How to design sales stages for your product

Stages do three different jobs, and the designers who conflate them end up with stages that feel busy but forecast badly. Fit answers whether a prospect could ever buy. Triggers answer why they are evaluating now. Criteria answer how they will choose. Walk through every recent won deal and label each piece of evidence with one of those three tags. You will find that fit evidence shows up in the first conversation, triggers show up early in discovery, and criteria harden through the middle stages. Once the tags separate cleanly, stages fall out of the data instead of being invented in a conference room. The point of this step is simple. If you cannot tell whether a signal is fit, trigger, or criterion, you cannot build a stage around it, and reps will not know which gate they are trying to pass. Design the taxonomy first and name the stages last.

• Pull the last five closed-won and five closed-lost deals and transcribe the moments where buyer behavior shifted
• Tag each shift as fit (could they buy), trigger (why now), or criterion (how will they decide)
• Confirm that fit evidence appears earliest, triggers next, and criteria through the middle, before promoting any to a stage
• Share the tagged examples with two frontline reps and resolve any disagreements before moving on

The first real stage is a fit gate, and it should be ruthless. A rep should be able to run the fit test inside a single discovery call and know whether to proceed, qualify out, or route elsewhere. Build the test from your ICP and your explicit disqualifiers: industry, company size, geography, tech stack, regulatory posture, and any dealbreakers from won-deal patterns. Keep it to five or six yes-or-no questions. Any more and reps start skipping items. Any fewer and the gate leaks. Store the answers as required picklists on the opportunity, not notes in the body of a call log. A fit gate that lives in free text is a fit gate that does not exist. Reps often resist disqualification because every lead feels like it could be the one. The data says otherwise. Across almost every B2B motion, deals that fail the fit test close at under five percent and consume twice the average cycle length. Faster disqualification is faster quota attainment.

• List every disqualifier from the last twenty closed-lost deals and group them into five or six buckets
• Convert each bucket into a yes-or-no picklist on the opportunity record
• Require all picklists to be answered before the opportunity can leave the fit stage
• Add a validation rule that routes any disqualified opportunity to a Not A Fit sub-pipeline for reporting

Deals exist for reasons. A new executive, a failed rollout of a competing tool, a funding round, a merger, a regulatory deadline, a public outage. If a rep cannot name the trigger on a specific date tied to a specific person, the deal is almost certainly a research exercise disguised as a pipeline opportunity. Make trigger a required field on every opportunity entering the discovery stage, structured as a picklist of trigger categories plus a free-text date and short description. Picklists let you run cohort analysis later: which triggers convert fastest, which produce largest deal sizes, which stall. Date and description keep reps honest. If the trigger entry says the champion joined last month and the LinkedIn record says three years, the manager has a conversation to start. Buying triggers are the second-most predictive field in a healthy pipeline, right behind economic buyer named. Treat them with that weight.

Decision criteria sit at the heart of the middle stages, and they are where most designers go wrong. Reps default to criteria that sound like a feature list: integrations, security certifications, reporting depth. Buyers actually choose on a smaller, harder set: does this solve the real problem, can we trust the vendor, will our team adopt it, and does the math work. Build four or five criterion fields on the opportunity that force the rep to document the buyer version, not the seller version, of each test. For each criterion, record who evaluates it, what evidence they want, and when they expect to decide. A rep who cannot name those three things for every active deal is guessing, and guesses do not forecast. Reviewing the criterion fields in weekly pipeline review is how you build the muscle. If a rep says the buyer cares about reporting, the follow-up is simple. Which specific report, who asked for it, and when do they want to see it live. Vague answers signal that the criterion is a seller assumption, not a buyer commitment.

• Interview three recent won buyers and ask what actually tipped the decision; look for patterns across the sample
• Convert the top four or five patterns into criterion fields with a who-evaluates picklist and an evidence-expected text field
• Require criteria to be filled before a deal advances past mid-funnel
• Audit filled criteria quarterly against actual won-reason data and retire any that do not predict outcomes

Each stage needs exit evidence that proves the deal has moved. Observable means something a camera could record. Economic buyer named means a specific person on a specific date who said in a meeting that they own the budget. Technical validation complete means a signed-off test plan or a decision document, not a rep writing down that the champion liked the demo. Three to four pieces of observable evidence per stage is the right density. Fewer than three and the gate is porous. More than four and reps either lie or stop updating the record. Store exit evidence as a short checklist on the opportunity with validation rules that block stage advancement until all boxes are checked. Reps will push back the first week. By week three, they will have adjusted their call plans to produce the evidence proactively, and your forecast accuracy will start moving in the right direction. The checklist is the single most durable artifact in the whole design. Everything else is window dressing on top of it.

• For each stage, list three to four pieces of evidence that must be observable to an outsider reviewing the record
• Convert each into a required field with a defined type: name, date, document link, or picklist
• Add a validation rule that blocks stage advancement when any box is empty
• Review the checklists every quarter and replace anything reps have found a way to game

Every stage carries a win probability and a velocity target. Both should come from your own data. For probability, calculate the historical close rate of deals that reached each stage, pulled from at least four quarters of closed deals. If you have less history, use benchmarks only as a starting point and recalibrate after your first full quarter. For velocity, measure the average time deals spent in each stage, excluding the deals that stalled past the eightieth percentile. The stall exclusion matters because a few sick deals skew the average and make your velocity target too generous. Publish both numbers next to each stage definition so reps and managers see them whenever they look at the pipeline. Benchmarks published by third parties are useful for sanity checks, never for primary targets. Your deal size, segment, product maturity, and sales motion all distort the industry average. Trust the data your own motion produced and update the numbers every quarter.

Pipelines fail because deals that should die stay alive. Build an explicit disqualification stage that reps advance deals into on purpose, with a loss-reason picklist that captures the real cause: no trigger, failed fit, no economic buyer, lost to competitor, lost to no decision, deprioritized, budget pulled. Make loss reason required at the moment of disqualification, not later. Later always means never. Set a monthly review of disqualification reasons at the segment level and feed the patterns back into the fit test and criteria design. If no decision is winning twenty percent of your losses, your criteria are not forcing a timeline. If competitor loss is clustered in one segment, your value proposition in that segment needs work. The disqualification stage is where your most valuable product insight lives. Treat it with the same discipline as closed-won analysis and the whole motion gets sharper every quarter.

• Build a loss-reason picklist with seven to ten categories drawn from your own historical losses
• Make loss reason required at the moment of disqualification and non-editable afterward
• Review loss patterns monthly at the segment level and feed changes back into fit and criteria design
• Run a quarterly deep dive on no-decision losses specifically; those are usually criteria failures, not product failures

The stages do not become real until they are measured. Build a dashboard that shows four numbers per stage: entry volume, exit volume, average time in stage, and conversion to next stage. Watch for the stage where deals pile up longest and the stage with the lowest conversion rate; those are almost always where your design is weakest. Review the dashboard weekly in forecast call and run a formal recalibration every quarter against ninety days of closed data. Ask three questions each cycle: did deals actually move through stages in the sequence you designed, did exit evidence predict wins, and did any stage become a graveyard. If a stage conversion rate falls below ten percent, either the gate is too loose or the stage should not exist. If it climbs above eighty percent, the stage may be redundant. Treat stage design the way a product team treats a shipping product: measured, iterated, and never finished. Strkr AI can surface the slowest stage and the lowest-conversion stage automatically each week, and Strkr Messaging can nudge reps whose exit evidence is missing before the review, so the hygiene work happens before the meeting instead of during it.

• Build a stage dashboard with entry volume, exit volume, average time in stage, and conversion to next stage
• Review the dashboard weekly in forecast call and tag any outlier stages for next-quarter work
• Pull ninety days of closed data every quarter and recalibrate probability, velocity, and criteria
• Communicate changes to the sales team in writing one week before they take effect, and never mid-quarter