Why is your B2B pipeline flat when lead volume is up? A four-check diagnosis

A lead is a form fill or a scored contact. Pipeline is a qualified opportunity with a buyer who has a problem and a budget. The two drift apart when volume is bought with offers that attract researchers, students or the wrong company size, because those leads count in the report and never turn into a meeting.

Timing adds to it. In 6sense’s 2026 study (vendor research), vendor contact happened about 59% of the way through a 10.3 month buying cycle, and the vendor the buyer favored before contact won 82% of the time. Many of your leads arrive after the shortlist exists, so a rising lead count says little about whether you are on it.

The diagnosis below takes about a day if your CRM has lead source, lifecycle stage dates and company fields filled in.

Run them in this order because the first one tells you where to look and the later ones explain why.

Split the rate by source, offer and campaign, and compare cohorts by lead creation month so late-stage conversions do not distort the read. A blended number is the most common way this problem stays hidden.

The illustration below uses assumed numbers. Leads grew 50%, the rate halved, and opportunities fell.

The report that showed leads up 50% hid a cost per opportunity up two thirds. This is why we report cost per opportunity next to leads, as described in our guide to paid media reports that show pipeline.

If the opportunity count holds but revenue does not follow, look at where opportunities sit. Export opportunities by stage with days in stage. A stack of aging deals in the first stage means the sales team accepted them without agreeing on need or budget, which is a handoff problem rather than a lead problem.

Agree the definition of an accepted opportunity in writing, with the fields that must be filled. Our marketing and sales SLA template sets out the acceptance rules and response times.

Count leads and opportunities by company size, industry and role against your ideal customer profile. Flag the share from your target account list. When fit falls, the cause is nearly always targeting that widened, such as broad match keywords, open LinkedIn audiences, or an offer that appeals to individuals.

Where several people from one company engage, the buying group is forming even when no single lead scores high. That is the argument behind buying groups vs MQLs, and it is why account-level reporting finds problems lead-level reporting misses.

List every source that grew. Gated reports, webinars and broad paid social generate volume cheaply, and a large share of that volume is research, not buying. This is not a reason to stop them. It is a reason to count them as priming, and to judge them by account engagement and later pipeline, not by lead count.

Dreamdata’s 2026 benchmarks (vendor data) put 81% of the B2B journey outside the sales pipeline, with an average of 88 touchpoints and 10 stakeholders, which is why one lead rarely tells you how a deal is going.

Fix one thing, then re-read the same cohorts four to six weeks later. The marginal dollar test in our $50,000 guide uses the same method for deciding where extra budget goes.

We run this diagnosis as the first step of a paid media or demand generation engagement: CRM fields and stage dates first, then reporting by source and cohort, then fixes in order of expected pipeline impact.

Our RevOps team builds the stage and source reporting. In the KNIME work this discipline produced 51% lower cost per lead with volume steady and 100% of paid leads traceable to closed deals.