Depth Review: weak clarity and performance (56/100)

Depth is an AI-assisted product analytics and user-feedback analysis product reviewed from a partial five-page crawl.

Fact	Evidence
Domain	depth.io
Category	AI-assisted product analytics and user-feedback analysis
Pricing model	Freemium
Stated price range	150–600 USD
Pages crawled	5 of 40
Crawl date	2026-10-09

Depth scores 56/100 because its product category and pricing are identifiable, but the captured experience leaves too much of the evaluation path unresolved. The pricing page presents four tiers and a broad audience ladder, while the sampled articles provide concrete product writing, including coverage of Bank of America and product analytics.

The main weaknesses are functional clarity, speed, and decision support. The homepage renders only 14 words from 467 raw words and reports a client-side exception, limiting first impressions and usability. A sampled article reaches a 10.9-second mobile LCP against the 2.5-second good threshold with a roughly 3.6 MB payload. The pricing surface lists prices and audience labels without the decision axes or defended recommendation needed to choose confidently.

Depth’s category is identifiable, but the homepage does not reliably make its audience or differentiated promise clear. The recorded H1 is "AI Product Manager", while the meta description names analytics, reports, user-session review, feedback analysis, and feature ideas; the homepage capture also reports a client-side exception. Pricing provides firmer evidence: Free is $0/mo, Startup $150/mo, Growth $600/mo, and Enterprise is contact sales, with six customer logos. The adjacent efficiency claim has no named result or measurement definition in the supplied evidence. Repair the hero first, name the primary product-team buyer, and define the metric behind the claim with a customer outcome.

A prospect can identify Depth as an analytics and product-feedback tool, but the supplied capture does not yet answer whether it fits a particular team or changes its workflow. "AI Product Manager" is memorable; the meta description is more concrete, naming automated analytics, reports, user-session analysis, and feedback analysis. The pricing page gives the clearest audience ladder: Free for small startups or individual developers, Startup for growing startups and small businesses, Growth for mid-sized companies, and Enterprise for larger needs. Usage limits, detailed feature differences, customer outcomes, and a credible comparison frame are not supplied. Connect the hero label to one product-team job, then add plan limits, included capabilities, and one measurable customer story with context, baseline, result, and timeframe.

The homepage is not a dependable entry point: its capture reports a client-side exception and an interactive inventory with no links, buttons, forms, or contact details. That prevents a first-time visitor from reliably understanding the offer, reaching pricing, or beginning conversion. Once reached, pricing has strong information scent through Free, Startup, Growth, and Enterprise tiers, but its exposed actions are Book time and Contact Us; no signup or trial form is shown. The blog path is structurally straightforward, while the sampled mobile article records 10.9 s LCP and 12.0 s Time to Interactive. Fix the homepage exception, expose primary navigation and a clear CTA, and replace the three # policy destinations on pricing with working URLs.

The sampled article pages provide a reasonable semantic baseline, but the homepage lacks the structure needed for orientation. Article audits report header, nav, and main landmarks, no missing image-alt attributes, and no positive tabindex values. By contrast, homepage DOM facts report landmarks=[], lang=null, headingSequence="h2", and hasSkipLink=false. Pricing reports four h3 headings without a preceding h1 or h2. CSS checks also found three animation declarations and zero prefers-reduced-motion rules. Add lang="en", one descriptive homepage h1, header/nav/main/footer landmarks, and a skip link; restore pricing hierarchy; then add a reduced-motion override. These are focused structural fixes, not a reason to discount the useful article semantics.

Depth has a responsive layout baseline, but mobile interaction and media delivery weaken the execution. The mobile facts report scrollWidth390=390 and a sane viewport meta, while the font inventory is limited to two Geist families with swap loading. Pricing reports 16 small tap targets; footer links are 16px high and several logo links are 80x40px, below a comfortable touch target. The Bank of America article records a 10.921-second mobile LCP on a 3,602 KiB payload, with 2,711 KiB of estimated image-delivery savings. CSS contains three animations but no reduced-motion rule, and homepage heading extraction includes wrapping-script text. Raise target areas to at least 44px, optimize the LCP image, add dimensions, and keep runtime text out of semantic headings.

The sampled article reaches LCP in 10.9 seconds in the mobile lab run, far above the 2.5-second good threshold. This is lab evidence; no CrUX field metrics returned, and homepage PSI failed. The clearest measured cause is payload: 3,602 KiB total, with 2,711 KiB of estimated image savings. PSI also estimates 133 KiB of unused JavaScript, 150 ms of render-blocking savings, and unsized images. Separately, the homepage diff records 467 raw words versus 14 rendered words and a JS-only share of 0.714. Prioritize LCP request discovery, right-size imagery, reserve image space, and defer non-critical JavaScript. Re-test against LCP ≤2.5 s, INP ≤200 ms, and CLS ≤0.1.

Depth’s strongest writing names concrete product and interface detail; its pricing copy is much harder to scan. The Bank of America article opens with a specific problem and names Activity Center, Alerts, Bill Pay, and Transfers. The product article frames a useful workflow around analytics, session review, feedback, and action. Pricing compresses 223 words into one paragraph with an average sentence length of 55.8 words. Its line "Spend ten times less time looking at analytics and more time iterating" also needs a grammatical rewrite. In the article recommendations, phrases such as "significantly enhance the user experience" and "more intuitive and user-friendly" should be replaced by the interface change and user task. Give the homepage, blog index, and pricing page distinct metadata.

The pricing page is easy to scan, but it lists four price states without the decision axes needed to choose confidently: Free at $0/mo, Startup at $150/mo, Growth at $600/mo, and Enterprise at Contact Us. Audience labels indicate scale, yet the supplied evidence does not show explicit tradeoffs, a recommended path, feature or capacity axes, or the threshold where one tier becomes preferable. The extracted page includes a Monthly/Yearly toggle, but the sampled text does not establish annual values or billing basis. Add a concise Choose this if line for every tier, including a trigger, benefit, and tradeoff. Then expose session capacity, analytics scope, seats, integrations, support, overage rules, and clear billing inputs, while keeping recommendations overridable.

The sampled copy is publishable, but a stricter editorial QA pass would catch three practical defects. Pricing places 223 words in one paragraph and averages 55.8 words per sentence, making plan comparison harder than it needs to be. The Bank of America article names useful screens and controls, then repeatedly falls back to "significantly enhance the user experience" and "more intuitive and user-friendly" instead of specifying the interface change and task. Finally, the crawler reports duplicate title and meta-description values across three sampled pages. Separate plan information, benefits, and calls to action into shorter blocks; rewrite recommendations around concrete tasks; and add duplicate-title and duplicate-description checks before publishing.

The five-page crawl did not establish a public documentation or self-serve help surface, leaving onboarding and evaluation support unresolved. This result is scoped: 5 of 40 pages were crawled, so login-gated or externally hosted help may exist outside the observed surface. Even so, Depth should expose a visible Help or Documentation link, a concise getting-started guide, task-oriented answers, an FAQ, and an escalation path.

Depth has a sound crawl baseline, but client rendering is the main technical SEO risk. Sampled pages return HTTP 200, robots.txt permits the public surface, the sitemap contains four URLs, HTTPS is enforced with HSTS, and the nonexistent-URL probe returns a genuine 404. The homepage diff is less reassuring: 467 raw words become only 14 rendered words, with js_dependent=true. The HTTP-to-canonical-host path also takes two 308 hops, while the blog trailing-slash variant takes one. Server-render primary homepage copy and metadata, redirect HTTP directly to the final HTTPS www URL, and use the canonical slashless blog URL consistently. Finally, write unique titles and descriptions for the homepage, blog index, and pricing page.

Depth is clearest when it states its analytics and feedback focus, shows a four-tier pricing surface, and publishes concrete article material. Those strengths give product teams enough context to identify the broad offer and intended audience.

The evaluation path remains weak. The homepage's 14 rendered words and reported client-side exception make the core promise difficult to verify. Mobile performance is also a material barrier: the sampled article records a 10.9-second LCP and roughly 3.6 MB payload. Pricing needs comparison criteria and a recommendation. The evidence supports further evaluation before a product team chooses a plan. This conclusion is based on a partial five-page crawl, not the entire domain.

This 13-dimension review uses the five pages crawled on 2026-10-09, the public score table, and the supplied crawl summaries. It covers positioning, messaging, usability, accessibility, design execution, performance, writing quality, decision-support surfaces, editorial QA, documentation and self-serve help, and technical SEO.

Review-content integrity (dimension 13) was excluded as not applicable. Risk & stability (dimension 17) was excluded because its check failed. Google Search Console was not connected, so no search-performance enrichment is included. The homepage rendering result and mobile lab measurements describe the sampled crawl rather than live user testing. Read How SiteList scores for the framework and scoring context.