Funnel conversion benchmarks. What good actually looks like, stage by stage.

The headline numbers. What a healthy B2B funnel actually converts at.

I'm Lauren Pearson, and clients send me the same screenshot at least once a month: a single "average conversion rate" pulled from a blog post, next to their own number, with a question mark. It is rarely a useful comparison, because a blended average hides five very different stages that each convert at a different rate for a different reason. The table below breaks the B2B funnel into the stages that actually matter for a pipeline review, with the current published ranges for each.

For a single "what counts as good" figure at the top of the funnel, our guide to what a good conversion rate actually is covers the context that number needs. This piece goes a stage deeper, into the parts of the funnel that decide whether that top-of-funnel number ever turns into revenue.

How the figures break down. Vertical and deal complexity move the number more than anything else.

Ruler Analytics built its 2026 figures from over 110 million website sessions and five million tracked conversions across 13 industries, and the spread inside that dataset is the real lesson. B2B SaaS converted visitors to leads at just 1.1 percent, while legal services converted at 7.4 percent, a sevenfold gap inside the same "B2B" label. Neither number is wrong. A SaaS buyer researching quietly for weeks before filling in a form behaves nothing like someone searching for a solicitor after a specific event, and the funnel has to be judged against a category that actually resembles it.

First Page Sage's mid-funnel figures show the same pattern in reverse: their SaaS-specific data puts lead to MQL at roughly 39 percent, ahead of the general 25 to 35 percent range, because SaaS buyers who fill in a form have usually already done enough self-directed research to qualify quickly. The opposite is often true further down the funnel in longer, more considered categories, where SQL to opportunity conversion can run higher than the general range because fewer, better-qualified deals make it that far in the first place.

Deal complexity is the other variable worth holding in mind, separately from vertical. Two SaaS businesses with identical products can have genuinely different funnels if one sells a single seat to an individual and the other sells a company-wide rollout with procurement sign-off. The single-seat business should expect a shorter, higher-converting path from SQL to opportunity, since there is one buyer to convince rather than a committee, while the rollout business will show a lower rate at that stage even when the underlying sales process is working exactly as it should. Reading the two against the same benchmark punishes the more complex, and usually more valuable, deal for being harder to close quickly.

The number benchmarks leave out. How long a deal sits at each stage.

A conversion rate on its own answers whether a deal moves. It says nothing about how long it takes to get there, and that second figure often matters more to a founder's cash position than the percentage does. Two funnels can both convert MQL to SQL at 20 percent and be in entirely different health: one where that conversion happens within a week of the lead landing, and one where qualified leads sit untouched for a month before anyone follows up. The rate looks identical in a monthly report. The revenue impact does not.

Time-in-stage is worth tracking alongside every rate in the table above, stage by stage, using the CRM's own timestamp on when a record entered and left each pipeline stage. A stage that is both converting well and moving quickly needs no attention. A stage converting well but moving slowly usually has a process gap, no defined owner, no service-level target, rather than a genuine quality problem, and it is a cheaper fix than anything that touches lead volume or scoring.

Where the bottleneck actually sits. A practitioner's read, not a formula.

In five years of running these reviews for founder-led B2B teams, the stage that quietly costs the most revenue is almost never the one clients ask me about first. They usually want to talk about the visitor to lead rate, because it is the number marketing reports on, but the stage that is actually broken is MQL to SQL more often than any other. It sits in the gap between two teams, marketing hands off a "qualified" lead and sales decides whether to believe it, and without a shared, written definition of qualification, that handoff drifts.

A worked example makes it concrete. A 35-person B2B software business I advised was converting MQLs to SQLs at 11 percent, well below the 13 to 26 percent range, while every other stage sat comfortably inside benchmark. The lead scoring model was awarding points for behaviour, a pricing page visit, a webinar signup, that correlated with interest but not with budget or authority to buy. Adding two firmographic questions to the lead form, company size and current tooling, and re-weighting the score around them lifted MQL to SQL to 19 percent within a quarter, without a single change to the top of the funnel. The lesson holds more broadly: a bottleneck in the middle of the funnel is usually a definition problem, not a volume problem, and it rarely needs more traffic to fix it.

How to read your own data. Benchmark against the right comparison, not the average.

Pull the last two full quarters of stage-by-stage counts from the CRM, not a single month, since most founder-led teams do not have enough monthly volume for one month's rate to mean much on its own. Segment by lead source and, where the sample allows it, by deal size, because blending an inbound enterprise enquiry with a small self-serve signup produces a rate that describes neither well.

Once the numbers are segmented, compare each stage against its own range above rather than a single blended target, and flag only the stage that sits meaningfully outside it, not every stage that is merely below the midpoint. A funnel with four stages inside range and one stage well outside it has a specific, fixable problem. A funnel with every stage a little below range usually has a lead quality problem further upstream than the funnel itself, which is worth diagnosing before touching the funnel mechanics at all. For the ecommerce-specific version of this exercise, checkout and cart figures behave differently again, covered separately in our ecommerce conversion rate benchmarks guide, since a B2B pipeline funnel and an online store funnel are not the same shape and should not be read against the same numbers.

A short checklist before acting on any of this: confirm the CRM's stage definitions have not drifted since the last review, since a redefined "MQL" halfway through a quarter makes the before-and-after comparison meaningless. Exclude test records and internal deals from the count, a smaller error than it sounds but a common one in founder-led CRMs where the founder's own accounts sit in the same pipeline as real prospects. Then look at the one stage furthest from range, agree a single change with a named owner, and hold the rest of the funnel steady for a full quarter so the effect of that one change is actually visible in the next set of numbers, rather than lost in a dozen simultaneous tweaks.