Someone on your pricing page has already decided they want the thing. They are working out whether they can justify it, and the ones who leave without acting are the most expensive audience on your site, because you paid to get them that far. Most teams measure the page and cannot see the stall, because the stall is a pattern across visits rather than an event within one.
This is about the four ways qualified leads get stuck, what each looks like in behavioural data, and why the page-level view that most analytics gives you is structurally unable to tell them apart.
The newsletter
Join our KISS newsletter
One short read a week on what actually moves revenue, in a free email. Read by 10,000+ operators and founders.
No spam. Unsubscribe in one click.
I.A pricing page visit is the strongest intent signal you get
And the standard reporting throws away almost everything that makes it useful.
A.Why the page-level view fails here specifically
Look at a pricing page in most analytics and you get pageviews, time on page, exit rate. All three treat every visit as equivalent and independent, which is exactly wrong for this page.
A first visit from an ad is idle curiosity. A fourth visit, nine days later, from a direct URL, by someone who has also read two case studies and the security page, is a person trying to build an internal argument. Those are opposite situations and page metrics average them into one number.
The information is entirely in the sequence, and the sequence is the thing session analytics discards.
What one stalled buyer looks like across three weeks
Activity report viewB.What a stall costs compared with a bounce
A visitor who leaves your homepage cost you a click. A visitor who leaves your pricing page cost you a click, everything you spent moving them through consideration, and the opportunity of a customer who wanted to buy.
They are also the cheapest group to recover, because no persuasion is required. They are already persuaded. Something specific is in the way, and the entire task is identifying what.
II.Four stalls, and how each one looks in the data
They are distinguishable by behaviour, and they need contradictory fixes.
A.The unit of pricing is unclear
The buyer cannot work out what they will be charged. Usage-based pricing does this constantly: priced per event, per seat, per workspace, per tracked user, with no way to estimate their own number.
Behavioural tell: long dwell on the page with heavy interaction on any calculator or plan-comparison element, then exit. Repeated returns to the same tier rather than comparison between tiers. Often a detour to documentation to find out what the unit even counts.
Fix: make the unit estimable before commitment. A calculator with realistic defaults, or a stated typical usage for a business of their size.
B.The tier boundary is in the wrong place
They need one feature from the tier above and cannot justify the whole jump. This is the most common stall in B2B SaaS and the most expensive, because these are your best-fit customers hitting a packaging decision rather than a price objection.
Behavioural tell: oscillation. The same person moving between two tiers repeatedly within a visit, expanding and collapsing feature lists, returning across multiple visits to compare the same pair.
Fix: move the feature, or sell it as an add-on. Not a discount, which solves a price problem they do not have.
C.There is an unpriced requirement
Something they need is not on the page at all. SSO, a data residency guarantee, an SLA, a specific integration, a compliance document. They cannot tell whether it exists, costs extra, or is impossible.
Behavioural tell: the detour. Pricing, then security or integrations or docs, then back to pricing, then exit. The path is the diagnosis and it names the missing thing precisely.
Fix: answer it on the pricing page. The detour is telling you which question to add.
4.The reader is not the buyer
An evaluator convinced and unable to authorise. They are not stalled on your page, they are stalled in their own organisation, and everything they need is ammunition for someone else.
Behavioural tell: multiple visits from one company domain, sometimes different people. Downloads of anything shareable. Long gaps between visits that match approval cycles rather than research cycles.
Fix: give them the artefact. A shareable summary, a quote they can forward, an ROI page addressed to a finance reader.
Telling the four apart
Diagnosis view| Stall | Signature behaviour | Wrong fix |
|---|---|---|
| Unclear unit | Calculator use, then exit | Lowering the price |
| Wrong tier boundary | Oscillation between two tiers | A discount |
| Unpriced requirement | Detour to docs or security | More social proof |
| Reader is not buyer | Multiple people, one domain, long gaps | Retargeting the reader |
III.Seeing it, which requires the person rather than the page
Every signature above is a sequence across visits, so any tool that resets at the session boundary cannot show you one.
A.What you actually need to record
Three things, and none is exotic. Pricing page viewed, with the tier in focus as a property. Any comparison or calculator interaction as an event. And the identity carried across visits so the fourth one is known to be the fourth.
The third is the hard one and it is the one that matters. Without it you have a pile of unrelated pageviews. With it you have a sequence, and every diagnosis above becomes legible.
B.The population worth building
Define it as: viewed pricing two or more times, no trial or purchase, most recent visit within 30 days. That is your stalled-qualified group, and for most B2B businesses it is larger than the trial pipeline and completely unattended.
Keep it live rather than exporting it monthly. People enter and leave it continuously, and a static export is stale within a week. Our segment-to-campaign workflow covers wiring that into outreach.
C.Why this is a person-level problem by construction
Kissmetrics resolves every event to a person across devices and months, so the fourth pricing visit is recorded as the fourth, and the detour to the security page three days earlier is on the same record. Opening the stalled population gives you the individuals, with their sequences, which is what turns a rate into a diagnosis.
It also makes the counterfactual checkable. Some share of stalled visitors return and buy without any intervention, and if you cannot follow the person past the stall you will credit yourself for all of them. Our guide to SaaS pricing models covers the packaging side of the tier-boundary problem.
One honest limit: behavioural signatures are strong hints and not proof. Oscillation between two tiers is consistent with a packaging problem and also with someone comparing you against a competitor in another tab. Use the data to decide who to ask, then ask them.
Verdict
Repeat pricing page visits without action are the highest-value unattended signal most businesses have. The visitors are pre-qualified, pre-persuaded, and stuck on something specific that is usually cheap to fix once you know which of the four it is.
Record the tier in focus, record comparison interactions, and carry identity across visits. Then build the stalled population and open it. The remedies are contradictory enough that guessing is worse than doing nothing, and the sequence tells you which one you have.
Continue Reading
The 3 Models of SaaS Pricing and How to Choose the Right One
Where you price yourself determines your entire business structure. A $20/month product requires completely different operations than a $2,000/month product. Here are the three models and how to choose.
Read articleMicro-Funnels: How to Find and Fix Hidden Bottlenecks in Your Customer Journey
Your main funnel shows the big picture. Micro-funnels reveal the hidden friction points within each stage. This technique consistently uncovers the small fixes that produce outsized improvements.
Read articleTrial to Paid Conversion: Strategies That Move Users Past the Paywall
The average SaaS free trial converts at 15-25%. The best companies convert at 40%+. The difference is not the product. It is how you guide users to value before the trial ends.
Read article