If you are asking why your sales cycle is so long and nobody in your company can give you a straight answer, the problem usually is not your reps. It is that your pipeline stages describe what your team did, not what the buyer actually decided. “Sent proposal” is an activity. “Champion confirmed and next step scheduled with the economic buyer” is a buyer decision. Those are not the same thing, and only one of them tells you anything real about whether a deal is going to close.
The tell
Ask your team to walk you through the last deal they won, stage by stage. Listen closely to the verbs. If every stage advancement sounds like “I sent the proposal” or “I had the call,” your stages are activity-based, and that is exactly why your pipeline cannot tell you anything trustworthy about what is actually going to close or when.
Here is the deeper problem with activity-based stages: they let deals sit. A rep can genuinely believe a deal is progressing because they did something, a call, an email, a proposal, when the buyer has not actually moved at all. The deal looks alive on your dashboard. It is dead. And because nobody is forcing a real buyer action before advancing the stage, dead deals clog the pipeline for weeks or months before anyone notices, which drags out your average cycle length and wrecks your forecast at the same time.
What this actually costs you
Poor pipeline data is not a rounding error. Gartner’s own research puts the average cost of poor data quality at $12.9 million a year per organization, and pipeline stage data, the thing your forecast is actually built on, is one of the most common places that cost originates (source).
Here is where most of that cost actually hides. Research by Matthew Dixon and Ted McKenna, based on an analysis of 2.5 million sales calls, found that 40 to 60% of B2B deals that fail to close are lost to customer indecision, not to a competitor. The buyer never said no. They just never said yes, and the deal quietly stalled while nobody was forcing the question (source). Activity-based pipeline stages make this invisible. A rep genuinely believes a deal is progressing because they did something, a call, an email, a proposal, while the buyer has already gone quiet. The deal sits in your forecast as “still working” for weeks or months, consuming pipeline coverage and sales attention, until someone finally closes it out as lost, usually logged as lost to a competitor when the real cause was indecision the whole way through.
This shows up differently by vertical, but the underlying mechanism is identical. Managed services contracts commonly run 30 to 180 day sales cycles, plenty of runway for a deal to sit “in proposal” for months while a buyer stalls on internal budget approval. An agency retainer pitch can sit in “sent proposal” for weeks after the prospect’s internal champion has gone dark. A staffing submittal can look like real movement while the hiring manager never actually confirmed the role is still open. In every case, the fix is the same: define what the buyer has to do or confirm before the stage counts as real progress, not what your team did.
That is the real cost of activity-based stages. It is not just a longer number on a dashboard. It is capital tied up in deals that were never going to close, reps spending time on the wrong things, and a forecast nobody can trust.
What good stage design looks like
Level 1: Your CRM still has the default stages it shipped with: Lead, Qualified, Proposal, Closed. Nobody can articulate what actually moves a deal from one to the next. Stage changes when the rep feels ready to change it.
Level 2: You have renamed the stages to match your business, but there is still no documented exit criteria. Advancement is subjective. Different reps interpret the same stage differently, which means your pipeline data cannot be used for reliable forecasting even though it looks organized.
Level 3 (Functional): Every stage has a written exit criterion phrased as a buyer action or a verified fact, not a sales activity. Your team actually knows the criteria and applies them.
Level 4: Exit criteria become required CRM fields, so a deal literally cannot advance until the criterion is met. Stage conversion rates get tracked and benchmarked. Deals that try to skip a criterion get caught and corrected before they distort the picture.
Level 5 (top):Stage design keeps improving based on real conversion and win/loss data. Exit criteria are validated by what buyers actually did, not asserted by what a rep believes, and increasingly that validation is not a manager double-checking a rep’s note, it is conversation-intelligence AI listening to the actual sales call or reading the actual email thread and confirming whether the buyer really said what the rep logged. Tools built for exactly this, Gong and Clari among them, flag the gap automatically when a deal’s stage does not match what happened on the call (source). Stage duration benchmarks flag stalled deals automatically, and the pipeline becomes something you can actually forecast from, not just a to-do list.
You do not need a dedicated tool to test whether this would help you. Pull the call recording or email thread from your longest-stalled deal right now, paste the transcript into Claude, ChatGPT, or Grok, and ask it directly: based on this conversation, did the buyer actually confirm what our exit criterion for this stage requires, or did our rep just believe they did? Most founders who try this on even one deal are surprised by the answer. That is the manual version of what conversation-intelligence software does continuously, and it costs nothing beyond a tool you likely already have. What “buyer confirmed” looks like will differ by business too. In field services, it might be a signed estimate inside ServiceTitan or Jobber rather than a verbal commitment on a recorded call, but the underlying discipline, verify against what actually happened, not what a rep believes happened, is identical.
Where to start
This competency depends on Revenue Lifecycle Design, the full map of how a customer moves from first contact through onboarding and expansion. If you have not mapped that yet, pipeline stage design will always feel like you are organizing symptoms instead of fixing the cause. Map the lifecycle first, then define your stages as the specific decision points a buyer crosses inside that map.
If you already have that map, the fastest next step is picking your worst-performing stage transition, the one where deals seem to sit the longest, and writing down, in plain language, exactly what has to be true for a deal to leave it. Not what your team has to do. What the buyer has to do or confirm.
