The name contains a claim.
A demand-generation campaign is not valuable merely because it is labelled “Demand Gen”. The strategic promise is that it introduces or strengthens demand that would not have materialised in the same way without the media.
That is a much harder standard than generating attributed conversions.
Platform attribution can answer the wrong question.
A campaign can report conversions because it touched customers who were already on a path to purchase. That does not make the media useless, but it does make “attributed ROAS” an incomplete test of whether new demand was created.
Likewise, a campaign can create consideration and still look modest in-platform because the customer later searches the brand, returns directly or converts through another campaign.
Define the expected pathway before launch.
Before running a demand-generation test, decide what change would make the campaign strategically interesting. More new customers? Higher category search? Growth in non-brand discovery? Incremental revenue in exposed geographies? More assisted journeys? Stronger conversion from first-time site visitors?
The campaign objective should determine the evidence you collect.
Protect the core while you learn.
New demand activity should generally have a defined test budget and a clear evaluation window. This prevents an immature experiment from destabilising the efficient part of the account and creates permission to stop the test when the evidence is weak.
A useful operating model is simple: core investment for proven demand capture, test investment for new questions, and explicit rules for graduation into the core plan.
Look for movement outside the campaign.
Incremental demand can reveal itself in places the platform does not fully own: branded search, direct traffic, new-customer mix, category traffic, organic searches, assisted conversions, geographic lift or changes in the performance ceiling of the rest of the account.
None of these signals is perfect in isolation. Together, especially when paired with controlled tests, they can create a more convincing causal story.
The takeaway.
The right question is not “Did Demand Gen get conversions?” It is “What changed in the business because we ran it?”