Skip to content
Sign inBook a demoStart free trial

Research

We measure the B2B data and GTM tools market and publish the raw data alongside the studies. It is all released under CC BY 4.0, so you can republish any figure, table or chart, commercially too, as long as you credit Cleanlist AI with a link.

We run these studies and show up in most of them, and on several metrics we lose. Each study page says so up front. We publish the data so you don't have to take our word for a count: download the CSV and run it yourself. If you get a different number, send it to support@cleanlist.ai.

Victor ParaschivCo-Founder & CMO, author of the studies below

12 studies, 11 open datasets.

Each entry opens with the result and its number. The chart beside it draws that number, and the CSV lets you check it.

Collected September 7, 20264 studies

  1. No. 08Collected September 7, 2026240 SERPs · 60 AI Mode answers · collected twice

    The shortlist census: what Google recommends across 240 B2B data buying queries

    Google writes an AI Overview on 232 of 240 buyer queries and names a median of 3 vendors in it. ZoomInfo is named in 121, Cognism 73, Clay 66, Apollo 19. Cleanlist AI is named in 6 to 11 across two identical pulls, while cleanlist.ai is the 7th most-cited source domain.

    Read the study CSVQuery-level census plus the run-to-run volatility table, CSV
    Fig. 8 AI Overviews naming each vendor, of 232
  2. No. 09Collected September 7, 202621 cited pages · 37 citing answers · 269 traced sentences

    269 AI answer sentences traced to our pages: 16 name us, 253 name somebody else

    Across 37 Google AI answers that cite a Cleanlist AI page, 269 answer sentences trace back to one of ours. 16 came from a sentence naming Cleanlist AI and 253 did not. Competitor name density does not predict naming, and the correlation runs the wrong way, so cutting rival mentions is not the fix.

    Read the study CSVSentence-level trace with per-page table, CSV
    Fig. 9 Traced answer sentences that name Cleanlist AI: 16 of 269
  3. No. 10Collected September 7, 2026292 AI answers · 42 vendor pricing pages · 376 attributed prices

    Google quoted 376 prices for 42 B2B data vendors, and we checked every one

    Five of the 40 vendors whose pricing page loaded publish a real number for every paid tier. Of 376 AI-quoted price statements, 48.7% confirm against a vendor page read the same day, 21.8% price a vendor that publishes no dollar amount anywhere, and 14.1% are contradicted by the vendor's own page.

    Read the study CSVStatement-level price audit plus the 42-vendor census, CSV
    Fig. 10 376 AI-quoted price statements, checked against the vendor's page
  4. No. 11Collected September 7, 2026240 SERPs · 2,400 classified slots · 296 PAA questions

    Who owns page one in B2B data

    Across 240 buyer keywords and 2,400 organic slots, no page format holds more than 54 of the 240 rank-one positions, and reddit.com holds 174 slots on its own. Cleanlist AI is absent from the top 50 on 149 of the 240.

    Read the study CSVFormat census and domain leaderboard, CSV
    Fig. 11 Buyer keywords where Cleanlist AI is absent from the top 50: 149 of 240

Collected September 1, 20267 studies

  1. No. 01Collected September 1, 2026208 queries · 199 AI Overviews

    The source-to-mention gap: 208 GTM buyer queries inside Google's AI Overviews

    Google's AI Overviews cite cleanlist.ai in 14 of 199 answers and speak the word "Cleanlist AI" in 2 of them. Apollo is named in 49. Being read and being named are close to unrelated.

    Read the study CSVQuery-level dataset, CSV
    Fig. 1 AI Overviews citing cleanlist.ai: 14 of 199
  2. No. 02Exported August 31, 2026827 pages · 1,000 queries · 16 months

    What AI Overviews did to our click-through rate: 16 months of Search Console

    7,458,554 Search Console impressions returned 40,566 clicks, a site-wide CTR of 0.54%. Dividing each page's clicks by the normal rate for its reported position leaves only 18.5% of that impression pool behaving like human search.

    Read the study CSVPage-level and query-level dataset, CSV
    Fig. 2 Impression pool behaving like human search: 18.5%
  3. No. 03Probed September 1, 202660 vendors probed directly

    The GTM MCP server census

    45 of 60 GTM and B2B data vendors ship a live MCP server, 39 of them confirmed on the wire rather than from documentation. Cleanlist AI is one of the 45. Any vendor claiming to be first, Cleanlist AI included, is wrong.

    Read the study CSVPer-vendor probe results, CSV
    Fig. 3 GTM and B2B data vendors with a live MCP server: 45 of 60
  4. No. 04Fetched September 1, 202638 vendors · 24 rankable

    The B2B data pricing index

    38 pricing pages fetched on one date and normalised to a cost per 1,000 verified emails. Six vendors publish no purchasable price at all. Across the 24 that can be ranked the range runs $13.00 to $490.00, a 38-fold spread, and Cleanlist AI is 10th of the 24. Nine vendors are cheaper than we are.

    Read the study CSVPer-vendor pricing dataset, CSV
    Fig. 4 Cost per 1,000 verified emails, 24 rankable vendors
  5. No. 05Fetched September 1, 202638 vendors · every plan page read on one date

    The B2B data free-tier census: what 38 tools actually give you free

    18 of 38 vendors publish no recurring free plan at all. Among those that do, the median allowance is 40 verified emails a month, and only 7 of the 38 let you upload a CSV in bulk without paying. The advertised credit count is rarely the binding constraint.

    Read the study CSVPer-vendor free-tier dataset, CSV
    Fig. 5 Vendors with no recurring free plan: 18 of 38
  6. No. 06Collected September 1, 2026411 queries · 4,027 classified results

    Reddit owns B2B software discovery

    reddit.com ranks on 147 of the 208 GTM queries (71%), and on the broader B2B software set it holds position 1 outright on 30% of them. Rival vendors' blogs take 46.5% of every organic slot measured, while Google's AI Overviews cite YouTube more than any other domain.

    Read the study CSVResult-level SERP anatomy dataset, CSV
    Fig. 6 GTM queries where reddit.com ranks: 147 of 208
  7. No. 07Collected September 1, 202688 SERPs · 22 vendors · 1,651 results

    The alternatives-page economy in B2B software

    Across 88 "[vendor] alternatives" SERPs covering 22 vendors, all 88 carry an AI Overview and all 88 name Apollo. The vendor named in the query ranks top 10 on 60% of its own SERPs and holds position 1 on 7 of the 88. Cleanlist AI is named on 1.

    Read the study CSVResult-level alternatives dataset, CSV
    Fig. 7 Named vendor at position 1 on its own alternatives SERP: 7 of 88

Collected Standing benchmark1 study

  1. No. 12Run July 2026, standing benchmark500 stratified B2B leads

    The Cleanlist AI 500-Lead Enrichment Benchmark, 2026

    On 500 stratified B2B leads run through a 25+ provider waterfall, 98% returned a verified email and 85% a phone number. Single-source providers returned 70% to 80% on email and 30% to 60% on phone against the identical list.

    Read the study Dataset publishes with the study
    Fig. 12 Same 500 leads: the waterfall against single-source providers

How these studies are run

Search results are collected through the DataForSEO SERP API against Google, United States, desktop, in English, unless a study says otherwise. Vendor pricing and plan pages are fetched directly, with a browser user agent for server-rendered pages and headless Chromium for tables that render client-side, and every figure carries the source URL and the date it was captured. MCP servers are probed by sending a real initialize call over the wire, not read off a vendor’s marketing page. Our own traffic figures come from Google Search Console exports. Classification rules are printed on each study page in full, including the ones that make our own numbers look worse.

6 limits apply to every study on this page:

  1. 01

    Most of it is one day. Search results move and prices change, so a re-run a week later will not reproduce every row.

  2. 02

    One country, one device, one language. United States, desktop, English. Results elsewhere will differ, and we don't know by how much.

  3. 03

    AI Overviews are personalised and non-deterministic. The same query can return a different answer, a different set of citations, or no AI Overview at all. Where a study re-pulls queries to measure that instability, it says so and reports the result.

  4. 04

    Code does the matching. Brand names and cited domains are matched by code against fixed rules, so some matches are false positives and some are missed. The rules are published so you can find our errors.

  5. 05

    We picked the queries. Query sets run in the hundreds, and we chose them to cover our own category. They describe one market. They are not a random sample of Google.

  6. 06

    We are an interested party. Cleanlist AI sells B2B data enrichment and competes with vendors named in these studies. Several of the vendors measured also supply data to our own enrichment waterfall. They are measured on the same rules as everyone else.

How to cite

All studies and datasets on this page are licensed Creative Commons Attribution 4.0 International. You may copy, redistribute, adapt and build on them for any purpose, including commercially, as long as you credit Cleanlist AI and link to the study. You don't need to ask first. Copy the line for the study you're citing.

The source-to-mention gap: 208 GTM buyer queries inside Google's AI Overviews
Paraschiv, V. (2026, September 1). The source-to-mention gap: 208 GTM buyer queries inside Google's AI Overviews. Cleanlist AI. https://www.cleanlist.ai/blog/2026-09-01-ai-overview-citation-study-b2b-data
What AI Overviews did to our click-through rate: 16 months of Search Console
Paraschiv, V. (2026, September 1). What AI Overviews did to our click-through rate: 16 months of Search Console. Cleanlist AI. https://www.cleanlist.ai/blog/2026-09-01-ai-overviews-destroyed-our-ctr-search-console-study
The GTM MCP server census
Paraschiv, V. (2026, September 1). The GTM MCP server census. Cleanlist AI. https://www.cleanlist.ai/blog/2026-09-01-gtm-mcp-server-census
The B2B data pricing index
Paraschiv, V. (2026, September 1). The B2B data pricing index. Cleanlist AI. https://www.cleanlist.ai/blog/2026-09-01-b2b-data-pricing-index
The B2B data free-tier census: what 38 tools actually give you free
Paraschiv, V. (2026, September 1). The B2B data free-tier census: what 38 tools actually give you free. Cleanlist AI. https://www.cleanlist.ai/blog/2026-09-01-b2b-data-free-tier-census
Reddit owns B2B software discovery
Paraschiv, V. (2026, September 1). Reddit owns B2B software discovery. Cleanlist AI. https://www.cleanlist.ai/blog/2026-09-01-reddit-owns-b2b-software-discovery
The alternatives-page economy in B2B software
Paraschiv, V. (2026, September 1). The alternatives-page economy in B2B software. Cleanlist AI. https://www.cleanlist.ai/blog/2026-09-01-alternatives-page-economy-b2b-software
The shortlist census: what Google recommends across 240 B2B data buying queries
Paraschiv, V. (2026, September 7). The shortlist census: what Google recommends across 240 B2B data buying queries. Cleanlist AI. https://www.cleanlist.ai/blog/2026-09-07-ai-shortlist-census-b2b-data
269 AI answer sentences traced to our pages: 16 name us, 253 name somebody else
Paraschiv, V. (2026, September 7). 269 AI answer sentences traced to our pages: 16 name us, 253 name somebody else. Cleanlist AI. https://www.cleanlist.ai/blog/2026-09-07-we-write-our-competitors-ai-answers
Google quoted 376 prices for 42 B2B data vendors, and we checked every one
Paraschiv, V. (2026, September 7). Google quoted 376 prices for 42 B2B data vendors, and we checked every one. Cleanlist AI. https://www.cleanlist.ai/blog/2026-09-07-ai-quoted-prices-audit
Who owns page one in B2B data
Paraschiv, V. (2026, September 7). Who owns page one in B2B data. Cleanlist AI. https://www.cleanlist.ai/blog/2026-09-07-who-owns-page-one-b2b-data
The Cleanlist AI 500-Lead Enrichment Benchmark, 2026
Cleanlist AI. (2026). The Cleanlist AI 500-Lead Enrichment Benchmark, 2026. https://www.cleanlist.ai/blog/2026-07-05-b2b-data-enrichment-accuracy-benchmark-2026

To cite the whole programme:

Cleanlist AI. (2026). Cleanlist AI Research: open datasets on the B2B data and GTM tooling category. https://www.cleanlist.ai/research

Run the benchmark on your own list.

On 500 stratified leads, the 25+ provider waterfall returned a verified work email for 98% and a phone number for 85%. See what your own list returns with 14 days of Pro, free: 250 credits, 3 seats, agents, Sequences and Clu in Slack.

Start free trialRead the benchmark

14 days of Pro, free: 250 credits, 3 seats, agents, Sequences and Clu in Slack.