The teams that get results from website-visitor identification are not the ones with the fanciest tooling. They are the ones with a workflow that fits on one page. Five steps, one owner per step, one output per step, one review per week. Everything else is decoration.
This is that playbook. It is deliberately tool-agnostic — every step names the artefact it produces, not the software that produces it, so a two-person team on a spreadsheet and a fifty-person team on a full stack can both run it. The template lines you can copy verbatim.
Step 1 — Identify: publish yesterday’s companies by 9:00
Every morning, one person exports the previous day’s identified companies to a shared location — a Slack channel, a Notion page, a CRM view, whichever the team already opens first. The list carries five columns: company name, domain, pages visited, session count, and first-visit date. Nothing else. If it takes longer than five minutes to produce, the workflow will silently die inside two weeks.
Set the deadline earlier than you need
Publishing at 9:00 means the SDR who opens their laptop at 9:15 sees a fresh list. Publishing "sometime in the morning" means the list is treated as optional. Deadlines create workflows; open ends destroy them.
Step 2 — Qualify against ICP: kill 60% of the list in ten minutes
The identified list is not the meeting list. Before anyone drafts a message, someone with the ICP definition in their head strikes through everything that does not fit. Industry mismatch, headcount too small, wrong geography, existing customer — all out. You are aiming to end this step with 30–40% of the original list.
The template line for this step is one sentence pinned above the list: “Keep only companies that would sign a €X annual contract if the timing were right.” If the answer is no, delete the row without looking at the pages they visited. Time spent on non-ICP is time not spent on ICP.
Step 3 — Prioritise by signals: sort the survivors into three buckets
The qualified list gets sorted by intent — not alphabetically, not chronologically, not by company size. Three buckets is enough: hot, warm, monitor. A rough one-page scoring table (pricing page = hot, comparison page = hot, deep product page = warm, blog only = monitor, return visit within a week promotes a bucket) will match most teams. Refuse to add a fourth bucket.
One trap here: the temptation to make prioritisation an algorithmic exercise. It is not. It is a triage step designed to answer one question — "who deserves a human today?" — in the time it takes to drink a coffee.
Step 4 — Route to an owner: names, not queues
Hot companies get a named owner within the hour. Warm companies go to the SDR queue by region or industry. Monitor companies land in a weekly digest that nobody has to action but marketing reads on Monday to see what content is pulling ICP-fit companies in.
Names beat queues at this step because a name creates accountability. A queue creates the fiction that "someone will handle it" and no one does. Publish an assignment map — industry X to person A, region Y to person B — and pin it in the same shared location as the daily list.
Step 5 — Outreach with context: the six-line message
The last step is the one that gets templated to death, which is exactly why it fails. A signal-based first touch is not a cold email. It is short, specific, references what the person actually did on the site, and asks for a fifteen-minute conversation — not a demo, not a deck, not a discovery call.
A working six-line template: (1) Hi [first name], I saw [company] spent time on our [specific page] this week. (2) Most companies who read that page are wrestling with [specific problem]. (3) One outcome we produce for teams like yours: [one metric, one sentence]. (4) Happy to be wrong about the timing. (5) Fifteen minutes next Tuesday or Wednesday? (6) [signature]. Personalisation is the [company] and [specific page]; everything else stays the same.
The workflow at a glance
| Step | Owner | Effort | Output |
|---|---|---|---|
| 1. Identify | RevOps / Marketing Ops | 5 min / day | Yesterday’s companies list, published 9:00 |
| 2. Qualify against ICP | SDR lead | 10 min / day | Filtered list (target 30–40% remaining) |
| 3. Prioritise by signals | SDR lead | 5 min / day | Three buckets: hot / warm / monitor |
| 4. Route to owner | SDR lead | Automated | Named owner per hot; queue per warm |
| 5. Outreach with context | Owning AE / SDR | 3 min / hot lead | Personalised first touch within 4 hours |
The weekly review that keeps the workflow alive
Every Friday, thirty minutes, three questions. What percentage of hot leads got a first-touch within four hours? Which pages produced the highest reply rates? Which ICP-fit companies did we let slip through qualify? Answer honestly, adjust one thing for next week, and stop. Two changes at once compound their failure modes.
Non-ICP is a discipline
The single biggest predictor of whether a team stays with this workflow after ninety days is how ruthlessly they kill non-ICP in step 2. Teams that keep working non-ICP leads dilute reply rates, burn out SDRs and quietly conclude that "signals do not work." Signals work; non-ICP does not.
Published by
lead.box Team
More articles
See lead.box on your own traffic
Start free — no card, no sales call required. Or book a 20-minute walk-through if you want the guided tour.