Start with a question you can finish
The easiest way to lose a morning is to open every dashboard without deciding what you are looking for. Try starting with one question: what changed for this app since the last review, and is there something we should investigate?
This is an editorial workflow, not a promise of revenue growth. Its purpose is to make a team’s reasoning visible. Pick one reporting window, keep a note of releases and campaign changes, and use the same starting view each time.
Read volume, value and experience together
eCPM is revenue divided by impressions, multiplied by 1,000. It is a unit rate, so it needs volume beside it. In an illustrative example, 10,000 impressions at $5 eCPM produce $50, while 6,000 impressions at $7 produce $42. A higher rate can coexist with lower total revenue. These numbers are hypothetical.
Next, look at where the volume came from. Did the audience change? Was an app release involved? Did one format stop receiving requests? Keep those as questions until the evidence supports an explanation.
Then return to the actual app. A report alone cannot describe whether an ad arrived at a sensible moment or whether an unavailable rewarded offer left a confusing button on screen. Add a short product check to the review instead of expecting a metric to stand in for the experience.
A review you can repeat next week
Our suggested review note has four fields. Keeping it short makes it easier to compare the reasoning across weeks and easier for someone else to take over.
- Observation: what changed, in which app and time window?
- Context: what releases, traffic shifts or data gaps might matter?
- Hypothesis: what explanation is worth checking first?
- Action: who will check it, and what evidence will they bring back?
Give portfolio growth the same discipline
If you promote another title inside your portfolio, apply the same routine to that campaign. Define the audience, destination and intended behavior before creating the placement. Check the creative in the app, not just in an asset preview.
The best output of a review may be a small configuration change, a follow-up investigation or a decision to collect more evidence. In each case, leave a clear record of why. Over time, that record becomes more useful than a collection of screenshots with unexplained arrows.