How SiteAssay works
The seven pipeline stages, from a keyword and a city to an approved outreach email.
One pipeline turns a trade and a city into a queue of businesses ranked by how much fixable technical pain they are in, each with a shareable diagnostic report and a drafted email waiting for your approval.
Two steps, not one
Finding businesses and auditing them are separate actions, and that is deliberate. A search returns a list; you decide which of those businesses is worth measuring. Auditing everything a search returns would spend most of a month's allowance on leads you had not looked at yet, and would point an active site scan at businesses nobody had chosen.
Each stage is a queued job. Stages are chained per prospect, so one dead website never stalls the other ninety-nine, and a run that loses a worker resumes rather than restarts.
Discovery: stages 1–3
1. Discovery
You give the engine a keyword and a location, such as "dentists" and "Austin, TX". It queries map and directory sources and search results, normalises the results, and deduplicates them against what your workspace has already seen. A business already in your workspace is recognised, not re-added, and not re-charged.
2. Segmentation
Every discovered business lands in one of two revenue segments. No website businesses are a build pitch: they are trading on a listing and a phone number alone. Technical debt businesses have a site, and the audit exists to find out how badly it is serving them. The two need completely different opening lines, so the pipeline splits here.
3. Enrichment
For businesses with a site, the engine collects published contact addresses and verifies which social profiles actually exist. Everything comes from the business's own pages and public listings. Discovery stops here.
Audit: stages 4–7
Started by you, one business at a time, from a prospect's page, or by pasting a URL if you already know who you want to pitch. Everything from here runs as one chain: measuring a site and not scoring, reporting or drafting on the result would be pointless, and only the measurement itself is expensive.
4. Audit
The audit engine measures the live site: performance, TLS, security headers, accessibility, SEO and schema, AI crawler access, broken links, technology fingerprint, known CVEs, review profile. Results are merged into one canonical payload alongside a tool-status map recording which checks ran, which were skipped and which failed. See the audit engine.
5. Scoring
The payload collapses into ten category scores and one weighted 0–100 composite with a letter grade. Lower is worse for the business and better for you. See scoring.
6. Report
A shareable web report and a matching PDF, written for a business owner rather than a developer: an executive summary with an estimated revenue impact, then each finding with what it means, why it matters and how to fix it. See reports.
7. Outreach
The engine picks the most sellable finding on that specific site, drafts a short message anchored to it, and stops. Nothing sends until you read it and click approve. See outreach and compliance.
Running a pipeline
From the dashboard, Lead Finder takes a trade keyword, a location, and a search limit to discover local businesses; Audit a website takes a URL and runs the full diagnostic suite directly.
Runs are tracked stage by stage, so a run in progress shows you exactly which prospect is at which stage, with complete visibility into every check performed.
Workspaces
Everything above lives inside a workspace: prospects, runs, audits, reports, outreach, suppression list and usage. Nothing crosses between workspaces, including in the aggregate benchmarks published on this site, which are computed over counts and averages only.