Most sales pipelines lie. Not on purpose, they just fill up with deals that felt promising at the time, sorted into stages that mean different things to different people, and they quietly become a wish list dressed as a forecast. Sales process and pipeline management is the unglamorous craft of fixing that: defining what each stage really means, and keeping the numbers honest enough to act on.

The quick version

  • A sales process is the repeatable set of stages a deal moves through from first contact to closed; a pipeline is the live view of every open deal sitting in those stages right now.
  • A stage is only useful if it is defined by buyer evidence, not seller optimism, an objective exit criterion (something the customer did) that lets a deal move forward.
  • Research links a defined, formal sales process to faster revenue growth, but the gain comes from managing the activities you control, not from staring at the forecast number you can only hope for.
  • The single most useful habit is a recurring pipeline review that removes dead deals and pressure-tests live ones, so the forecast reflects reality rather than wishful thinking.

The idea in depth: a stage is a buyer milestone, not a seller mood

The common failure is to name pipeline stages after what the salesperson is feeling. "Interested." "Hot." "Verbal yes." None of these can be checked by anyone else, which means the pipeline becomes a record of mood rather than progress. The fix, drawn straight from how disciplined sales organisations build their funnels, is to define each stage by an objective exit criterion, a thing the buyer has demonstrably done, before a deal is allowed to advance. Practitioner guidance for tools like HubSpot's CRM makes the same point: every stage should have explicit entry and exit conditions that the whole team understands the same way, or the data is worthless.

In practice that means rewriting your stages as evidence. Not "Qualified" but "prospect confirmed budget, authority, need and a timeline." Not "Proposal" but "buyer agreed to review a written proposal by a named date." Each label answers the question what did the customer do that proves they moved? When you can answer that for every stage, two things happen: forecasts stop being a popularity contest, and your weekly review can challenge any deal that has not earned its position.

The case for bothering is empirical. In a 2015 Harvard Business Review piece, Jason Jordan and Robert Kelly reported that companies with a formal, structured sales process generated more revenue than those without one, their analysis associated a defined process with roughly 18% higher revenue growth. That is the headline, and it is worth taking seriously. But it is also widely over-read: a tidy process does not sell anything. What it does is make selling visible and manageable, which is the real subject of the next section.

flowchart LR
  L(["New lead"]) --> Q(["Qualified
buyer confirmed need + timeline"])
  Q --> D(["Discovery done
buyer shared real requirements"])
  D --> P(["Proposal reviewed
buyer agreed to evaluate by a date"])
  P --> N(["Negotiation
buyer engaged on terms"])
  N --> W(["Won"])
  Q -. "no evidence?" .-> X(["Disqualified / parked"])
  D -. "stalled" .-> X
  P -. "stalled" .-> X
Each stage gates on something the buyer did, not how the rep feels. The dotted exits are where pipeline hygiene happens. Leaders Loop

Manage the activities you control, not the result you can only hope for

The most useful idea in modern pipeline management comes from Cracking the Sales Management Code (Jason Jordan and Michelle Vazzana, 2012). Studying 306 sales metrics in use across sales forces, the authors found that the vast majority were not actually manageable. They sort metrics into three tiers: business results (revenue, market share) that you can only hope for; sales objectives (pipeline coverage, win rate) that a manager can influence; and sales activities (calls made, meetings booked, opportunities created) that a salesperson can directly control. You cannot manage revenue. You can only manage the activities that, stage by stage, tend to produce it.

You can't manage a number you can only hope for. You can only manage the activities upstream of it.

So you run the pipeline backwards. Start from the revenue you need, divide by your average deal size to get the number of wins, then walk back up your own stage-to-stage conversion rates to find how many qualified opportunities, and therefore how many discovery calls, and how many leads, that implies. Now you are managing inputs you can actually change this week, not a quarterly outcome you can only watch arrive. This is also what "pipeline coverage" really means: not a vanity multiple, but whether the controllable activity upstream is enough to hit the result downstream.

An honest limitation. Process discipline has a real failure mode. A pipeline is a model of the buyer's journey, and models drift the moment the market shifts, a process tuned for last year's buyer can enforce confident motion through stages that no longer match how people buy. Stage definitions also invite gaming: reps learn to park deals in a flattering stage, or sandbag the ones that make the quarter look easy. And the headline research is correlational, the companies that build a formal process may simply be better-run in ways the process merely reflects. Treat the 18% as encouragement to get disciplined, not a promise that a tidy CRM prints money. The process is a lens for seeing the deal clearly. It is not the deal.

Pipeline hygiene: the boring habit that does the heavy lifting

If there is one practice that separates a trustworthy pipeline from a hopeful one, it is regular hygiene, the deliberate removal of deals that are dead, stalled, or have no genuine next step. Industry guidance on pipeline management converges on the same warning: stalled deals with no recent activity and outdated close dates inflate the pipeline's apparent value while contributing nothing to revenue, and forecasts built on them become optimism dressed as data. A pipeline that looks busy is not the same as one that will close.

The fix is a standing rule: every open deal must have a scheduled next step with a date, or it does not belong in the active pipeline. No next step means the deal goes to a "parked" or "disqualified" state, not deleted, not hidden, just out of the forecast where it was lying to you. Run this every week and your forecast stops drifting upward by inertia. The discomfort of moving a once-loved deal to "parked" is exactly the discomfort that keeps the number honest.

A worked example

Take a small B2B software team, call the rep Maya, with a target of £240,000 in new business this quarter. (Illustrative figures throughout; this is a teaching example, not real accounts.) Her average deal is an illustrative £20,000, so she needs roughly 12 wins. Walking back up her own historical conversion rates, say an illustrative 25% of qualified opportunities close, and 40% of discovery calls reach qualified, she needs about 48 qualified opportunities and around 120 discovery calls in the quarter. Suddenly the target is not a scary number on a slide; it is a weekly rate of calls she can actually run.

Now the hygiene bites. Maya's CRM shows £600,000 of "open" pipeline, comfortably more than her £240,000 target, so on paper she is fine. But the Friday review asks one question of each deal: what did the buyer last do, and what is the dated next step? Six deals worth an illustrative £180,000 have had no buyer activity in three weeks and no next meeting booked. They are not pipeline; they are hope. Parked. Her real, evidence-backed pipeline is £420,000, still healthy, but honest, and now she knows she must create new qualified opportunities rather than coast on a number that was never real.

flowchart TD
  T(["Target: £240k this quarter"]) --> A{"÷ £20k avg deal"}
  A --> B(["≈ 12 wins needed"])
  B --> C{"÷ 25% close rate"}
  C --> D(["≈ 48 qualified opps"])
  D --> E{"÷ 40% discovery→qualified"}
  E --> F(["≈ 120 discovery calls"])
  F --> G(["A weekly activity rate Maya controls"])
Running the pipeline backwards turns an uncontrollable target into a controllable weekly activity rate. Figures illustrative. Leaders Loop

The process did not close anything for Maya. It told her the truth twice: that her target was a manageable rate of calls, and that a third of her pipeline was fiction. Both are more useful than a comforting £600,000.

Frequently asked questions

How many pipeline stages should we have?

Enough to mark genuinely distinct buyer milestones, and no more. Most B2B teams land on four to seven. The test is not a magic number; it is whether each stage has a clear, buyer-evidenced exit criterion that everyone applies the same way. If two stages share the same exit test, merge them. If a stage is just "we feel good about this," cut it.

What is the difference between a sales process and a sales methodology?

The process is the stages a deal moves through in your CRM, your map of the journey. A methodology (MEDDIC, SPIN, Challenger, solution selling) is how a rep sells within those stages, the questions they ask and the qualification they apply. You need both: the process tells you where a deal is; the methodology helps move it forward.

What pipeline metrics actually matter?

Start with stage-to-stage conversion rates (where deals leak), sales cycle length (how long each stage takes), average deal size, and win rate. Together these let you run the pipeline backwards from a target. Avoid managing total pipeline value alone, it is the easiest number to inflate and the hardest to act on.

Why do forecasts come in wrong even with a good process?

Usually because of pipeline hygiene, not process design. Stalled deals with optimistic close dates inflate the number; reps sandbag or pad to manage expectations; and any model lags a shifting market. A weekly review that enforces a dated next step on every deal fixes most of it. The rest is the honest uncertainty of predicting other people's decisions.

Do small teams really need this, or is it just for big sales orgs?

Small teams benefit most, because they have the least room to waste effort on dead deals. You do not need expensive software, a shared spreadsheet with defined stages, exit criteria, and a dated next step per deal delivers most of the value. The discipline is the point, not the tooling.

Related in the Toolkit

A pipeline only works if the deals entering it are the right deals, which is why territory, segment & quota design sits upstream of everything here, and why funnel & conversion optimisation is the natural next step once your stages are honest enough to measure.

Where to go next