Heap

Heap’s founding idea inverts the usual tracking-plan workflow: its SDK autocaptures every click, form submit, and pageview from day one — including single-page-app navigation triggered by pushState, replaceState, popstate, and hashchange — and analysts later define events on top of that raw interaction stream, retroactively, against data collected before the event existed. Labeling is point-and-click via a visual labeling tool, so non-engineers can turn “clicks on this button” into a named event without a code deploy. That removes the instrument-wait-analyze loop for common questions, at the cost of heavier collection volume and event definitions that lean on the page’s markup and can drift as the product’s UI changes.

Implementation is a single snippet placed before the closing </head> tag; the current SDK sends data via POST requests rather than the classic pixel-style GETs. Autocapture pulls text from page elements by default but excludes the contents of input fields; if even element text is too sensitive, a disableTextCapture setting turns it off globally. Alongside autocapture you still get a conventional API — identify() for user identity, addUserProperties() for persistent traits, and track() for deliberate custom events — which most serious deployments end up using for anything autocapture can’t see (server-side outcomes, revenue, subscription state). Other switches worth knowing about: cookie vs. localStorage persistence, cross-domain session stitching via a supportedDomains URL-decoration option, secure-cookie hardening, and a disableSessionReplay toggle.

Since its 2023 acquisition by Contentsquare, Heap sits inside a broader digital-experience suite alongside session replay and experience analytics; replay is sold as an add-on from the Pro tier. Warehouse sync (Heap Connect) pushes the event stream to Snowflake, BigQuery, Redshift, or S3 — one destination of each type per environment — on a default 24-hour cadence (some destinations support 2-hour syncs), with exports as retroactive as the dashboard itself. Note that Heap Connect is gated to the top Premier tier, so “we’ll just model it in the warehouse” is not a cheap escape hatch here the way it is with some competitors.

Pricing is session-based: the free plan covers up to 10k monthly sessions with 6 months of data history, and exceeding the cap blocks you from viewing data until you upgrade — worth knowing before you put it on a high-traffic marketing site. Growth and above are quote-based. Migrating away is harder than migrating in: retroactive, autocaptured data doesn’t map cleanly onto tools that expect explicit instrumented events, so plan on rebuilding a tracking plan if you leave.

Use it when product and marketing teams need retroactive funnel and behavior analysis without maintaining an upfront tracking plan for every question. Look past it when you want strictly governed, intentional event schemas, predictable costs at high interaction volumes, warehouse access without a top-tier contract, or an open, self-hostable pipeline.