It's Not "What's My ROAS?"
Every founder asks about ROAS. Every agency has a polished answer ready.
That's exactly why it's the wrong question.
The question most agencies cannot answer well is far simpler:
"Which of my products should we stop advertising?"
Watch what happens when you ask this.
Why This Question Is So Revealing
A good agency should be able to point to specific SKUs or categories that are hurting your overall profitability.
They should have an opinion on which products are conversion traps: they generate sales, but those sales cannibalise margin, inflate costs, or drag down algorithmic efficiency.
Understanding which products to cut is central to our audit process.
If your agency cannot answer this clearly, they either:
- Haven't looked at your data beyond surface metrics
- Don't understand your margin structure
- Are afraid to recommend reducing their managed spend
None of those are acceptable.
The Bad Answers
"We should advertise everything and let the algorithm decide." This is abdication, not strategy. Algorithms optimise for the objective given, not for your business outcomes.
"All your products are contributing positively." Statistically improbable. Every catalogue has margin drags.
"That's more of a merchandising question." No, it's a commercial efficiency question. If spend is going to unprofitable products, that's an advertising problem.
"We'd need to look into that and get back to you." Acceptable once. If they still don't have an answer after a week, they aren't managing commercially.
The Good Answer
A good answer sounds something like:
"Based on your margin data and performance, here are the products we've either paused, negatively targeted, or moved to low-priority campaigns. We review this monthly."
And then they can actually show you the list.
This is what profit-led governance looks like in practice.
What This Question Exposes
Agencies are incentivised to increase managed spend. More spend, more fees (usually).
Recommending that you stop advertising certain products runs against that incentive.
The question tests whether your agency prioritises your profitability or their revenue.
It also tests whether they understand your business beyond the ad platform. Margin structures, inventory constraints, seasonal dynamics... all of these should inform which products get pushed and which get pulled.
Agencies that understand commercial reality act differently.
The Follow-Up Questions
If they pass the first test, keep going:
"How do you decide when a product moves from active to paused?" There should be a framework, not just intuition.
"What would cause you to recommend we reduce overall spend?" If the answer is "nothing," they're optimising for volume, not outcomes.
"Show me the products you've cut in the last 90 days." Evidence, not claims.
Why Most Agencies Fail This
Managing commercially requires work that doesn't scale.
It means understanding margins at SKU level. Reviewing performance against profit, not just ROAS. Making recommendations that might reduce spend.
Most agencies operate on efficiency models that don't allow for this depth.
We work with fewer clients because this approach doesn't scale infinitely.
The Point
You can tell more about an agency from how they answer this question than from any case study or proposal.
If they squirm, deflect, or give a non-answer, that tells you something important.
If they answer with specificity and evidence, that tells you something too.
Book a strategy call if you want to work with an agency that has opinions.
Curious what we'd cut from your account? Request an audit and we'll tell you directly.