next step — not a 40-slide deck that produces a meeting.
We're not sure which metrics matter for our business model. Where do we start?
With the unit economics: what does it cost to acquire a customer, how much do they generate in the first 12 months, and how long until acquisition cost is recovered? Everything else in the reporting structure flows from those three numbers.
Book a 30-minute diagnosis.
We'll audit your current reporting setup. We'll tell you which dashboards are signal and which are theater. We'll show you which metric is driving your decisions — and whether it should be.
Book the Diagnosis →