Target ROI is the single instruction you give GMV Max. It is not a forecast or a hope – it is a constraint, and the campaign will decline to spend anywhere it cannot meet it. Set it wrong in either direction and the campaign behaves in ways that look like faults but are not. The mechanics behind that are in how GMV Max works.
Two starting points cover most Malaysian shops.
Where to start
A brand new product with no sales history: start at 1.1. You are asking the campaign to find buyers for something unproven, with no reviews and no purchase data. A demanding target gives it nothing to work with. Start near break-even, accept that early spend is buying data rather than margin, and raise it once the product has history.
An established offline brand moving online: start at 5. The brand has recognition, the product is proven, and there is usually enough margin to support a real target. Start there and move up if it performs.
The difference between the two is not ambition. It is how much the algorithm already knows.
When it will not spend
Set the target above what your creative can deliver and the campaign simply declines. This is the most common cause of a GMV Max campaign that will not spend.
The fix is to step the target down until spend reaches a rate that uses the budget over the period you intend. Step is the important word. Cut it sharply and delivery can freeze for days while the algorithm reassesses everything. Move it in small increments and wait between moves.
When it spends fully and still misses
This is the case that catches people out, and the answer is the opposite of what instinct suggests.
If the campaign spends its entire budget and the return comes in below target, raise the target. Do not lower it.
The logic follows from what the target actually does. It is the bar a video must clear before the campaign will spend on it. A campaign spending everything at a poor return is finding plenty that clears a low bar – including a lot of weak creative. Raising the bar forces it to be more selective. It will spend less, on better videos, at a better return.
Lowering the target in that situation does the opposite. It widens the pool of acceptable spend and buys more of the traffic that was already underperforming.
Move it slowly, and rarely
The mechanism that makes the target work is also what makes it fragile. Every change sends the algorithm back to reassess what it has learned. A sudden rise, a sudden fall, or changes made every day will scramble that learning and freeze spending altogether.
So: small increments, and leave the campaign alone between them. Set a target, give it several days, read the result, move once. A campaign that is edited constantly never leaves learning, and the operator ends up attributing the resulting mess to the platform.
This is the least popular advice we give, because it asks an operator under monthly pressure to do nothing for days at a time. It is also the difference between a campaign that stabilises and one that never does.
Target ROI is not your margin
Worth stating plainly, because it is a frequent confusion. Target ROI is a revenue-to-ad-spend ratio. It takes no account of your cost of goods, affiliate commission, platform fees, fulfilment or returns.
A campaign hitting a return of 3 may be losing money once everything else is counted. Work out the ratio your business actually needs before you set the number, rather than choosing something that sounds respectable. Malaysian sellers in low-margin categories – food and beverage in particular, where unit prices and gross margins are both thin – need considerably higher returns than sellers in categories with room in the price.
When the target is not the problem
Before adjusting anything, rule out the two inputs that no target setting will fix.
If the creative pool is thin, the campaign has nothing to select between and the target is irrelevant. If the listing converts badly – short description, too few photos, weak buyer feedback – the traffic arrives and leaves regardless of how well the ads are tuned.
Adjusting the target on a campaign with either of those problems is rearranging a symptom.
Common questions
How often should I change the target?
Rarely. Days between moves, not hours. If you cannot remember what you last changed, you are changing it too often.
Is there a right target for my category?
There is a right target for your margin structure. Two shops in the same category with different cost bases need different numbers.
Should I set different targets for product ads and LIVE ads?
Usually yes, because they behave differently and the economics of a livestream session are not the economics of a product page.
We run social commerce management for Malaysian brands. If your campaign is spending without returning, that is a diagnosis worth doing properly. Tell us the brand and the goal.