Duolingo's search surface is a second product
A language app with an enormous indexed footprint. What the free content layer is doing for acquisition, and what it costs to keep standing.
Topic
16 articles.
A language app with an enormous indexed footprint. What the free content layer is doing for acquisition, and what it costs to keep standing.
The most copied page format in technology. What Apple's product pages are structurally doing, and why the imitations get it wrong.
Speed metrics became a compliance exercise. What the numbers do for rankings, what they do for conversion, and which sites can safely ignore them.
Both publish numbers; only one expects you to buy from the page. What each pricing surface reveals about how the company actually sells.
Most business sites cannot detect the effects they are testing for. How to know when a test is pointless, and what to do instead.
Realistic timelines by starting position, with the leading indicators worth watching in the meantime and the point at which to call it.
Three ways to count visitors, three different numbers. What each one actually measures, what consent banners do to them, and which to trust.
Stripe publishes a headline rate and a long tail of everything else. What the structure of that page says about who it wants and who it tolerates.
Widely called the worst site in travel, and consistently profitable. What the booking flow optimises for, and why the criticism misses the target.
Two challenger banks with visibly different public sites. What each one leads with, who each is trying to reassure, and where the strategies diverge.
A retailer that gives editorial equal billing with product. What the structure of Patagonia's site commits them to, and where the tension shows.
A catalogue of thousands of items, sold to people who mostly know what they want. What IKEA's site optimises for, and what it visibly gives up.
A blank-canvas product has an onboarding problem by construction. What Notion's first-run flow does about it, and where it still loses people.
Basecamp's public site sells a position rather than a feature set. What that buys them, what it costs them, and who it deliberately turns away.
A repeatable method for judging any company from its public surface alone: site, pricing page, onboarding, search footprint. No insider numbers required.
Most website reporting measures activity, not effect. A measurement stack that answers whether the site changed anything, and admits when it cannot.