Monthly Agency Strategy Audit
A client project doesn't end with the pitch - it runs for months. The monthly strategy audit compares the last planned state against the current situation - using the local JSON scenario export as a fixed reference point. A vague "it's going fine" becomes a traceable target-vs-actual comparison that agency and client have in front of them together.
Why the one-off look back is weaker
No reference point, no comparison
Whoever looks only at the current numbers in the monthly meeting sees a state, but no movement. Whether the mix is running better or worse than planned can't be said without the recorded starting state - you discuss impressions instead of deviations.
What the recorded state changes
The last exported scenario is the target: the assumptions the month was planned with. Using the "compare plans" function, the tool places the key results of target and actual right next to each other - the deviation becomes visible instead of being pieced together from two separate views. That sets the agenda for the conversation.
How the monthly audit works
- Load the last state. Read the scenario saved at the previous audit - the target state - back into the model.
- Capture the actuals. Enter the actual figures from the past month - budgets, rates, result. Depending on how you work, as a separate actuals plan or directly in the existing scenario; the previous version stays as an archive JSON.
- Compare plans. Use the compare function to place the key results of target and actual side by side in the tool and read off where plan and reality diverge.
- Carry the scenario forward. Adjust the assumptions for the coming month based on what you've learned - recalibrate instead of guessing.
- Save anew. Export the carried-forward state as JSON again - it is the target for the next audit.
The JSON export as a shared artifact
The file sits locally, not in a cloud - and both sides can hold it. The client keeps their own backup and can edit it themselves; work on the scenario is shared between client and agency. No login, no exclusive agency access, no black box: one state both can open and follow.
Export a scenario and read it back at the next audit: try the tool
Talking points for the conversation
- "We're not comparing gut feeling, but the planned state." Shifts the meeting from impressions to demonstrable deviations.
- The file belongs to the client too: the backup sits with them, not only with the agency - that builds trust instead of dependence.
- A deviation means "let's look" first, not "who's to blame": the comparison opens the conversation instead of judging it upfront.
- Treat the carried-forward state as an agreement - both sides know what the next audit measures against.
Common thinking traps
- Confusing deviation with cause. The comparison shows that actual and target diverge - not why. What caused the deviation remains an interpretation, not a measurement from the model.
- Reading the scenario as measured truth. The target state is a set of assumptions, not causal proof. It is only as good as the numbers that went into it - a planning layer, not a data-driven MMM.
- Running the audit as a pure agency ritual. If only the agency holds the file and the client never opens it, the real benefit - shared traceability - is lost.
Frequently asked questions about the strategy audit
Why a monthly audit instead of a continuous dashboard?
A comparison needs a fixed reference point. The JSON export records the planned state the next month is measured against - a continuous dashboard shows the current state, but not the deviation from the plan.
How does the target come about?
From the last exported scenario. It holds the assumptions the past month was planned with and serves as the benchmark for the actual figures at the audit.
Can client and agency both work on the scenario?
Yes. The JSON file sits locally and can be passed along - the client keeps their own backup and can edit it themselves. Work on the scenario is shared between client and agency.
Does the comparison prove that marketing caused the deviation?
No. The audit shows where plan and reality diverge, not why. The cause remains an interpretation - the model is a planning layer, not a causal proof of effect.
Does the audit need a cloud or a login?
No. The scenario export is a local JSON file; the user data is not transmitted to a server. The comparison runs entirely in the browser.
