I judge ad campaign performance by bringing the platform reports back to the business. I want to know how much we spent, how much revenue actually came in, and whether the customers we acquired are the customers we intended to reach. A strong number inside an advertising dashboard is a reason to investigate further.
During a recent advertising review, I explained that I still use reported return on ad spend. It helps me compare campaigns and spot changes. But the decision to increase budget comes from several pieces of evidence agreeing with each other.
Start with the total
For an ecommerce account, I look at advertising spend across the platforms alongside the revenue recorded by the business. That gives us a shared view before we get into arguments about which channel deserves credit.
In a service-business review, we also brought the agency fee into the acquisition calculation. That was useful because the owner needed to understand the marketing investment as a whole. The number still excluded other operating costs, which meant we could not describe it as profit.
This is why I ask what is included in a return figure. A business can cover its media spend while struggling once the rest of the costs are considered. The target needs to make sense for that business.
Revenue divided by total ad spend is a blended efficiency measure. It includes revenue that may have happened without those ads, so it does not prove incremental impact. Keep its scope clear and use consistent revenue definitions and date ranges when comparing periods.
Read the campaign in the context of its job
A new-customer campaign has a different job from a campaign encouraging existing customers to buy again. I want to know which one we are judging before deciding what good performance looks like.
I have raised this in creative reviews too. A product can sell strongly without proving that the associated ads are bringing many new customers into the business. If acquisition is the brief, we need to examine new and returning customer purchases as well as overall sales.
Frequency, cost per click, cost per conversion and engagement help me understand the pattern. For example, a campaign repeatedly reaching the same audience needs a different discussion from one bringing in relevant new visitors. None of those metrics tells the entire story by itself.
Look for a chain of supporting evidence
When we previously scaled an influencer campaign, I described the reasoning in practical terms. The traffic was arriving at an affordable cost. We could see relevant social traffic in analytics. Engagement looked useful, and business revenue had moved upwards.
That combination gave us a reason to increase the budget and observe what happened next. I was looking for revenue to follow the extra investment. The low click cost alone would not have been enough.
I still want to know whether the pattern survives the next change. A campaign that looked attractive at one spend level may not behave the same way with more budget, a different creative or a different audience mix.
An increase after a campaign launch is supporting evidence, not proof of causation. Promotions, stock availability, seasonality and other channels can move at the same time. Where practical, a controlled experiment gives a stronger answer about the campaign's additional contribution.
Ask what happened after the first sale
For a service business, I also want to know whether acquired customers return. In one review, we compared initial acquisition costs with rebooking by source. That opened a more useful conversation than treating every first appointment as the final result of the campaign.
Future value matters, but it needs to become actual repeat business. I would rather watch what those customers do than use an optimistic lifetime-value assumption to explain away weak results.
If your advertising reports look healthy but the business feels under pressure, put the platform numbers beside revenue, new-customer purchases and repeat purchases. That is the conversation I would have before approving the next budget increase.