The Shopify lift number your agency will accept
The argument with your agency is never really about the number. It is about who chose the method. Four things to agree in advance so the result survives contact.
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 Joris van Huët, Founder & CEOUpdated 3 min read
Run the numbers for your store: the free holdout test planner.
An agency will not accept a lift number that arrives after the fact from a method they did not see chosen. That is not obstruction, it is the correct response to being measured by an instrument someone else selected once the results were known.
The fix is boring and it works: agree the instrument before the period, not after.
The four things to settle in advance
| Item | Agree it before | Why it matters afterwards |
|---|---|---|
| The window | The period starts | Otherwise the window becomes the argument |
| The channel grouping | Any read is run | Regroupings change results and look like manipulation |
| The success threshold | You see any numbers | Post-hoc thresholds convince nobody |
| Who runs the read | The first disagreement | An independent read is worth more than either party's |
The third row is the one that saves the relationship. "Above break-even" is not a threshold, because break-even depends on margin and on which costs you count. Write the actual figure down.
Why the method matters more than the tool
Agencies push back hardest on estimates that arrive without a stated design.
Given a confident single figure from a black box, the sensible response is scepticism, and you would want your agency to be sceptical of a black box pointed at their work. That reasoning is set out in why vendors should explain methodology openly, and the practical mediation is in settling attribution debates with your agency.
The one asymmetry to name out loud
Your agency's reporting comes from platforms that sell media. That is not an accusation, it is a description of the incentive, and it is the reason a third read exists at all. Saying it plainly once, at the start, is better than implying it repeatedly later.
The counterpart is worth conceding too: your agency knows things about your account that no aggregate model sees, including creative fatigue, auction dynamics and what they tried last quarter. A lift number is an input to that conversation, not a replacement for it.
What to bring to the meeting
One page. Then the two questions that make it a conversation rather than a verdict: what would change your view, and which channel should we test properly next.
When the disagreement is real
Sometimes the agency is right and the model is wrong, usually because something structural happened in the window that aggregate data cannot see. That is exactly the case a geo holdout settles. Agreeing in advance to run one on the contested channel converts an argument into a plan.
Related answers
Frequently asked questions
How do I stop attribution arguments with my agency?
Agree the window, the channel grouping, the success threshold and who runs the read before the period starts.Should my agency run the attribution read?
An independent read carries more weight than either party's own, precisely because neither side selected the instrument to suit a result. If the agency runs it, agree the method in advance and make the raw output visible to both sides.What if my agency disputes the lift number?
Ask what would change their view, then run a geo holdout on the contested channel. A dispute about an observational estimate is usually a dispute about design strength, and a holdout is the instrument that settles it rather than escalating it.
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
- AttributionAttribution identifies user actions that contribute to a desired outcome and assigns value to each. It reveals which marketing touchpoints drive conversions.
- Multi-Touch AttributionMulti-Touch Attribution assigns credit to multiple marketing touchpoints across the customer journey. It provides a complete view of channel impact on conversions.
- ShopifyShopify is an ecommerce platform for creating online stores and selling products. Attribution modeling shows which marketing channels drive traffic and conversions within Shopify.