Eight months in, the reports look busy and the meetings are pleasant. You genuinely cannot tell whether any of it is working. This is the most common position a marketing buyer occupies. The reason is almost always that someone designed the reporting to describe activity rather than to answer a question.
The reports that tell you nothing
Impressions and reach. Both depend on budget, not skill. Follower growth, which correlates with revenue so weakly that we have stopped putting it on the first page. Engagement rate as a single blended percentage, which conceals whether the engagement came from buyers or from other marketers.
Keyword rankings with no traffic attached, which are easy to improve by targeting terms nobody searches. And the activity dashboard: posts published, keywords optimized, ads launched, hours logged. Activity is an input. Presenting it as an outcome is the oldest trick in agency reporting. It works because it is hard to argue with.
Four checks that tell you whether your marketing agency is working
One, blended acquisition cost. Take total marketing cost, including the retainer and every fee, and divide it by new customers acquired. Use your own system for the customer count, not a platform. Track it as a rolling quarter. It is crude and it cannot be gamed, which is exactly why it belongs on the first slide.
Two, branded search. Are more people typing your name this quarter than last? Nobody searches for a company by accident. That makes it the cleanest available proxy for whether the agency creates demand rather than merely harvesting it. Google Trends gives you a free first read. The signal moves slowly, six to nine months behind sustained upper funnel work, so review it quarterly and never monthly. We cover it in detail in brand search is the metric that survived.
Three, the quality of what arrives, not the count. Look at average order value, cash on delivery refusal rate, lead to meeting rate and close rate. An agency can double lead volume and halve lead quality. The top line number will still look like a triumph for two months.
Four, a holdout. Once or twice a year, switch one channel off across a matched set of governorates for three weeks. Then compare revenue against the control regions. It is the only test that shows what a channel actually adds, and a good agency will propose it before you do.
The relationship checks matter as much as the numbers
Do they bring you bad news before you discover it yourself? Do they ever say no? Are they still asking questions about your business in month nine, or has the curiosity stopped? Does the monthly meeting produce two decisions, or is it a slide read out?
A team that is still arguing with you at month nine is engaged. A team that agrees with everything has quietly moved you to maintenance. If you are still choosing a partner, our twelve questions to ask before you sign help you set these expectations on day one.
If the answer is no, do this before you fire anybody
Check upstream first. In a meaningful share of the underperforming accounts we review, the constraint sat on the client side. Approvals took three weeks. Nobody was allowed to change the website. Sales did not call leads back the same day. Replacing the agency does not fix any of that, and the next one will fail identically.
If the constraint really is the agency, put a written sixty day reset in front of them. Give it three specific outcomes and a named owner for each. Then act on the result without a second extension. Sentiment is not a metric, and neither is how much you like the account manager.
If you want an independent read before that conversation, the free audit looks at the accounts and the tracking rather than the reporting. That is usually where the honest answer lives. Our paid media planning work starts from the same checks.
