Product feed use inside Google Demand Gen campaigns reached 35% in 2026, up from 26% in 2025 and 16% in 2024. The figures come from Smarter Ecommerce (smec), an Austrian software provider and performance agency, and were reported by PPC Land on 27 July 2026. smec's own reading of the data is that feed-driven Demand Gen now competes in the same auction territory as Performance Max, which puts channel overlap on the table for retailers planning Q4.

Before those percentages travel any further, one thing has to sit alongside them: smec sells software in the category it is measuring. That does not make the numbers wrong, but it does make them vendor-reported rather than independent measurement, and a vendor reporting growth in the behaviour its product supports has an interest in the result. Read the figures as a supplier's account of its own book of campaigns.

The sample matters just as much. smec drew on thousands of active campaigns among European retailers, including more than 4,000 Performance Max campaigns. None of it covers Thailand or the wider Asian market. The 35% number describes European advertisers and should not be repeated as if it described Thai ones.

What smec reported on feed adoption

The trend line is the whole story here. Feed use in Demand Gen, as counted across smec's European campaign base:

2024: 16% of Demand Gen campaigns used a product feed
2025: 26%
2026: 35%

That is more than double in two years. smec's conclusion from the trend is that Demand Gen has moved from a creative-led channel into feed-based product advertising, and that the resulting campaigns bid for the same kind of placements that Performance Max is already buying. The overlap concern follows from that, and it is smec's inference rather than a Google statement.

Why overlap gets expensive in Q4

A retailer running feed-based Demand Gen and Performance Max at the same time may be putting two of its own campaigns into the same auctions, for the same products, in front of the same users. In a quiet month that shows up as mild inefficiency. In peak season, when auction pressure and cost per click both climb, the same structural problem costs more per day and takes longer to notice, because rising costs look like normal seasonal behaviour.

The practical version of the question is not "should we run both" but "can we tell them apart in reporting". If Demand Gen and Performance Max both carry feed products and both report conversions, an advertiser with loose measurement and attribution setup will struggle to see which campaign actually earned the sale. That ambiguity is what makes budget decisions guesswork in December.

The ROAS finding is a different study

PPC Land's article also cites a separate piece of research from Fospha and Google. It looked at 25 retail ecommerce brands across fashion, beauty and consumer goods, covering 28 market deployments between 29 October and 16 December 2025. Brands that put 10-20% of their Google budget into Demand Gen recorded double the return on ad spend of brands that allocated under 5%.

Two cautions on that one. It is a different study with a different sponsor and a much smaller sample, so it does not corroborate the smec adoption figures. And it measures budget allocation, not feed use, so it says nothing about whether feed-based Demand Gen overlaps with Performance Max. Twenty-five brands is a small base for a headline ratio, and the research was produced with the platform whose channel it evaluates. Keep the two studies apart when quoting either.

What this means for Thai marketers

Nothing in either study covers Thailand, so the honest position is that the local adoption rate is unknown. What travels is the question, not the percentage.

Thai retailers running Shopping or Performance Max alongside Demand Gen through Q4 and the 11.11 and 12.12 peaks can test their own overlap without waiting for regional data. A few checks that use data an advertiser already has:

Compare the product sets feeding each campaign type. If the same best-sellers sit in both, the overlap is real in that account, whatever the European average says.
Look at impression and conversion movement in one channel when the other changes budget. Cannibalisation usually shows up as a swap rather than a gain.
Check whether reporting can separate the two at product level before peak season, not during it.
Decide in advance which channel owns which part of the catalogue, so a mid-December budget call is not made blind.

For Thai ecommerce teams, this sits with the rest of peak-season planning for ecommerce campaigns and the wider Google Ads account structure. The work is boring and it happens in October, which is exactly why it gets skipped.

Questions advertisers are asking

Does 35% feed adoption apply to Thai accounts?

There is no evidence either way. The smec sample is European retailers, and no Thai or APAC breakdown was published.

Is the overlap between Demand Gen and Performance Max confirmed?

It is smec's conclusion drawn from its own campaign data, not a finding from Google or an independent auditor. Any account can check whether it holds locally.

Should a retailer cut one of the two channels?

The data does not support that call. It supports measuring the overlap first, since the answer depends on catalogue size, margin and how much of each budget is chasing the same products.

The takeaway

A vendor-reported European trend is still useful, as long as it is labelled correctly. Feed use in Demand Gen has grown fast in smec's sample, and the overlap question that follows is worth answering in any account that runs both channels into peak season. The answer will come from the advertiser's own data, not from a percentage measured 9,000 kilometres away.

Relevant Audience works with Thai retailers on Google Ads and ecommerce performance. If Q4 planning is already underway and channel overlap is an open question, that is a good conversation to have before budgets are locked.