The report is for a solution, not a collection of graphs.
A good monthly report answers three questions: what has grown, why has it happened, and what to do in the next period. If, after twenty graphs, the decision is not clearer, the report is overloaded.
Twelve metrics
- Expenses by source.
- Clicks.
- Registry.
- FTD.
- Confirmed by FTD.
- Cost per raw FTD.
- Cost per confirmed FTD.
- Approval rate.
- NGR, if available.
- Affiliate income.
- Profit/margin after expenses.
- Retention or fixed cohort, if available.
Main table
| Source | GEO | Spend | Confirmed FTD | Cost/confirmed FTD | Revenue | Result |
|---|---|---|---|---|---|---|
| paid-social | KZ | … | … | … | … | … |
| seo | RU | … | … | … | … | … |
Do not mix sources in one line if different decisions are made on them.
Add the change block
Next to the numbers, list what has changed: new creative, landing, cap, rate, GEO mix. Without this, it is impossible to link the result with the cause.
Separate block of experiments
For each test, four lines are sufficient: hypothesis, change, result, solution. This connects team work to business metrics.
What to take upstairs
Three short conclusions: the best source, the main risk, the action for the next month. A manager or partner doesn’t need to understand twenty tabs first.
Why we need to keep history
After three to six months, monthly reports become the seasonality base. You can see if quality deteriorates when scaling, how cost changes, and which GEOs are more stable.
Final principle
Do not optimise the report for the sake of beauty. Optimize it so that the next step is obvious: stop, fix, continue the test, or scale.
For the construction of such a structure, materials about unit-economy, cohorts and segmentation.
Add a comparison to the plan, not just last month.
Growth to the past period may look good, but it does not meet the goal. If the plan assumed the cost per confirmed FTD is not higher than $90, and the actual plan is $ 105, the increase in volume does not solve the problem. In the report, it is useful to have plan / actual / variance columns.
Separate the controlled factors
CPM growth is an external variable, and the weak structure of the creative test is an internal one. Separate them in the conclusions. This helps to understand what the team can fix next month and what needs to be factored into the forecast.
Three levels of reporting
The first screen is the management result. The second is sources and GEO. The third is detailed campaigns and experiments. This structure allows you to quickly read the report and, if necessary, fall into the details.
Complete with concrete actions
Do not “improve KZ” but “stop campaign-07, move 20% of the budget to campaign-03, check the new landing on iOS and request an increase in cap while maintaining the approval rate.” A good report ends with decisions with owners and deadlines.
Add a comment to the anomalies
If one metric has changed by more than a selected threshold, there should be a comment next to it: a known cause, hypothesis, or test task. This makes dashboard a working document.
Don’t overload the first screen.
Five key KPIs and three inferences are usually more useful than fifteen charts. Details should be available below, but do not interfere with the main story of the month.
Link the action to the following report
Each task from the final block should have a metric by which the result will be evaluated in a month. Reports then become a consistent management system rather than independent presentations.
