Quick Answer: For every metric you report, answer three questions: what changed, why it changed, and what to do next. If a line in your report does not answer the third question, it is description, not advice.
Here is a sentence from a typical client call:
"Your cost per click went up 12% and conversions were flat."
It is true, and the client can do nothing with it.
Compare it with this:
"Cost per click rose 12% while conversions held flat, mostly because two campaigns are bidding against each other on the same terms. We will pause the weaker one this week and expect cost per click to ease."
The numbers are in both. Only the second one is a decision.
(The figures above are illustrative, not from a real account.)
The what, why, next structure
Use it for every metric that matters:
- What: The metric and the change, with the number
- Why: The most likely cause, said honestly
- Next: One specific action
If you cannot name the cause for certain, say so. "Most likely" is honest. Overstating is how you lose trust the day the guess turns out wrong.
Three habits that make it work
1. Lead with the action. Clients read the first line. Put the decision there and the numbers underneath.
2. Pick one action per metric. Three recommendations feels like none.
3. Keep an internal version and a client version. You will sometimes need to be blunt inside the agency ("this landing page is hurting us") and diplomatic outside ("there is an opportunity to improve the landing page"). Same facts, different tone.
Where agencies get stuck
The structure is simple. The effort is in finding the why, which means looking across Google Ads, Meta, analytics and the CRM for each client.
That is the part LedgeSpace's RUSHA AI is built for:
- Every answer has the three parts. RUSHA is set up so each response states what happened (the metric and percentage change), why (the root cause) and what to do (a specific action, with numbers where it can)
- Numbers are checked. Numbers in an answer are checked against the raw integration data. If one cannot be matched, the response is rejected and retried
- Two voices. RUSHA writes a blunt internal version and a diplomatic client version, and a programmatic check confirms every number in the internal version also appears in the client version
- Careful wording. It is barred from absolute words like "confirms," "proves" or "definitively"
- You stay in control. Proactive anomaly and correlation alerts arrive as drafts. A client sees nothing until you approve it, and you only use an insight credit when you approve
RUSHA is an optional add-on: $26/mo for 225 insights, $46/mo for 600, $99/mo for 1,500, billed monthly. Every platform plan has a 14-day free trial that includes 20 RUSHA insights.
Try it on one call
Before your next client call, take the three metrics that moved most. Write the what, why and next for each. Then watch how the conversation changes.
To see how RUSHA does it from your own client data, start a free trial (no credit card), or read how RUSHA handles client calls.
