SimilarWeb Insights
/similarweb-insightsPass an array of domains and get back total visits, engagement rates, top traffic sources and competitor performance for each, in one call.
Overview
SimilarWeb Insights is a SyncGTM MCP skill that takes an array of website domains and returns SimilarWeb traffic and audience data for each one: total visits, engagement rates, top traffic sources and competitor performance. It is the audience and competitor member of the traffic family, so reach for it when you care where the traffic comes from rather than how much of it there is. Billing is 2 credits per call and not per domain, so an account and its competitors go into a single array and come back compared side by side for the price of one lookup.
What it does
- 1
Takes one parameter, `domains`, and it is an array. Pass ["acme.com"] even for a single domain — a bare string is the most common failure on this tool.
- 2
Returns per domain: total visits, engagement rates, top traffic sources and competitor performance, as one record each.
- 3
Bills 2 credits per call, not per domain, and the docs set no cap on array size, so an account plus three competitors in one array costs 2 credits instead of 8.
- 4
Does not return organic keywords or backlinks (that is `semrush_insights`, one domain per call), site technologies (`find_company_techstack`), ad creative (`b2b_ads_search`), or any contact, email or phone data.
- 5
Has no date range, no country split and no spend figure. `top_traffic_sources` comes back as a ranked list of channels, so it gives you the order, not the percentage split.
- 6
Low-traffic domains sit below SimilarWeb's reporting threshold and come back thin or empty. That is a coverage limit, not proof of zero visitors, so fall back to `find_company_website_traffic` for a provider-agnostic number.
How to use it
- 1
Download the skill file — it saves as similarweb-insights.zip.
- 2
In Claude, open Settings → Capabilities → Skills and upload the .zip (or unzip the similarweb-insights/ folder into .claude/skills/ for Claude Code).
- 3
Connect the SyncGTM MCP server using the setup guide for your client. Sign-in runs through the browser, so there is no API key to paste.
- 4
Run /similarweb-insights and paste your domains as a clean list: bare registrable domains, no https://, no www., no paths or tracking parameters.
- 5
Tune one thing — how many domains go in the array. Put the account and every competitor you want it measured against in the same call. Billing is per call and the docs set no cap on array size, so a long list belongs in one call rather than in tidy batches; each extra call is another 2 credits. The skill reports the call count and credit total before it spends.
Use cases
Competitor benchmark
Put an account and its two closest rivals in one call to see who leads on visits and who is gaining. Gives an AE a defensible number for a displacement pitch.
Channel-mix qualification
Read whether a prospect is paid-heavy or organic-heavy before writing the first line. A paid-heavy site gets the CAC pitch, an organic-heavy one gets the scale pitch.
Inbound lead triage
Separate real demand from noise on last week's inbound queue by checking where each signup's traffic actually comes from. SDRs then work the accounts with a genuine audience instead of all of them.
Pre-call account prep
Pull visits, engagement and traffic sources the morning of a discovery call. Opens the call with a specific observation about their acquisition mix instead of a generic question.
GTM workflow examples
Monday competitive displacement brief
- 1Run `enrich_organization` on the target domain at 2 credits for industry, headcount, funding and technologies — this is what tells you which two competitors are the right comparison set.
- 2Put the account and both competitors in a single `similarweb_insights` call. One 2-credit call returns total visits, engagement and top traffic sources for all three, side by side.
- 3Where the account's top sources come back paid-led, run `b2b_ads_search` on it at 0.5 credits to pull the live ad creative and targeting they are currently paying for.
- 4Find the buyer with `find_people_within_company` at 1.5 credits flat, passing `job_title` as ["VP Marketing", "Head of Growth"], then `find_work_email` at 1 credit and `verify_email` at 0.3 — `find_work_email` finds an address, it does not check that it is deliverable.
- 5Write the opener from the measured gap: quote the visits ratio, which is real arithmetic on `total_visits`, and name the leading channel from `top_traffic_sources`. Do not invent a percentage split — the tool returns ranked sources, not shares.
Outcome: A one-page displacement brief per account: visits versus two named competitors, the channel-mix gap, their live ad creative, and a named contact with a work email checked by `verify_email`.
Triage last week's inbound queue before SDRs touch it
- 1Run `find_company_website_traffic` on every domain in the inbound queue — it takes one `domain` per call at 1 credit, so 20 signups cost 20 credits — and rank them by monthly visits.
- 2Put the top ten domains in one `similarweb_insights` call, 2 credits total, for engagement and top traffic sources. That separates accounts with a real audience from ones that spiked on a single referral.
- 3Run `enrich_organization` at 2 credits on the domains that clear the bar to confirm headcount, industry and funding before routing.
- 4Add the second and third stakeholder at each qualified account with `find_people_within_company` (`job_title` filtered to the buying roles), then `find_work_email` and `verify_email` on each address.
Outcome: A ranked inbound queue with a channel-mix read on the top ten, so SDRs work the handful of accounts with a genuine audience — each carrying visits, traffic sources and a checked contact — instead of every raw signup.
Segment an outbound list by acquisition channel
- 1Source the list with `find_companies` at `limit` 30, filtered on `industries`, `employee_count_ranges` ("51-200") and `countries` — 0.3 credits per company returned, so 9 credits for 30.
- 2Run all 30 domains through a single `similarweb_insights` call for visits, engagement and top traffic sources. Billing is per call, so 30 domains cost the same 2 credits as one; splitting them into three calls would cost 6 for identical data.
- 3Split the list organic-heavy versus paid-heavy, then run `semrush_insights` on the organic-heavy segment only. It takes one `domain` per call at 2 credits, so budget per domain here, not per segment.
- 4Attach the growth or demand-gen owner to each row with `find_people_within_company`, then `find_work_email` and `verify_email` at 0.3 to confirm the address is deliverable.
Outcome: Two segmented outbound sequences with distinct value propositions, each row carrying a traffic figure, a channel mix and a contact whose work email passed `verify_email`.
Prompts
Paste any of these into Claude once the skill is installed and the SyncGTM MCP is connected.
Get SimilarWeb insights for the domains ["acme.com"]. Show total visits, engagement and top traffic sources.Get SimilarWeb insights for ["acme.com", "competitor-a.com", "competitor-b.com"] in one call. Compare total visits and engagement, and tell me which one is growing fastest.For ["acme.com"], break down the top traffic sources. Tell me whether they are paid-heavy or organic-heavy, then write a two-line cold email opener based on the answer.Run find_company_website_traffic on each of these 8 domains (one call per domain, 8 credits), then batch the top 4 by visits into a single similarweb_insights call and tell me which account to prioritize: acme.com, beta.io, gamma.co, delta.com, epsilon.io, zeta.com, eta.co, theta.com.Use find_companies with limit 25 to pull companies in the "Software Development" industry with employee_count_ranges "51-200" in the United States, then run similarweb_insights on the first 8 domains in one batched call and rank them by total visits.Best practices
Always pass an array. ["acme.com"], never "acme.com" — a bare string is the single most common failure on this tool.
Put the whole comparison set in one call. Billing is 2 credits per call regardless of array size and the docs set no cap on how many domains go in, so five separate calls cost 10 credits where one array costs 2.
Screen long lists with `find_company_website_traffic` first — 1 credit per domain, one `domain` per call — then spend 2 credits here only on the domains that survived the cut.
Say "audience", "traffic sources" or "competitors" in your prompt. That phrasing routes the request here instead of to `semrush_insights`, which answers a different question at the same price.
Do not retry a thin result. A domain below SimilarWeb's reporting threshold returns the same thin record on the second call and costs another 2 credits, so switch to `find_company_website_traffic` instead.
Ask for the comparison, not the raw table. "Which of these three is growing fastest and why" is the sentence that goes into an email. Compare on `total_visits`, which is a number; `top_traffic_sources` is a ranked list, so name the leading channel rather than quoting a percentage the tool never returned.