On July 29, 2026, Google published Demand Gen Drop #11 and expanded YouTube Checkout Links into nine new markets — Switzerland, Australia, South Korea, Indonesia, Mexico, France, Poland, Israel, and Argentina. In the same post it pointed retailers toward a second, much heavier path to the same outcome: the Universal Commerce Protocol, which is still United States only.
Those two things sit one bullet apart in Google’s announcement, and a lot of Michigan retailers are about to treat them as the same project. They are not. One is a URL change you can ship in an afternoon. The other is a catalog and payments integration. Picking the wrong one first is how a Q4 budget gets spent driving traffic to a page that still asks the shopper to start over.
What actually shipped in Google’s July 29 Demand Gen Drop?
Three things: Checkout Links went live in nine additional countries, Target ROAS bidding for Demand Gen was upgraded to stop being “overly cautious early on,” and Google published a 2026 Demand Gen Retail guide alongside an affiliate partnerships boost for creator content.
The number Google leads with is a 6% average conversion lift for advertisers who supplied checkout URLs on Demand Gen, measured on US data from September 1–30, 2025. It also cites an Ipsos study it commissioned (October 2025 through January 2026, n=4,391 US holiday shoppers) reporting that 94% of holiday shoppers who used YouTube took a further step toward purchase after watching a related video.
Worth noting separately: on July 13, 2026, Google made YouTube reach and frequency optimization for video campaign groups available globally, citing a Meridian marketing-mix study of roughly 600 US brands using 2023–2025 data that found an optimal frequency of 2.7 exposures per week produced a 19% ROI lift. That matters here because Checkout Links only pay off if the same person sees enough of your video to act.
How do Checkout Links and UCP agentic checkout actually differ?
A Checkout Link is a destination URL. UCP is a transaction. With Checkout Links you hand Google a URL pattern that lands the shopper on your own cart or checkout page with the product already loaded. With UCP, the purchase completes inside Google’s surface — AI Mode, the Gemini app, or YouTube — and your systems receive the order through a protocol integration.
The mechanism that makes Checkout Links work is subtraction. A normal Demand Gen click lands on a product page, where the shopper still has to pick a variant, add to cart, and open the cart. A Checkout Link collapses those steps into the landing itself. So the lift scales with how many steps you removed and how much ambiguity lived in them.
That gives you a way to predict your own result before you build anything. A single-SKU direct-to-consumer brand with a one-page checkout has almost nothing to subtract, and should expect close to zero. A 40,000-SKU aftermarket catalog where the shopper has to confirm year, make, model, and submodel before the cart will accept the part has an enormous amount to subtract — and will see far more than 6%, provided the link can carry the fitment selection with it. If it can’t, the shopper lands on a cart that rejects the item, which is worse than the product page.
Why is that 6% conversion lift smaller than it looks?
Because it is an average across advertisers who chose to implement checkout URLs — and the ability to produce a clean, stateless checkout URL is itself evidence of a well-built cart. Google is measuring a group that was already selected for cart hygiene.
This is the part the announcement doesn’t say out loud. If your cart depends on a session variable — a stored vehicle in a garage widget, a logged-in dealer price tier, a shipping-zone cookie set on a previous page — then a cold Checkout Link arrives without that state. The shopper gets an empty cart, a wrong price, or a fitment error. You will not get 6%. You will get a higher bounce rate on more expensive traffic, and you will blame the creative.
Google also measured this in September 2025, before the nine-market expansion and before the tROAS upgrade shipped. Treat 6% as a floor for clean carts and a warning for messy ones, not a forecast.
Which one should a Michigan retailer build first?
Build Checkout Links first, in essentially every case. UCP earns its integration cost only when your catalog is large, your inventory turns fast, and you are losing the sale to a competitor inside an AI surface you never see.
The decision rule that holds up: if your cart can accept a stateless URL today, ship Checkout Links this month and revisit UCP after the holiday. If your cart cannot, fix the cart — that fix is a prerequisite for both paths, so it is never wasted work.
This lands hard on a specific kind of Michigan business. Powersports and marine dealers in Sterling Heights and Rochester Hills, industrial supply houses in Warren, and automotive aftermarket sellers across Livonia and Southfield all run fitment-critical catalogs where the part number is meaningless without the vehicle or hull it attaches to. A Livonia powersports retailer putting $8,000 a month into Demand Gen through Q4 spends roughly $24,000 delivering shoppers to a cart. The difference between a cart that holds the right part and one that throws a fitment error is not a rounding error on that spend.
How should a small team roll this out?
Seven steps, all of which a three-person in-house team can finish before Labor Day.
- Take your ten highest-margin Q4 SKUs and manually build a cart URL for each in an incognito window. If any lands empty, wrong-priced, or fitment-blocked, you have found your real project.
- Strip session dependencies from the cart. The URL must carry SKU, quantity, and any variant or fitment identifier as parameters — nothing inherited from a prior page.
- Add the checkout URL to your Merchant Center product feed for those SKUs, then expand outward once the pattern proves stable.
- Set up a separate conversion action for checkout-link sessions so you can compare them against product-page sessions rather than pooling both into one number.
- Run two weeks with the old destination and two weeks with the new one on the same creative and budget. Do not change creative mid-test — that is how these tests get thrown away.
- Group your video campaigns and set a weekly frequency target near 2.7 before you trust any of the results. Uncoordinated campaigns will over-serve a small audience and under-serve everyone else.
- Only then evaluate UCP, using the two-week data as the input: if checkout-link sessions convert well but you are still invisible in AI Mode product comparisons, UCP is the next investment.
Step 6 is the one teams skip. If you want the mechanics of coordinating video reach without collapsing your formats, our earlier breakdown of the June 2026 YouTube ads formats and Demand Gen follow-on views covers how the pieces fit together.
What dates should be on the calendar right now?
- August 15, 2026 — Woodward Dream Cruise. Aftermarket and detailing demand across Royal Oak, Ferndale, and Birmingham spikes for roughly a week on either side. A useful live test for cart URLs before the stakes get high.
- August 17, 2026 — Google’s target-based bidding change for budget-limited campaigns takes effect. Campaigns that had been outperforming their targets will drift back toward them. This interacts directly with the upgraded Demand Gen tROAS, and we covered the account-level implications in our note on the August 17 budget-limited bidding change.
- Mid-September 2026 — the practical deadline for feed changes to stabilize before holiday bidding. Structural feed edits made in November get learned against your most expensive traffic.
- November 26–30, 2026 — Thanksgiving through Cyber Monday. Whatever your cart does on November 27 is what it has been doing all year; nobody discovers a fitment bug on Black Friday and fixes it that day.
Frequently asked questions
Do Checkout Links require Merchant Center?
Yes — Google’s Demand Gen retail guidance ties Checkout Links to product feed integration, so the checkout URL travels with the product data. If you are running Demand Gen without a feed, that is the first gap to close. The feed is also what makes your inventory eligible for the AI shopping surfaces you would target later with UCP.
Is UCP available to Michigan retailers yet?
UCP purchasing on YouTube ads is US-only as of the July 29, 2026 announcement, so Michigan retailers are eligible in principle. Google has said a simplified onboarding flow is rolling out through Merchant Center over the coming months. Unless you already have engineering capacity idle, waiting for that onboarding is the reasonable call.
Will the upgraded tROAS bidding change my current targets?
Google describes the change as reducing early caution so campaigns reach their targets faster, not as changing the target itself. Expect faster spend in the first days of a new campaign or a large target edit. Pair that with the August 17 budget-limited behavior change and watch daily pacing closely for the first two weeks rather than judging on week-one cost per acquisition.
Does a frequency target of 2.7 per week apply to every advertiser?
No. That figure comes from a Meridian marketing-mix model across roughly 600 US brands using 2023–2025 data, which means it is an average across large advertisers with broad reach. Treat it as a starting point, not a rule. A regional dealer with a narrow addressable audience in metro Detroit will typically need a lower cap to avoid fatiguing the same households.
The short version
Google’s July 29 drop offered two roads to the same destination and priced them very differently. Checkout Links cost you a URL audit; UCP costs you an integration. For most Michigan retailers the honest answer is that neither one is the real work — the real work is a cart that can accept a cold link and still know which part fits which vehicle. Fix that, and both roads open. Skip it, and you will spend the fourth quarter buying traffic to a page that asks your customer to start over.
Comments are closed