Methodology 2026.11-v2

A diagnosis you can check.

Deterministic checks run first. Every finding points to the evidence behind it. Where we can't see enough, we say "insufficient evidence" instead of guessing. The score is a decision-support signal, not a prediction of success, revenue or funding.

8
weighted dimensions
39
deterministic checks
90
automated score ceiling
100%
reproducible
The pipeline

Seven steps, no shortcuts.

  1. 01

    Collect evidence

    Your public pages, fetched through an isolated, SSRF-safe browser.

  2. 02

    Run checks

    Deterministic checks: the same evidence always gives the same result.

  3. 03

    Interpret

    Plain-language findings, each pointing to the evidence behind it.

  4. 04

    Validate

    Anything not backed by evidence is dropped before scoring.

  5. 05

    Score

    Weighted, versioned scoring with honest handling of thin evidence.

  6. 06

    Freeze

    The report is stored immutably with its methodology version.

  7. 07

    Notify privately

    Only your workspace hears about it. Nothing is published.

Principles

Rules we hold ourselves to.

Honest by construction

A few passing checks are weak evidence, so every dimension starts from neutral and only broad evidence moves it. Automated baselines can't score above 90; higher needs an analyst-reviewed investigation.

Missing evidence ≠ average

A dimension needs at least 50% of its checks to have evidence before it is scored. Unscored dimensions are excluded and reported, never filled in.

Reproducible

Each report stores its methodology version, evidence hashes and check results, so its score can be recomputed exactly. Changes ship as new versions; old reports never change.

Private by default

Only your workspace sees your score and report. It never appears on public pages, in search, in ranking, in our API or in notifications.

What's scored

Eight dimensions, openly weighted.

Weights are a proposed default, to be calibrated with the founding cohort. Confidence reflects evidence type: observed (1), third-party (0.8), founder-reported (0.6), inferred (0.4).

01Value proposition & positioningClarity of what the product does, who it serves and why it is different.15% · 6 checks
  • observed
    Descriptive page titleIf it fails: Write a page title of 10–70 characters that names the product and its core benefit.
    w2
  • observed
    Meta description explains the productIf it fails: Add a 50–170 character meta description stating what the product does and for whom.
    w2
  • observed
    One primary headline (H1)If it fails: Use exactly one H1 that states the product's core promise in plain language.
    w3
  • observed
    Headline is specific, not genericIf it fails: Replace generic headline copy with a specific outcome for a specific audience (e.g. “Ship invoices in one click for freelancers”).
    w2
  • founder reported
    Founder-supplied one-line positioningIf it fails: Add a one-line description (20+ characters) that says what the product does.
    w2
  • observed
    Homepage explains the product in enough depthIf it fails: Add sections that explain features, outcomes and how it works — at least a few hundred words of server-rendered copy.
    w2
02Problem–solution fit signalsWhether the product addresses a clear, meaningful problem for a defined audience.10% · 4 checks
  • founder reported
    Clear problem statementIf it fails: Describe the specific problem the product solves, for whom, and what happens today without it.
    w3
  • founder reported
    Defined target audienceIf it fails: Name the primary audience (role, segment or situation) the product serves first.
    w3
  • observed
    Use cases explained in sectionsIf it fails: Structure the homepage around 2–4 concrete use cases or jobs-to-be-done.
    w2
  • observed
    Public customer evidenceIf it fails: Publish genuine testimonials, case studies or usage evidence (with permission).
    w2
03User experience & usabilityClarity, navigation, accessibility and core-flow coherence on public surfaces.15% · 5 checks
  • observed
    Mobile viewport configuredIf it fails: Add <meta name="viewport" content="width=device-width, initial-scale=1"> and test on phone widths.
    w3
  • observed
    Page language declaredIf it fails: Set the lang attribute on <html> (e.g. lang="en").
    w2
  • observed
    Images have alternative textIf it fails: Give every meaningful image descriptive alt text; use alt="" for decorative images.
    w3
  • observed
    Logical heading outlineIf it fails: Use one H1 followed by H2 section headings to give the page a scannable outline.
    w2
  • observed
    Reasonable HTML sizeIf it fails: Reduce inline payload (large inline scripts/styles/data) so the initial HTML is under ~1 MB.
    w2
04Functionality & reliabilityReachability, broken links, performance and reliability signals within permitted scope.15% · 4 checks
  • observed
    Primary URL responds successfullyIf it fails: Make sure the primary URL returns HTTP 200 to ordinary browsers and well-behaved bots.
    w6
  • observed
    Responds within 2.5 secondsIf it fails: Investigate server response time (caching, CDN, cold starts) — target under 1 second for the HTML.
    w4
  • observed
    Short redirect chainIf it fails: Point links directly at the final URL; keep redirect chains to one hop.
    w1
  • observed
    Sampled internal links workIf it fails: Fix or remove the broken links listed in the evidence.
    w4
05Trust, safety & credibilityTransparency, security/privacy signals, contact/support clarity and policy disclosures.15% · 7 checks
  • observed
    Served over HTTPSIf it fails: Serve the product exclusively over HTTPS and redirect HTTP to HTTPS.
    w3
  • observed
    HSTS header setIf it fails: Send a Strict-Transport-Security header once HTTPS is solid everywhere.
    w2
  • observed
    Privacy policy publishedIf it fails: Publish a privacy policy describing what data you collect and why, and link it from every page footer.
    w3
  • observed
    Terms of service publishedIf it fails: Publish terms of service and link them in the footer.
    w2
  • observed
    Contact or support routeIf it fails: Provide a visible contact, support or help route (page, form or email).
    w3
  • observed
    security.txt for vulnerability reportsIf it fails: Publish /.well-known/security.txt with a contact for security researchers.
    w1
  • observed
    Who is behind the productIf it fails: Add an About or Team page that says who builds and operates the product.
    w1
06Market & competitive contextCategory fit, differentiation, alternatives and market clarity.10% · 3 checks
  • founder reported
    Specific category chosenIf it fails: Choose the most specific category that fits rather than “Other”.
    w2
  • founder reported
    Stated differentiationIf it fails: Articulate what makes the product meaningfully different from the alternatives your audience uses today.
    w4
  • founder reported
    Known alternatives identifiedIf it fails: List the main alternatives (including “do nothing” or spreadsheets) your users compare you with.
    w4
07Growth & discoverability readinessSearch/social presentation, distribution readiness and conversion path.10% · 6 checks
  • observed
    Social share preview (Open Graph)If it fails: Add og:title, og:description and a 1200×630 og:image.
    w3
  • observed
    Twitter/X cardIf it fails: Add <meta name="twitter:card" content="summary_large_image">.
    w1
  • observed
    Canonical URL declaredIf it fails: Add <link rel="canonical"> pointing at the preferred URL.
    w1
  • observed
    robots.txt availableIf it fails: Serve a robots.txt that allows public pages and references your sitemap.
    w1
  • observed
    XML sitemap availableIf it fails: Publish /sitemap.xml listing your public pages.
    w1
  • observed
    Clear call to actionIf it fails: Add one primary call to action above the fold (e.g. “Start free”, “Download”, “Book a demo”).
    w3
08Business & sustainability readinessPricing clarity, business-model coherence and support/maintenance signals.10% · 4 checks
  • observed
    Pricing is discoverableIf it fails: Publish a pricing page (even “free while in beta”) and link it from the main navigation.
    w4
  • founder reported
    Pricing model declaredIf it fails: Declare the pricing model (free, freemium, paid, usage-based, open source).
    w2
  • observed
    Help or documentationIf it fails: Provide help docs, a FAQ or a getting-started guide.
    w2
  • observed
    Signs of active maintenanceIf it fails: Keep visible dates current (copyright, changelog, blog) so visitors know the product is maintained.
    w2
Reading a score

Score bands.

Scores are only comparable within the same methodology version and category. When too little can be seen, the result is "Insufficient evidence" rather than a number. We don't publish score leaderboards. Want a score above 90? Get investigated.

How diagnosis works: our evidence-based method · Science of Apps