Programmatic SEO in cold email takes three forms. Data-exhaust directories turn a contact or company database into one page per row (ReachInbox, Clay, Hunter, PlusVibe). Ecosystem directories list tools, agencies or experts (Salesforge, Saleshandy, lemlist). Libraries of templates and glossaries are covered in Glossaries, template and prompt libraries.
Size does not predict results. The largest family in the market, ReachInbox's people directory, belongs to a site Similarweb puts at ~106.9K visits over three months (a panel estimate). Clay's 10,345 English company dossiers are fewer, longer and richer, and in our checks they came up first for their exact queries. The one company that ran a people directory at real scale, Apollo, has switched it off: apollo.io/people/<first>/<last>/<id> and /companies/… now return 410 Gone (our probe, 5 Oct 2026).
Who runs what, and at what scale
Counts come from each site's sitemaps, crawled on 5 October 2026 (scoreboard) and re-pulled in full. "EN" strips language copies, which can double or quadruple a family.
| Site | Family | URLs | What one page holds |
|---|---|---|---|
| ReachInbox | /people/<first>/<last>/<id> | 1,337,500 read from 107 of 1,584 sitemap files; ≈19.7M declared | A named person's title, employer, masked email and phone, list of colleagues; ~200 words |
| Clay | /dossier/ | 10,345 EN + 10,332 FR | Company email format, funding, CEO, HQ or executives; ~3,100 words with FAQ |
| Clay | /gift/ | 3,599 EN (6,835 total) | Personalised ABM gift landing pages, noindex, nofollow |
| Clay | /integrations/ | 1,868 EN | One page per action or data point |
| Hunter | /technologies/ | 6,228: 1,583 tech hubs + 4,646 tech×country | "924 websites using GoDaddy Website Builder in United Kingdom…" |
| Salesforge | /directory/ | 1,637 EN (859 tools, 727 agencies, 45 extensions); 5,599 with de/fr/pt | ~900 words of AI-style feature prose per tool or agency |
| PlusVibe | /number-of-employees/ | 991 unique (each listed twice → 1,982) | Headcount, LinkedIn followers, HQ, then a 100-name link list |
| PlusVibe | /cold-email-tools/ | 753 unique | Tool name in the H1, then the PlusVibe site chrome |
| Saleshandy | /agencies/ | ~340 profiles + 233 service×country pages | Agency partner profile, or "Best … Agencies in |
| lemlist | /experts/ | 143 profiles × 4 languages = 572 | Certified-partner profile |
PlusVibe's sitemap shows 6,084 URLs, but the www and apex sitemaps list the same 3,042 URLs twice. Read PlusVibe's family counts as half of what data/seo.json shows. The brief's "glossary 1,282" is 641 unique pages.
Thin or rich: what the templates contain
We fetched a sample page from every family (5 October 2026).
Clay's dossiers are the richest. Aegon Asset Management Email Format carries industry, headcount band, HQ, founding year, revenue band, the pattern (firstinitiallast@), named executives and a five-question FAQ. It is marked up as FAQPage and ProfilePage, and each fact links into Clay ("Find the current person →"). The families are split by intent. Of the English dossiers, 2,873 are -email-format pages, 2,102 -funding, 835 -ceo and 390 -headquarters-office-locations.
Hunter's tech pages are honest lists drawn from its own TechLookup data (Hunter: search teardown).
Salesforge's directory is the content-farm variant. The LinkedRadar page is about 940 words of numbered feature prose ("Key Features of LinkedRadar 1. LinkedIn Auto Connect…") under the title "Overview (2026) – Features, Pros, Cons & Pricing". Every directory page we sampled, five agencies and tools, declares the same AggregateRating of 4.6 from 132 reviews in its JSON-LD. A rating that is identical on unrelated businesses cannot describe reviews of any of them.
PlusVibe's families are the thinnest. eGain – Number of Employees has about 60 words of data followed by 100 links to other companies, and its title still ends "| pipl.ai", PlusVibe's old brand. Its JSON-LD declares a 4.9 rating from 8,863 reviews for "pipl.ai". /cold-email-tools/pocus has the title "PlusVibe", an H1 with the tool's name and almost nothing else: the rest is navigation and the "pipl.ai is re-branding to plusvibe.ai" banner. Each family went up in one batch: 990 of the 991 headcount pages carry a June 2024 lastmod and all 753 tool pages carry May 2026 (PlusVibe: SEO teardown).
Saleshandy's country pages count on the agency profiles. Best Lead Generation Agencies in Ukraine 2026 is 350 words and lists one agency; its meta description reads "List of best 1 Lead Generation Agencies in Ukraine".
ReachInbox's people pages hold personal data. Our fetch of a sampled profile returned a named employee's job title and employer, masked email and mobile with "verified" badges, a list of named colleagues and two testimonials: about 200 words. curl gets a bot challenge (403), so we cannot tell whether Googlebot sees the same page (ReachInbox: SEO teardown).
ReachInbox's robots.txt lists 1,584 sitemap files on a public storage bucket, each holding up to 12,500 URLs. We read 107 (1,337,500 URLs) and spot-checked seven more. The ≈19.7M is our arithmetic, not a count.
Does it rank?
Our search tool is not a logged-out Google results page, so read these as signs of indexation, not rankings.
| Query | What came back | Read |
|---|---|---|
| "Aegon Asset Management email format" | Clay dossier #1, a second Clay dossier at #4 | Rich programmatic pages win exact-match queries |
| "eGain number of employees" | Revelio, Unify, two Clay /number-of-employees/ URLs, PitchBook. No PlusVibe page | Clay has since retired that family: both URLs now return 404 |
| "LinkedRadar features pros cons pricing" | Capterra's country sites plus two Salesforge directory URLs, one with a ?via= affiliate tag | The directory is indexed for tool long tail |
| "best lead generation agencies in Ukraine" | Saleshandy's one-agency page #1, ahead of Clutch, DesignRush, The Manifest | Thin pages can still rank on low-competition geo queries |
The best first-party evidence comes from Salesforge's content lead in a Surfer case study (Sep 2026): "All the programmatically built pages bring in a ton of traffic into our funnel and allow us to bid on competitor-branded keywords." The same piece reports organic traffic up more than 100% in 12 months. Surfer is an SEO-tool vendor writing about a customer, and the study names no measurement source. A page about a rival tool doubles as a landing page for ads on that rival's brand.
We could not get page-family traffic for anyone. No public source splits Clay's, Hunter's or Salesforge's organic traffic by directory and blog, and we did not reach Google Search Console data or paid rank trackers. Whether PlusVibe's 1,744 headcount and tool pages earn any clicks is unknown.
Indexation risk
Google announced on 5 March 2024 that it would act against "producing content at scale to boost search ranking — whether automation, humans or a combination are involved". It later reported "45% less low-quality, unoriginal content" in results. Site reputation abuse enforcement began on 5 May 2024. The current spam policy gives these examples of scaled content abuse: "scraping feeds, search results, or other content to generate many pages … where little value is provided" and "creating many pages where the content makes little or no sense to a reader but contains search keywords".
Mapped onto the families above:
- High risk: PlusVibe
/cold-email-tools/(keyword in the H1, no content) and the headcount pages (a database row plus a link farm). ReachInbox/people/adds legal risk to search risk. Under GDPR, a public profile of an EU resident built from scraped data needs a lawful basis and Article 14 notice. The CNIL fined KASPR €240,000 over LinkedIn-scraped contacts (Data sourcing law). - Structured-data risk: a sitewide
AggregateRatingstamped on third-party pages (Salesforge's 4.6/132; PlusVibe's 4.9/8,863 for "pipl.ai"; Instantly's 4.8/3,838 on its expert and prompt pages). Google's review snippet guidelines expect ratings to describe the item on the page. The usual penalty is lost rich results. - Lower risk: Clay dossiers and Hunter tech pages. Each page answers a specific question with data the vendor actually holds.
The pruning record matters as much as the policy. Apollo, with by far the biggest people database, now answers its old people and company URL shapes with 410. Clay deleted its /number-of-employees/ family even though search results still listed it. The two strongest data brands chose to cut their thinnest families.
What this means for an entrant
- Do not publish a people directory. For an EU-based company it is a GDPR Article 14 problem before it is an SEO question, and Apollo has already shown the end state: 410. Company-level pages (email format, not people) carry far less risk.
- Build one rich family on data only you hold. The reply desk (The reply desk) creates aggregate data nobody else publishes: reply-classification rates, out-of-office patterns and bounce codes by provider and country. A few hundred "Cold email to
: legal regime, reply benchmarks, provider mix" pages built from your own anonymised data plus the EU/EEA country matrix would match Clay's richness at a hundredth of its scale. - Make every page answer one query completely. The Clay template works because the email pattern, executives and FAQ all sit in one place. Count unique words per page before shipping a family; if the template is most of the page, don't ship.
- Never stamp a sitewide rating into JSON-LD. Put a rating only on pages that show the reviews behind it.
- Use directories as ad landing pages, not as an SEO bet. Salesforge's own quote is about bidding on competitor brands. A 200-page directory of tools you integrate with (Smartlead, Instantly, EmailBison, Clay) supports the overlay positioning in The wedge and gives each paid campaign a relevant page.
- Prune on a schedule. Each quarter, send 410 for template pages with zero impressions after six months, as Apollo and Clay have.