A quarter of paid clicks land on a grid.

Where do consumer ads actually send people in 2026?

Five destination shapes cover all 216 rated ad and landing page pairs in the library, with nothing left over, and two of them take 172 of the 216 between them (The Handoff, 243 published teardowns, August 2026). The product detail page carries 115, 53.2%. The collection or category grid carries 57, 26.4%. After that the tail is short: 25 campaign pages, 15 homepages, 4 retailer listings.

Every pair was captured by hand, the creative as it ran in feed and the page behind the click, opened signed out from a US IP. 216 of the 243 published entries carry a rating on whether the destination kept the ad's promise: 55 keep it, 110 weaken it, 51 lose it. That is the 74.5% base rate everything below is measured against.

Shape is assigned from the URL path, with eleven hand overrides for custom stacks. Each pair gets exactly one shape, so unlike format or placement these shares sum to 100%.

Which landing page shapes do consumer brands actually use?

Just over half of all paid clicks land on a single product's own page, 115 of 216 at 53.2%. A quarter land on a grid of many products, 57 at 26.4%. Campaign pages, homepages and retailer listings account for 44 pairs put together, 20.4%.

The retailer row stays a count rather than a rate: 4 pairs is below the reporting floor. The interesting number is the second row. A quarter of paid consumer traffic in 2026 is pointed at a grid, and a grid is the one shape built to present options to somebody who has not decided yet. The visitor arriving on it has already decided. They clicked a specific thing.

Does the destination shape decide whether the handoff survives?

More than anything else in this dataset. The four shapes with a reportable sample run from 64.0% failure on a dedicated campaign page to 87.7% on a collection grid, a spread of 23.7 points, and the two shapes carrying 79.6% of all traffic sit 19 points apart from each other.

Read the two tables against each other and the ranking flips with the question. By failure rate the campaign page wins at 64.0%. By break rate it is the second worst thing you can do, 9 breaks in 25 pairs. A built page either answers the ad exactly or answers a different ad entirely. There is not much middle ground available to it.

The product page has the opposite temperament: 36 Continuous, 17 Breaks, 62 Partials. Landing on the right object is close to a guarantee against catastrophe and close to no guarantee of anything else.

Why is the collection grid the worst place to send a click?

Because it is where routing failure lives. The grid fails 50 of its 57 pairs, 87.7%, and produces 7 Continuous ratings, 12.3%, against 25.5% across the library. 34 are Partial and 16 are outright breaks, more than any shape except the product page, which has twice the volume.

The taxonomy calls the mechanism the ad sold one thing and the page shows the whole shop: 38 entries, 24 of them breaks. That single mode owns 24 of the 51 outright breaks in the library. The grid and the homepage carry 24 breaks between them, 16 and 8. Those come from different fields, so I am not claiming they are the same 24 entries, but there is not much room for them to be different ones.

Bataleon ran a rider pulling the yellow lace tab on a black boot, under type ending in top performing boots are on sale. The click opens Bataleon Sale, 38 items, all of them snowboards and bindings. Boots are one click away in the nav, which is exactly the click the ad was paid to remove.

Ridge sold the RIDGE 5MM CURB CHAIN under FOR GUYS WHO DON'T DO FLASHY JEWELRY and landed on a page headed MEN'S RINGS & WEDDING BANDS, 28 Results, none of them the chain. Zesty Paws named one probiotic and landed on a Gut Health grid where the advertised powder at $33.97 is the fourteenth tile, five rows down. Gymshark printed THE ARRIVAL TEE over Starting from $22 and landed on THE BESTSELLERS, 293 Products, no price in the first screen.

A grid does not have to be misrouted to fail. Cozy Earth ran a group text about menopause night sweats and the click genuinely delivers sheets. The page title is the single word Sheets, and nothing above the 18 items says cooling, temperature, night sweats or menopause. Right shelf, wrong shop assistant.

Is landing on the product page enough?

It is the floor, not the finish. The product detail page fails 79 of its 115 pairs, 68.7%, six points better than the library and second best in the set. But 62 of the 115 are Partial, so the commonest outcome in the entire dataset is a visitor who arrives at exactly the right object and still loses the reason they came.

Jones Road ran a gel liner ad on one promise: hooded eyes, meet your new liner. The page opens on the same pot, Navy preselected, add to cart at $32, and the description sells glide, staying power and water resistance while all four FAQs cover longevity, brush choice and water resistance. Eye shape, the thing the ad promised to solve, appears nowhere. Nothing is broken. The argument is simply gone.

Boll & Branch printed LINEN SHEETS. COOLER THAN SATEEN. STRONGER THAN BAMBOO, and the page answers cooling with a Temperature line and a hot sleepers callout, then never mentions sateen or bamboo. Branch burned IN MONOCHROME MINT into the frame and opened the correct Verve Chair with the COLOR field reading Galaxy.

Then there is the tax that is not the shape's fault. Parachute routed a duvet ad to the exact Dusk Stripe page, then put a full screen You've Got An Exclusive Discount panel over the title, price, review count and size selector. SKIMS covered the correct collab page with a NEVER MISS A DROP email capture. That mode runs to 36 entries. Getting the URL right earns you the right to have the modal be your problem.

What happens on the two shapes that skip the catalogue?

They produce the best result and the worst break rate in the set, for the same reason. The dedicated campaign page is the strongest shape at 64.0% failure across 25 pairs. The homepage is 80.0% across 15, with 8 outright breaks, 53.3%. Neither shape has a product page underneath to catch a bad link.

The campaign number needs a caveat beside it. Those 25 pairs come from 11 brands, and 8 are Grüns alone, 32.0% of the shape. This is not the market discovering landing pages. It is a handful of operators who build a page per angle, measured against everybody else's catalogue.

When it works it is unambiguous. Blueland ran an autocomplete mock topped by Detergent free from microplastics, closing on Your Search Ends Here., and the page opens on Quit Cleaning with Microplastics. David Protein claimed the most protein for the fewest calories and the page restates it in the eyebrow, 28g protein. 150 calories. 0g sugar.

When it fails, it fails because a built page can be built for a different campaign. Grüns ran a serif editorial card headlined One of the Most Common Mistakes People Make When Cutting Calories, and the click lands on a direct response sale page headlined You Have Nutrition Gaps with a Save 61% button. Nothing in the viewport is an article. Oura ran Oura Ring 4 above 10% off with Verify with ID.me, and the preferred pricing path resolved to a homepage selling Ring 5: no percentage, no ID.me wall, no occupation named.

The homepage fails the same way with less excuse. Warby Parker introduced a named sunglass, MEET BIX, and landed on a homepage leading with New metals that raise the bar.

Which destination shapes does nobody actually ship?

Four plausible shapes were searched for across all 216 rated pairs and not one appears as a destination: an advertorial or listicle article, a quiz or product finder, an app store listing, and a link in bio page. Zero, in a library where five shapes account for every single click.

That is the finding I did not expect to write. The advertorial has been the most discussed paid social destination of the last three years, and nobody here sent a click to one. Neither did anybody send one to a finder quiz, the shape most CRO decks open with.

The quiz does exist here. It is never the destination. Craighill sent a PRE-ORDER NOW. STOCK IS LIMITED. ad to the correct Swingtop Lighter page, then covered the hero, the price and the buy button with a full screen YOU'VE GOT A MYSTERY OFFER panel asking the visitor to pick a category first, Daily Carry, Desk, Bar and Puzzles & Play among them. A quiz bolted onto a product page, turning the urgency the ad built into an interest survey.

The gap between what the industry discusses and what it ships is the useful part. The shapes that get conference sessions are not in the data, while the shape nobody discusses takes a quarter of paid traffic and loses 87.7% of it. The case for a page per angle is not that advertorials are hot. It is that those 25 pages are the only shape under 65%, and the six ways a handoff fails are all things a purpose-built page can be checked against before launch.

What should you change first?

In the order the numbers justify. The first is free, takes an afternoon, and is worth up to 23.7 points on its own, more than any creative decision measured here.

1. Repoint every creative that names one product and links to a collection. That mode sits behind 24 of the 51 outright breaks. Send it to the product's own URL, or to a hand-sorted grid where the advertised item is the first tile.
2. Send homepage traffic only from ads that ask for nothing. 8 of 15 homepage pairs are breaks, and both that worked were brand ads with no product and no button.
3. Put the ad's strongest line on the product page before you launch. 62 of 115 product page landings are Partial. The right object buys you nothing if the page reverts to the evergreen argument on arrival.
4. Suppress every modal for paid arrivals. The overlay mode runs to 36 entries and hits correctly routed pages hardest, because those are the ones with something worth covering up.
5. Build a page per angle only if you will maintain it. Campaign pages score best at 64.0% and break at 36.0%. A built page serving last month's campaign is worse than a product page, and 9 of the 25 here were that.

One item is deliberately absent: test more templates. Nothing here says the winning shape is one nobody uses. Most brands are choosing between two shapes they already own and picking the worse one a quarter of the time. Every figure above comes from the same 216 pairs, each open to read with its ad and page captures attached. To see the read on your own funnel rather than someone else's, the free store audit does it in a few minutes.

Common questions

The five questions this cut of the data gets asked most, answered with counts rather than a preference. Every figure comes from the same 216 rated pairs.