Every swipe file shows you the ad.
This one shows you the page.

The handoff is the moment an ad passes a visitor to a landing page. Each pair here gets one of three ratings. Continuous means the promise arrives intact: the product in the creative is the product on the page, the offer is the same offer, and the register does not lurch. Partial means the promise arrives, but the reader has to do work the ad never warned them about, such as reconciling a renamed sale or hunting the product out of a collection. Break means the promise does not arrive: the thing the ad sold cannot be obtained from what is on screen. The rating is a read on that specific pair on that specific day, not a score for the brand.

The rating reads the first view of the destination, signed out, before anything is dismissed. An overlay costs the rating as much of the ad's promise as it hides, and no more. A cookie bar across the bottom of a page whose product, price and size run are all visible does not change the rating; it belongs in the watchout instead. An email capture that opens over the product, on a page where the thing the ad showed has no counterpart on screen, is a break. The test is never whether a modal exists. It is whether the promise is still obtainable from what is on screen.

A promotion the ad did not make is not automatically a fault. What matters is whether the ad's own offer is still obtainable, and whether the page still makes the ad's argument. A page that carries the offer through, alongside site-wide promotions that do not contradict it, stays continuous. A page that carries it but renames it, changes the number, or surrounds it with offers that argue against the ad's case, is partial. A page where the ad's offer cannot be obtained at all is a break.

The Pattern facet groups the failures into the six named modes on the patterns page, and those six cover 169 of the 226 rated pairs, every Partial and Break except 3 held at Partial only because the destination is inferred. The remaining 54 are Continuous, so they carry no pattern at all rather than a catch-all one: there is no failure to name, and filtering on the absence of an editorial judgement is not a finding. It is the same reason the 27 page teardowns carry no handoff rating.

Pair captured means the ad and the page it clicked through to were both captured first-hand, and the destination recorded is the page the ad pointed at. Homepage destination means the ad was captured and its own recorded destination was followed, and that destination is the brand's homepage because that is where the advertiser points the click, not because the page was picked afterwards. Destination inferred means the ad was captured but the recorded destination is the brand's homepage rather than a click that was followed, so the read is of that page and not of a confirmed click-through; those entries are badged on their card. Ad not captured means the landing page was captured but no ad creative could be paired with it, so the entry carries no rating. You can filter on any of the four under Filter, then Evidence.

A swipe file shows you the ad. This library shows you the ad and the page it clicked through to, captured together and dated to the month, side by side. It is also indexed by what the ad was built to do (lead gen, subscription, conversion, install), by funnel stage, and by persuasion mechanic, rather than only by format or brand.

No. Brands appear because their advertising is being analysed as commentary and criticism. No brand featured here has any affiliation with, or has endorsed, Taylor Sicard Consulting. Brand names and marks belong to their respective owners.

New pairs are added on a rolling basis. Each entry records the date the ad and the landing page were captured, because landing pages change constantly and a pairing is only meaningful as a snapshot of a specific moment.