Your Funnel Does Not Leak in the Middle. It Leaks at the Seams.

Ask a CRO where revenue is lost and they will point at conversion rates. But conversion rates measure stages, and stages are usually fine. The losses concentrate at the seams, the moments where one function hands a buyer to another and accountability briefly belongs to no one.

Why seams fail when stages do not

Inside a stage, one team owns the buyer and its own number. At a seam, ownership transfers, and transfer is where systems fail: context gets dropped, urgency resets to zero, and the buyer experiences a gap precisely when their attention is highest. The structural cause is that companies design organisations by function and measure by function, so nobody's dashboard shows the seam at all. Marketing reports MQLs, sales reports pipeline, success reports retention, and the space between them reports to no one. This is the systems problem we describe across the whole revenue architecture: each function can hit its number while the company misses its own, because the losses live in the handoffs.

The marketing-to-sales seam

MQLs rot because sales does not trust them, and sales is often right. The fix is not a better lead score; it is a shared pipeline definition. One meeting, both leaders, one question: what evidence makes an account worth a seller's time? Write it down. Route on it. Review misses weekly.

Making the shared definition stick

The definition decays unless it has an operating rhythm around it. Three mechanisms keep it alive. First, a weekly thirty-minute review of every account sales rejected, with marketing present, until the rejection rate stabilises below an agreed line. Second, a service-level agreement in both directions: marketing commits to evidence quality, sales commits to touching every accepted account within a defined window, and both numbers are on the same dashboard. Third, a quarterly re-derivation of the definition from closed-won data rather than from opinion, because the market moves and last year's qualification evidence goes stale. Companies with a clean CRM find this almost mechanical. Companies without one discover that the seam argument was really a data argument all along.

The sales-to-success seam

The deal closes, the AE moves on, and everything they learned (the champion's real motivation, the promised outcomes, the political map) dies in a handoff call that never happened. Expansion revenue is won or lost in the first 30 days post-sale. Treat the handoff as a deliverable with a template, not a courtesy.

The handoff document that actually gets used

A usable handoff template fits on one page and answers five questions. Why did they buy now, in the champion's own words? What outcome did we promise, with the number attached? Who signed, who will use it daily, and who was against the deal? What did we learn about their politics and their timeline? What was left unsold that maps to a future expansion? The AE completes it before commission is finalised, and the success lead walks the customer through the promised-outcome line in the kickoff call. That single act, repeating the promise back to the customer in week one, sets up the expansion motion better than any QBR deck, because renewal conversations become a review of a commitment both sides remember making.

Instrument the seams

Add two metrics to your weekly review: acceptance rate (what share of marketing-sourced accounts does sales actively work?) and time-to-first-value (how long from signature to the customer's first realised outcome?). Both are boring. Both predict revenue better than any stage conversion rate you currently track.

The third seam nobody names

There is a quieter seam running backwards: customer success to sales and marketing. Expansion opportunities that success spots but nobody sells, renewal intelligence that never reaches the forecast, happy customers nobody asks for references, case-study material that dies in a QBR deck. Because the value flows against the usual direction, no process carries it. The fix mirrors the forward seams: a defined route for expansion signals into a real expansion pipeline, a monthly slot where success briefs marketing on the phrases actual customers use, and a standing reference request tied to every success milestone. Companies that instrument this reverse seam effectively get a second demand engine from accounts they already won, at a cost of two meetings a month. It is the cheapest pipeline in the business, and most teams simply never plumb it.

What good looks like after ninety days

Progress on seams is measurable quickly if you baseline before you start. By the end of the first month you should have the shared pipeline definition written and routed, the one-page handoff template live on every closed deal, and both seam metrics reporting weekly. By day sixty the rejection-rate review should be shrinking its own agenda, because marketing is correcting sourcing against real feedback. By day ninety, acceptance rate should be trending toward the agreed line, time-to-first-value should have a number leadership recognises, and the forecast conversation should reference seam metrics unprompted. If none of that has moved, the usual cause is not the mechanics but ownership: a seam still reports to no one. Assign each seam a single named owner with the authority to convene both sides, and the metrics start moving within two cycles.

Where to start if everything leaks

You cannot fix both seams at once with the same energy, so sequence by arithmetic. If win rates are healthy but pipeline feels thin, start at marketing-to-sales, because you are paying to generate demand you then discard. If new logos land but net revenue retention sits below 100 per cent, start at sales-to-success, because you are refilling a leaking bucket. If you genuinely cannot tell, that is itself the finding: your data will not answer a basic question about your own funnel, and a structured look at the whole system, like a diagnostic before any tactic purchase, will pay for itself in the first quarter. Seams are unfashionable work. They are also the highest-yield fix in most GTM systems we see.