When a campaign loses performance, the common reflex is to duplicate ad sets, cycle audiences, or kill ads in a hurry. I prefer a different order: ask about the business context first, only then touch the tool.
If this were your business, would you spend money on this page?
Before sending paid traffic to any page, I have to answer honestly whether I'd risk my own money on that experience. I put it directly to my students: if the page were yours, if the business were yours, would you actually do this?
It's common to hear media buyers say the CTR was great, CPC was low, but the client's page just didn't convert, as if that weren't their concern. I reject this stance: a strategic media buyer isn't driven by excuses, they're driven by results. If the landing page was weak, the responsibility, as a business partner, was never to have sent money there in the first place. If the client insists on running ads without fixing the identified issues, the right move is to state clearly that results aren't guaranteed while that obstacle remains, and in that case there's no basis to charge as if the job were complete.
Are you evaluating offer strength or just delivery quality?
I carefully separate two concepts that get confused: delivery is what happens after payment (support, fulfillment, product robustness); the offer is the communicated perception, the clarity of the promise, how easily someone understands the benefit. The best product in the world, poorly communicated, will sell less than an average product that is extremely well communicated.
Many ads fail because the copy lists features instead of solving a pain point. Before touching budgets or audiences, it's worth reviewing whether the message talks about modules and specs or clearly communicates what the person gains. I work different angles and hooks for different awareness levels, since a cold audience needs a different approach than an audience that already follows the brand or has visited the site.
The Diagnosis I Run Before Proposing Any Marketing Plan
What I expect from a media buyer in the AI era
Growing an online business: what I learned about margin, team, and AI
What is the real revenue generated by each ad, not just cost per lead?
Optimizing only for the lowest cost per lead is a classic mistake. The example I gave in class is simple: a campaign with a hundred-euro budget generated a hundred leads at one euro each, but closed zero sales. Another campaign, also with a hundred-euro budget, generated only ten leads at ten euros each, but those resulted in actual sales revenue.
Whoever optimizes purely by cost per lead would have killed the second campaign, the one actually bringing money into the business.
To avoid this mistake, I recommend linking each lead to the sale it generated, even if that starts simply in a Google Sheet. Answering a question in class about clients without a CRM, I said plainly: you build a CRM for them, and that CRM can start in Google Sheets. The goal is being able to ask the client who they closed deals with that month, then trace that sale back to its origin: which ad, which creative, which channel.
What historical data supports the change you want to make?
In class I stressed the importance of comparing the last three days of a campaign against seven, fifteen, or even thirty days of history before deciding to cut budget or stop a creative. Platform auctions have normal fluctuations in CPM and CTR; reacting to a brief dip of a few hours introduces noise and doesn't let the algorithm learn.
As a hypothetical illustration, I described an agent built in a tool like N8N that compares time periods and sends an email alert when a creative underperforms over the last three days compared to the fifteen- or thirty-day average. This is an example of how one could automate this comparison, not a description of a process I have necessarily implemented exactly this way for every client. The advantage of using such tools is speed of analysis; the same date-range filter can also be done manually inside the ad platform itself.
Is this campaign meant for direct profit or to feed the value ladder?
Demanding immediate profit from a campaign designed to acquire new customers is a failure to see the business as a whole. I gave my own example in class: the ads for the training I was promoting weren't built for direct profit, they were built to bring in new clients; profitability comes afterward.
That campaign could even run at a loss rather than simple break-even, if the back-end math, with the clients acquired through that product, works out later.
A strategic media buyer asks the client what happened afterward with those leads: how many upgraded to the main offer, what was the return in the following months. Without that view, you risk pausing an effective acquisition campaign just because, viewed in isolation, it looks unprofitable.
Can you explain this decision in plain business language?
Whoever overcomplicates things usually doesn't understand them. A strategic media buyer can explain what they did and why in terms any business owner understands, without hiding behind technical jargon. The client doesn't care about CPM or relevance scores; they want to know if their money is safe and what return it's generating.
Before touching any campaign settings, try stating the reason for the change in simple sentences: which ad brought real customers, which one brought only curious contacts, and that's why you're shifting the budget. A client who understands the logic trusts you more than a confused client who assumes the media buyer "must be good" simply because they didn't understand a word.
Bring a concrete decision to the meeting
One way to prepare is to write down the change you are considering, the signal that prompted it and the result you expect to observe. Include the data you have and identify what still needs confirmation from the client. For example, a fall in sales might coincide with fewer enquiries, unavailable stock or delays in sales follow-up.
If you change the ad before understanding which situation occurred, you may alter something that was working. Use the meeting to confirm the explanation and agree on responsibilities. Then record the decision and review the results at an agreed time. This small record helps the team learn from the changes it actually made.
Source note
This article draws from the class I taught in Sextas Ímpares #120, "Como ser um Gestor de Tráfego Estratégico em 2025?", available at youtube.com/watch?v=vvGfLEDE3vo. Worth watching: the distinction between offer and delivery around the 19-minute mark, the lead-vs-revenue campaign example around 53 minutes, and the point about plain business language around 58 minutes.