Leaks at Arrival: When the Wrong Visitors Arrive

When visitors leave in the first seconds, suspect the arrival before the page. Two causes dominate: the audience is wrong (the ad, the ranking or the link brought people the offer does not serve), or the promise is broken (the right people arrived and the page's first screen does not confirm what they came for). Separate them by source: if one source bounces everywhere, it is the audience; if every source bounces on one page, it is the promise or the page. Fix audiences in the targeting and the copy that fetches them; fix promises on the page's first screen. Redesigning a page to treat an audience problem is the most common wasted month in small-store conversion work.

Audience problems show a pattern: the bounce is concentrated in one source, one campaign, one keyword cluster or one geography, while other traffic converts on the same pages as usual. Promise problems invert the pattern: all sources degrade on the affected page, often from the date the ad, email or snippet changed. Tracking problems mimic both: a mistagged campaign dissolves good traffic into 'direct' and the source view lies. Check the tagging before the targeting; the cheapest arrival fix in the catalogue is a corrected campaign tag restoring your ability to see.

Keyword drift (your page ranks for a neighbouring query with different intent), broad ad targeting that outruns the offer, a viral or directory mention in the wrong community, an affiliate or influencer describing the product to their audience in terms you would not use. Each brings visitors whose honest verdict on arrival is 'not for me', and no page survives that verdict - nor should it; a page that could convert the wrong audience would be lying to someone. The repairs live upstream: negative keywords, tighter audiences, corrected descriptions, and where the wrong traffic is large and permanent, a routing page that splits audiences kindly rather than a homepage that serves neither.

When the audience is right and the bounce is high, walk the chain the visitor walked: the ad or result they saw, then the first screen they met. The landing bench's message match audit is the instrument. Common breaks: the advertised product is below the fold or out of stock, the price on the page differs from the price in the ad, the page leads with a seasonal campaign the ad never mentioned. Each break is small in the telling and decisive in the first three seconds, which is where arrival leaks live.

A class of arrival loss never sees the page at all: the visitor abandons during load, and analytics records either a bounce or nothing. Slow first paint on the landing pages that receive paid and social traffic masquerades as an audience problem in every report except the speed field data. Check LCP for the top landing pages before concluding anything about audiences; the load bench's methods apply, and the fix order is images, third-party scripts, then theme weight.

Work the causes in the order they can be known: verify the tracking, read the source split, audit the promise chain for the worst pages, and only then touch page design. Log which cause each repair addressed; the log entry is what stops next quarter's arrival dip from being misdiagnosed as this quarter's page. Arrival is the stage where discipline saves the most money, because its misdiagnosis is paid for in full redesigns of innocent pages.

Arrival problems recur because landing pages accumulate: every campaign, partnership and seasonal edit adds destinations nobody owns. Keep a simple inventory - each significant landing page, the sources meant to feed it, its baseline by source, and the date its promise was last audited against those sources. Review it monthly in ten minutes. The inventory is what turns arrival diagnosis from archaeology into lookup: when a source's quality shifts, you can name the pages it feeds; when a page's rate falls, you can name the promises made on its behalf. Stores without the inventory rediscover their own site quarterly, at campaign prices.