The B2B Buyer Journey Funnel: A Behavior-Based Stage-Scoring Framework

You can tell which buyer-journey stage a visitor is in without knowing their name. The signal isn't a form fill, it's the combination of which pages they hit, how many times they've come back, and how confidently you can identify the account behind the visit. Most "buyer journey funnel" content assumes you're already working a known contact through email opens and call notes. That assumption breaks for roughly 95% of B2B traffic, which never fills out a form at all.

This is a refresh of one of our oldest posts. The original version listed four generic funnel stages. This version replaces that with a scoring framework you can actually run against real visitor data this week.

Why Funnel Stage Has to Start With Anonymous Traffic, Not Known Contacts

Classic funnel models (Awareness, Consideration, Decision, Purchase) were built for an era when engagement meant email opens and gated content downloads. They still work as labels. What's missing is how you assign a visitor to one of those labels before they've told you who they are.

Website visitor identification changes the input. Instead of waiting for a form, you get two things the moment someone lands on the site: a resolved company (and often a resolved person) behind the session, and a full record of what they looked at and for how long. Stage scoring is just the discipline of turning that behavior into a consistent, repeatable read on buying intent instead of a gut call.

Which Pages Actually Signal Buying Intent?

Not every page view carries the same weight. Treat page-level behavior like a ranked signal, not a binary "engaged or not":

🔴 Very High - Pricing page, "book a demo" page, or a comparison/alternative page (e.g. "X vs Y") visited more than once in a session or across repeat visits.
🟠 High - Product or feature pages, integration docs, and case studies matched to the visitor's industry or company size.
🟡 Medium - Multiple blog posts across different topics in one visit, or a return visit within 7 to 14 days after a first touch.
🟢 Low-Medium - A single blog post or resource page, no return visit yet.
⚪ Low - One page, short session, no return visit, entering from a broad organic or social query.

Map those tiers to funnel stage this way: ⚪ and 🟢 map to Awareness, 🟡 maps to Consideration, and 🟠 and 🔴 map to Decision. Purchase is no longer a website behavior question at all, it's a CRM stage.

Does Identification Confidence Change the Stage You Assign?

Yes, and this is the part most funnel content skips entirely. A 🔴 Very High signal from a session you can only resolve to the company, not the person, is a different action than the same signal from a session resolved down to a named buyer. Person-level identification typically resolves a smaller share of sessions than company-level identification (Knock2's own published benchmarks: 93%* of engaged sessions resolve to a company, 62%* resolve to a named person, U.S. traffic).

Practically: a company-level-only Decision-stage signal should trigger account research and a multi-thread outreach attempt, not a single 1:1 email to a name you don't have. A person-level Decision-stage signal can go straight to a rep with the name attached. We cover this distinction in more depth in Person-Level vs. Company-Level Website Visitor Identification.

What Should You Actually Automate at Each Stage?

Stage scoring is only useful if it changes what happens next. Here's the pattern that holds up in practice:

Awareness (⚪/🟢 signals): No sales touch. Feed the account into a nurture or ad-retargeting audience at most. A rep reaching out here reads as cold, not warm.
Consideration (🟡 signals): This is where routing rules earn their keep, sending the account to the right rep by territory or segment before intent cools off. See Lead Routing Rules for Identified Website Visitors for the rule architecture.
Decision (🟠/🔴 signals): Time-sensitive. This is exactly the trigger condition covered in The BDR Playbook for Website Visitors, score and route fast, follow up while the pricing-page visit is still fresh, and treat repeat Decision-stage visits from the same account as a buying-committee signal worth escalating.

None of this requires a bigger team, just a scoring rule tied to page behavior and identification confidence, feeding a routing rule, feeding a Slack alert or sequence enrollment. It's the same mechanic behind the six core plays we lay out in Six Plays for Turning Anonymous Traffic Into Pipeline.

What Your Website Traffic Can't Tell You About the Buyer Journey

Here's the honest caveat most vendor content in this space leaves out: your website is one surface, not the whole journey. A meaningful share of B2B research now happens somewhere you can't instrument at all, in peer Slack groups, review sites, competitor comparisons, and increasingly inside AI answer engines. A visitor who lands directly on your pricing page with no prior blog history isn't necessarily "skipping the funnel." They likely did their Awareness and Consideration-stage research somewhere else entirely.

The practical implication: don't treat a thin on-site history as evidence of low intent. Treat identification confidence and on-site behavior as a floor, not a ceiling, on what you know about where an account really is.

How to Build This Without a Data Team

You don't need a custom model to start. A simple, defensible starting rule set:

Score = page-tier weight (⚪ 0, 🟢 1, 🟡 2, 🟠 3, 🔴 4) x visit-recency multiplier, summed across the account's sessions.
Set a routing threshold that matches your team's actual capacity, not an arbitrary "lead score of 80."
Route company-only matches to research/multi-thread motions, and person-level matches directly to a named rep.
Re-score on every new session so a dormant account moving from ⚪ to 🔴 gets caught immediately, not at the next weekly review.

This is the same logic behind Knock2's lead scoring and workflow automation, resolve the visitor, score the behavior, route on the result, without a data team building it from scratch.

Frequently Asked Questions

Q: How is this different from a standard MQL lead score?
A: MQL scoring almost always starts after a form fill. This framework scores anonymous, pre-form behavior so you can act before a visitor ever converts, which is most of your traffic.
Q: Can you assign a funnel stage with company-level identification only, no name?
A: Yes. Page-tier and recency signals work at the account level. What changes is the action, company-only Decision-stage signals should trigger research and multi-thread outreach, not a 1:1 email to a name you don't have.
Q: What's a realistic identification rate to plan around?
A: Knock2 publishes 93%* account-level and 62%* person-level identification, both measured against engaged sessions (10+ seconds or 2+ pageviews, U.S. traffic for person-level). Actual results vary by traffic profile, geography, and industry.
Q: Do repeat visits matter more than the specific page visited?
A: Both matter, but they're not interchangeable. A single pricing-page visit outranks several repeat visits to a single blog post. Recency and repetition mostly move an account up within a stage, they rarely override what the page itself signals.

*Identification rates measured against engaged sessions (Google Analytics definition: a visit lasting 10 seconds or longer, or including 2 or more pageviews). Results may vary by traffic profile, geography, and industry.