gtmskills.com

Reference file

Domain

domain.md

Instructions

Lightweight by design. Combine with GTMmotion for the full GTM picture.

Step 1 — Pull traffic data

swan-website-traffic for the domain. Returns SEMrush-style metrics: monthly visits, organic share, paid share, top organic keywords, top traffic-driving pages, traffic trend.

If the company is small or niche, traffic data may be sparse or unavailable. Note that and move on — don't burn fetches trying to compensate.

Step 2 — Classify scale

Use traffic volume to anchor the rest of the analysis:

Monthly visits Likely posture
< 5K Pre-scale, founder-led marketing, sales-led acquisition
5K – 50K Building marketing function, mixed motion
50K – 500K Established demand-gen, content + SEO investment
500K – 5M Mature inbound machine, dedicated growth team
> 5M Brand-scale, likely PLG or massive paid spend

This single number anchors everything else.

Step 3 — Top pages reveal what works

From the traffic payload, look at the top 10 organic landing pages. Categorize:

  • Homepage + product pages dominating → brand-led traffic
  • Blog content dominating → content/SEO machine
  • Comparison pages ("X vs Y") → bottom-of-funnel SEO investment
  • Glossary / definition pages → top-of-funnel SEO play
  • Tool pages / free calculators → PLG funnel into product
  • Case study pages → trust-driven conversion

The top pages tell you what their content team prioritizes — and what's working.

Step 4 — Keyword posture

Top organic keywords from the traffic payload. Look at:

  • Branded vs non-branded — high non-branded share = real SEO investment
  • Category keywords — ranking for the category they sell in? If not, big SEO gap
  • Competitor keywords — ranking for competitor names? Suggests displacement strategy
  • Long-tail vs head — long-tail dominance = content scale; head dominance = brand-driven

One paragraph summary. Don't dump all 100 keywords.

Step 5 — Content depth check

swan-fetch-scraped-url on <domain>/blog or <domain>/resources. Quick read on:

  • Publishing cadence (weekly, monthly, dead?)
  • Content type (long-form thought leadership, short SEO posts, product updates)
  • Author depth (single founder writing, content team, guest contributors)

Stop after one fetch. You're checking maturity, not auditing the blog.

Step 6 — Paid posture

From the traffic payload, paid share matters:

  • 0–5% paid → organic-first, low ad investment
  • 5–20% paid → balanced, performance team in place
  • 20–50% paid → paid-led, heavy spend on demand-gen
  • > 50% paid → ads-dependent, vulnerable to CAC inflation

High paid share is an inferable budget signal — they're spending on growth and may be open to tools that reduce CAC or improve conversion.

Step 7 — Trend matters more than absolute

A site at 100K visits/mo going from 200K is a different conversation than one going from 50K. From the traffic payload, look at 6m or 12m trend:

  • Up sharply → growing investment, momentum to ride
  • Flat → plateau or limit of current strategy
  • Down sharply → algorithm hit, team turnover, real opportunity for refresh

Step 8 — Compose the read

<Domain> — Digital Footprint

SCALE
  - <monthly visits>, <trend over 6m/12m>

ACQUISITION MIX
  - <organic / paid / direct split>

WHAT'S WORKING
  - <top organic pages and what they reveal>

GAPS
  - <missing keyword categories, weak content, paid-dependent>

WHAT THIS MEANS
  - <one paragraph on positioning angle>

Step 9 — Use in qualification

swan-update-company with the scale tier. For SEO/martech sales, a sub-5K-visits company is unlikely to need enterprise marketing tooling; a 5M-visits company likely already has the basics covered and you'll need to position against incumbents.

Rules

  • MUST cite the metric source. Don't say "they get lots of traffic" — say "120K monthly visits per swan-website-traffic."
  • NEVER state traffic figures as absolute truth; SEMrush is an estimate. Use "approx" or "est" for precision claims.
  • NEVER run this on companies that obviously don't have a web presence (very early-stage or pure offline). Note and skip.
  • If swan-website-traffic returns no usable data, say so plainly and stop. Don't pivot to web scraping to manufacture metrics.
  • If a tool result is truncated, read the JSON from files/tool-outputs/swan-website-traffic_<callId>.json in swan-execute-code.