A high ROAS can look like a reason to spend more.
The tension between ROAS and incrementality becomes clearer when paid GMV grows but total-store GMV does not follow at the same rate. But there is a point where additional paid investment stops translating into proportional growth in total-store GMV.
That is when I stop asking whether the media is performing well. I ask:
What is actually constraining incremental growth?
GMV is the outcome of a system
In Digital Commerce, a transaction rarely comes from one activity. A consumer may discover a brand through awareness media, arrive on a platform through offsite traffic, and convert through onsite advertising. Along the way, the same transaction may also be influenced by promotions, livestreams, videos, CRM, membership or affiliate activity.
These activities overlap.
Media can sit on top of other growth activities, while different media solutions can also overlap depending on how attribution is assigned. So when a dashboard says that paid media generated a certain amount of GMV, the number can be useful. But it does not necessarily tell us what actually created the growth.
Measure the job, not just the channel
The way I evaluate media changes with the job it is meant to do. At the awareness stage, the objective is to build awareness and visibility. At consideration, the objective shifts toward traffic. At conversion, the focus becomes GMV, conversion rate and ROAS.
That distinction matters because ROAS is a conversion metric, not a universal score for every media dollar. And even when ROAS is high, increasing paid investment does not automatically mean the business will grow at the same rate.
When paid growth stops translating into total growth
Consider a simple situation. Paid GMV is growing. ROAS is high. So we increase ad spend. But total-store GMV does not increase proportionally.
The easy conclusion is: Paid media is working. We should spend more.
I don’t think that conclusion is strong enough. The next question is:
Is the audience pool the constraint, or is something else in the commerce engine?
Those two situations can look similar in a dashboard. But they require very different decisions.
The audience pool may be the constraint. A business can scale faster than its organic audience pool. In that situation, paid media may genuinely be doing its job, but the existing demand pool is not expanding quickly enough to support further growth.
The answer is not simply to buy more traffic. The business may need to create and recruit more demand through broader visibility, brand activity, CRM or other upstream growth initiatives.
Media is not necessarily the problem. The available audience may be. Or the rest of the commerce engine may be underperform in.
The opposite can also happen. Paid media may be contributing normally while other growth levers are contributing less than expected. Organic traffic, content, livestream, CRM, affiliate or commercial activity may not be generating enough demand or conversion.
In that situation, paid media can appear disproportionately important. The answer may not be to optimise ROAS further. It may be to understand what is happening elsewhere in the system.
Why attribution makes this harder
The same transaction can be touched by multiple growth levers.
A consumer might see a brand campaign, engage with content, arrive through an affiliate link, encounter a promotion, watch a livestream and eventually convert through paid onsite advertising.
So which activity created the sale?
Attribution can tell us which activity receives credit under a particular attribution rule. This is where the relationship between ROAS and incrementality matters. It does not automatically tell us which activity created incremental demand.
That is why I find it useful to distinguish between three possibilities.
Demand Creation
The consumer buys through social commerce while maintaining their normal purchasing elsewhere. The transaction represents additional demand.
Channel Shift
The consumer buys through social commerce instead of making the purchase through another channel. The transaction has moved, but total business volume may not have increased.
Demand Pull-Forward
The consumer purchases earlier than they otherwise would have. Current-period ROAS can look strong while some future demand has simply moved forward. The same reported ROAS can therefore represent very different things for the business.
The question I come back to
The question is not simply:
Which channel generated the GMV?
It is:
Would the sale have happened without the media?
That question changes how I think about performance. ROAS is still useful. Attribution is still useful. But neither should automatically be treated as proof of incremental business growth.
The goal is not perfect attribution for every transaction. The goal is to understand what the business needs next.
What this looks like at business level
Over three consecutive years, the broader DCOM business doubled year on year. During that period, ROAS reached 2× the market benchmark and recruitment exceeded 50%.
Those results came from a much broader growth systemgrowth system — including growth strategy, GTM, team capability, daily operations and execution. This measurement and planning work was one part of that system: helping diagnose media’s role and decide where further investment could create value.
That distinction matters.
Because a business does not grow from one media framework. It grows when strategy, GTM, people, operations and execution work together — and when the business can understand what is actually constraining growth.
