GMV Max asks whether you are ready to stop, not to start

Once Product GMV Max runs, video shopping ads and product shopping ads using the same products are paused across every ad account, or those products are removed from them. Switch on the LIVE variant and LIVE shopping ads for the shop official account stop, whichever account created them.

Do not read pause as a temporary switch. If the same products were running across several accounts, those campaigns clear at once, and reversing it means another round trip of removing and re-adding products.

There is one structural constraint on accounts. Only one ad account per shop can use GMV Max, and it has to be the primary ad account designated in Seller Center.

Moving the primary account later is not a settings change. When a different account becomes primary, access is revoked from the previous one and existing campaigns for that shop are paused.

That adds one more question to answer beforehand. Has anyone written down, in numbers, what the campaigns about to stop were producing.

Counting the blast radius in advance removes the surprise. Put the product list and the campaign list on one page and mark which campaign uses which product.

There is no entry threshold, though.

Official guidance states that it works regardless of initial GMV scale or available creative volume, and the platform blog states that there is no minimum creative requirement.

Having no threshold is what makes the timing hard. With no starting condition on the platform side, the brand sets the moment itself, and without a rule for that, the rule becomes let us just switch it on.

Decide about the LIVE variant in the same sitting. LIVE pauses against the shop official account rather than against a product list, so its blast radius is a different shape.

Starting and stopping happen in the same instant, and that is the core of the decision.

The start date is the day you are ready to stop what is already running.

GMV Max attributes orders from the advertised products to its own campaign, organic and affiliate orders included. The campaign figure on the first morning therefore already contains orders that would have happened anyway.

Stating an incremental effect needs something to compare against, and that comparison cannot be built after the fact. Writing down seven or fourteen days of organic and affiliate orders beforehand is the whole of this step.

Write the baseline down in a document of its own. A screen changes every time it is reopened, and a changed number cannot later prove what it said that day.

Writing it down does not take long.

It is four numbers copied from two screens, and they carry the next few months of judgement.

Record the dates alongside them. The same seven days read differently depending on which weekdays they include.

The platform states that advertisers using GMV Max reported a twenty percent lift in GMV during early tests, and labels the basis as internal data. That figure is one sentence the platform reported from its own data, not an expectation you can plan against.

It stays out of your plan because the sample is not yours. The only number that belongs in the plan is your own shop two weeks before switching on.

Do not record the baseline as revenue alone. Writing down order count and average order value beside it lets you separate more buyers from bigger baskets later.

Shop ads attribute on a seven-day click and a one-day view. Last interaction decides, and a click outranks a view.

The attribution key being the shop rather than the product is what unsettles creative judgement. An order for product Y placed after an ad for product X still lands on that campaign, so reading campaign results straight across as creative results gets it wrong.

Judging creative needs a different place to look. The campaign picks products, creative and placements on its own, so one campaign becomes the sum of several pieces, and unwinding that sum piece by piece is not something the report does for you.

This is also where Seller Center and Ads Manager diverge. The first screen uses the order date as its axis, while the second uses the ad interaction and gathers orders inside the attribution window. Without agreeing internally which screen counts, the same argument returns in the same place every week.

Pixel and Events API defaults are also seven-day click and one-day view, with click selectable at 1, 7, 14 or 28 days and view at off, 1 or 7 days. That choice cannot be changed once the ad group is published.

Changing it later means building a new ad group, and at that moment the comparison with the earlier period breaks. That is what irreversible actually costs here.

Which of the two screens is correct is not something this page rules on. Both are right inside their own rule, and what the brand picks is which rule the margin gets read under.

Name the source screen on the report itself. Without that note, two people read the same table differently.

Keep the reporting cycle no shorter than the window. Judging every three days against a seven-day click window counts orders that have not landed yet as absent.

Reversibility decides the order of the setup. Nail down the first two rows below, then fill in the rest.

The third column is what makes this table worth keeping. Reversible items can be corrected after launch, while correcting either of the first two rows destroys what was there before.

The setup screen runs product selection, ROI target, daily budget, start time, a suggested commission reduction module, and publish. If the product already sits in a video or product shopping ad or another GMV Max campaign, it has to be removed there before the flow continues.

If shop ads are already running, separate migration rules apply. Reading that document before opening the setup screen removes a round trip of pulling products out and putting them back.

The ROI target and daily budget are reversible, but without recording the date of every change you lose the ability to explain a shift in results. Reversible and traceable are two different properties.

Keeping the order lowers how much has to be undone. With attribution and the primary account nailed down first, everything else can be adjusted after launch without shaking the axis of comparison.

Make one page before setup. Four lines — attribution value, primary account, baseline period, products removed — let you retrace what changed.

No threshold does not mean any moment will do.

What has to be in place before launch belongs to the checklist in the next section. These two items are a different kind of thing: the day is wrong even when the preparation is right.

If the affiliate commission is being adjusted at the same time, count the grace period in. A decrease does not take effect immediately: creators already promoting that product keep the previous rate for thirty days, while an increase applies at once.

A day of delay usually costs very little. A month run without a baseline, by contrast, leaves every judgement made in that month without grounds.

There is also a case for not waiting at all. With the product shopping ads already stopped and the baseline and attribution value settled, that day is the start date.

One item never improves by waiting. A baseline does not appear on its own, somebody has to write it, so if the baseline is the reason for delaying, today is the cheaper day to write it.

Published tempo guidance lists GMV Max campaign optimisation inside the operations scope. That states a scope of work rather than promising a result, so what to ask an agency is not the outcome but who decides the two irreversible settings above.

Even with an agency running the campaign, the baseline and the attribution value are better held by the brand. Keep both in a brand document and the axis of comparison survives a change of agency.

Everything below takes about five minutes immediately before publishing. The first two cannot be undone afterwards, so keep the order.

Two numbers were never invented for this page. Official documentation does not specify a minimum budget or a learning period, so both cells stay empty, and the list of available markets is not transcribed here because official pages were found to state it differently.

The first three cannot be replaced by memory. The attribution value, the primary account name and the baseline period take minutes to write down, so keep them in the same document and leave it open beside the publish screen.

For the first few days afterwards, start by checking what stopped. The automation picks products, creative and placements on its own, so what a person should look at is the gap the previous campaigns left behind.

Week one is a consistency check. Three things: that the campaigns which should have stopped did, that no product is entered twice, and that the report is coming from the screen you agreed on.

Do not rush the first judgement. Until a stretch as long as the baseline period has passed, only half the comparison exists.