A product teardown framework

A reading of a product teardown starts with a question whose subject is visible within the sample’s captured page set. A product teardown reads a product’s public pages and draws each finding from them. Write the question first, then name which of the captured pages the reading covers and where it stops. The Supabase sample, captured on 10 August 2026, supplies a starting observation: a usage calculator sits directly below the pricing tier grid. For a reading exercise using that sample, the question is how the pricing page positions tier information and the calculator relative to each other. The scope is the captured pricing page, not the account area or the result of using the calculator. Although the public sample covers three pages, this question selects the pricing evidence and does not imply a review of every page in the sample. The brief includes the product name, the page under discussion, the sample’s capture date, and the question about placement. It also names what stays outside the exercise: input changes, calculated results, and billing after sign-up. Those exclusions follow from the question and the still image rather than from a judgment about the product. The stopping point is a sourced description of the visible relationship, accompanied by any unanswered questions. A topic that needs another page or an interaction belongs in a separate scope instead of silently widening the first reading.

Evidence stays attached to an observation through page identity, capture date, and the context visible around it. For the Supabase example, the source is the public sample with a capture date of 10 August 2026. The observation describes the usage calculator sitting directly below the pricing tier grid. Keep the sample link and that page context beside the sentence, so a reference to the calculator does not become a statement about every Supabase screen. When working from a capture of your own, include its page URL, a screenshot reference when the finding has one, and any visible state relevant to the question. If the material is public sample text, state that source instead of presenting the exercise as direct inspection of the underlying image. In the notes, preserve the difference between a visible detail, an unavailable source, and an open question. The calculator’s placement is the cited detail here, while its response to an input change remains outside the still capture. The coExploro methodology page sets out the same relationship between a finding and its source. That relationship supports checking a statement about the image, and supplies no evidence for an interaction the image omits.

The review checks whether each sentence stays within the evidence named beside it. Read the observation against the starting question, then examine its subject, verb, and scope. In the Supabase sample from 10 August 2026, the supported statement in this product teardown is that the pricing page places a usage calculator directly below the tier grid. Placement is visible page context. A claim about the calculator’s effectiveness needs evidence that this observation does not contain, so remove that claim and move it to the open questions list. Apply the same check to absence: a feature outside the frame is not evidence that the product lacks that feature. If a sentence names another page, date, or state, check that it has its own source before including it. Where notes conflict, retain the conflict with its source context instead of choosing a description without support. If a quoted phrase is vendor copy, preserve its wording and identify it as page content rather than a verified outcome. End the review when the retained observations answer the scoped question about visible placement and each has its source beside it. The remaining questions mark the boundary of this reading; they do not count as findings about screens or actions that the capture does not show.

A handoff preserves the question, scope, date, and unresolved work alongside the observation. For the Supabase pricing example, the handoff names the 10 August 2026 capture and asks how the tier grid and usage calculator sit in relation to each other. Its answer states that the calculator sits directly below the grid, with the public sample as the source. The scope line limits the exercise to the captured pricing page and identifies the text as a reading of the sample. A separate note states which source is open to the reader, so the next person knows what the public page can confirm. The open question concerns what the calculator shows after an input changes; the handoff does not supply an answer from the still image. Continuing that question requires evidence from the relevant interaction and its context, under a scope that includes both. Keep the capture date attached when quoting the observation elsewhere, because a later page visit is another source rather than confirmation of every past state. Include any missing access or source context so the next reader sees where checking stops. This closes the working sequence without converting page placement into a product recommendation. The handoff ends with a sourced answer to the original question and an explicit question for the next investigation.