Skip to content

Telling the team a channel is wasting spend

The person running it will hear a verdict on their work, because that is what it sounds like. Four framings that keep it about the budget instead.

Three platforms, three answers, one budget to defend. Your GA4 export already holds the answer, and the first finding is free and never leaves your browser. Get your first finding free

By , Founder & CEOUpdated 2 min read

Run the numbers for your store: the free break-even ROAS calculator.

Say a channel is wasting spend and the person running it hears a verdict on their work.

The four

Instead ofSay
"This channel is wasting money""The evidence for this channel is weaker than for the others"
"The numbers say cut it""The interval sits below break-even, which is our agreed cut condition"
"Reported ROAS is wrong here""Reported ROAS answers a different question; here is the gap"
"We are cutting it""We want to test it before deciding"

Each right-hand version is more accurate, not just softer.

Explain the mechanism before showing the channel

If the channel in question is branded search or retargeting, the finding is entirely predictable from how proximity-based credit works, and saying so first turns it into a structural observation rather than a performance review. The mechanism is in platform attribution overcounting.

Predictability is the point. A finding that could have been forecast before the analysis is obviously not about the person.

Use a pre-agreed threshold

If the team agreed in advance that an interval sitting entirely below break-even is a cut condition, the conversation is about whether the condition was met rather than about whether the channel deserves to exist.

Give them the next move

Ask what they would test, and then test it. The person closest to the channel knows things the model cannot see: creative fatigue, auction dynamics, what changed in the window. They are often right, and inviting the objection produces better analysis as well as a better conversation.

Do not attach the budget decision

Presenting the finding and the cut together guarantees the finding gets contested, because the finding is the part that can be argued with. Separate them by at least a week, as argued in showing the gap to your team.

The material

Show the widest interval yourself.

Frequently asked questions

  • How do I tell someone their channel is underperforming?
    Say the evidence is weak rather than the channel is wasting money, name the pre-agreed threshold that was crossed, and explain the mechanism first. If it is branded search or retargeting, the finding was predictable before the analysis.
  • Why agree cut thresholds in advance?
    Because it turns the conversation into whether a condition was met rather than whether a channel deserves to exist. That is entirely a matter of sequencing and it makes a difficult discussion considerably easier.
  • Should the channel owner get a say?
    Yes, and ask first. They know creative fatigue, auction dynamics and what changed in the window, none of which an aggregate model can see. Inviting the objection improves the analysis as well as the conversation.

Go deeper: Incrementality testing, explained.

Sixty-second versions of these ideas: Causality Engine on YouTube Shorts.

Have an audience of ecommerce operators? Earn 20% recurring commission by recommending the €99 read.

Keep reading

Terms in this article

Browse the full glossary

Your platforms guess.
We run the math.

Upload a GA4 export and see what each channel caused, next to last-click, in 1–2 minutes. The read is yours to keep.

The free finding runs in your browser. Your file never leaves it. The full read is €99, refundable within 30 days, no questions asked. Prices exclude VAT.