B2B sales funnel: how to design stages that predict revenue

A B2B sales funnel predicts revenue when each stage records a decision the buyer made, with entry and exit criteria anyone on the team can check in the CRM, a probability calculated from the company’s own historical conversion and time measured in each phase. For new business, that fits in five to seven stages. The funnel that comes from the CRM’s default setup, with stages like “proposal sent” and probabilities rising ten points at a time, measures the salesperson’s activity and hands leadership a number nobody can explain. The redesign usually fails after it is finished, when ten people have to apply the same criterion to the same deal at the end of a tight month, and that behavior is built through practice and a management routine.

When I open a company’s pipeline for the first time, I ask for a simple report: how many deals entered each stage last year and how many of them ended up won. In many operations, the real conversion of “proposal sent” sits well below the 60% or 70% the CRM assigns to that stage, and the negotiation stage holds deals with no contact for months. The sales manager knows this and applies a mental discount to the forecast, a symptom of stages that describe the salesperson’s work.

What is a B2B sales funnel and what should each stage record?

A sales funnel is the sequence of stages an opportunity goes through between the first qualified interest and the purchase decision, and the name comes from the shape: many deals enter at the top and some of them come out at the bottom. A sales pipeline is the same structure seen as a portfolio, the list of open deals with value, stage and date. Both readings depend on the same stage design.

The stage has to record a buyer milestone because the B2B buyer does not move through a purchase in a straight line. In a Gartner press release from 2018, the firm described six jobs buyers complete to finish a purchase: problem identification, solution exploration, requirements building, supplier selection, validation and consensus creation. According to the same research, most buyers revisit nearly every one of these jobs at least once before buying, and validation and consensus run throughout the entire process. For whoever designs the funnel, this has two consequences. The stage marks a commitment the buyer made, such as validating the scope or taking the proposal to the decision maker, instead of trying to reproduce the order in which the buyer thinks. And whatever happens all the time, such as mapping and winning over the buying committee, becomes a deal field reviewed at every stage.

The test of a good stage

I approve a stage when it passes two questions. Did the buyer do something I can point to in the CRM, such as a logged meeting with the decision maker or a reply to a draft scope? Would two managers, reading the same deal without talking to each other, place it in the same stage? “Proposal sent” fails the first, because the salesperson sends the proposal alone. “In negotiation” fails the second, because each person understands negotiation differently. “The decision maker received the proposal in a meeting with us present” passes both. In RevOps and CRM in B2B funnel management I show a nine-stage bureaucratic funnel cut down to four; here the focus is what each stage requires so that conversion can support a forecast.

How many stages should a B2B sales funnel have?

For new B2B business, I work with five to seven open stages, plus won and lost. With fewer than five, a single phase lumps together different buyer decisions, such as confirming the problem and approving the budget, and that stage’s conversion no longer shows where the deal stalls. Above seven, the extra stages tend to be salesperson tasks promoted to phases, such as “contact attempt” or “proposal in progress”, and the funnel ends up measuring clicks. The exact number depends on the cycle: a three-week sale with a single decision maker fits in five stages, and a nine-month sale with IT, legal and procurement involved needs six or seven.

Create another pipeline when the buyer’s decision changes, as in renewal and expansion, where the customer decides about value already delivered. A pipeline per salesperson, per region or per source channel multiplies reports without adding information, and those cuts work better as filters.

B2B sales funnel model: stage, buyer, evidence and fields

The model below is a starting point for consultative sales with a cycle of a few months and a buying committee. Each row describes the stage’s entry criterion, which is also the exit criterion of the previous stage. Adapt the names to your market’s language and keep the buyer column as it is.

Stage	What the buyer did for the deal to enter	Evidence in the CRM	Required fields to enter
1. Problem recognized	Accepted a discovery conversation about a problem they described themselves	Meeting held, logged with the contact associated to the deal	Source, main pain from a closed list, primary contact
2. Diagnosis confirmed	Confirmed the pain, the impact and the timing, and explained how the company decides purchases of this size	Discovery summary sent to the customer and answered by them	Impact stated by the customer, timing event, committee roles already known
3. Scope validated	Validated requirements and decision criteria, with the technical team involved when the purchase requires it	Customer reply to the draft scope or minutes of the technical meeting	Decision criteria, alternatives the customer is considering, estimated value
4. Proposal with the decision maker	The economic buyer received the proposal in a meeting or confirmed in writing that they are evaluating it	Meeting with the decision maker associated to the deal	Decision maker associated, proposal value, decision date agreed with the customer
5. Formal approval	Started the internal contracting process, such as supplier registration, legal review or purchase requisition	Registration request, contract draft under review or requisition received	Remaining approvals, owner on the customer side, signing date
Won	Signed the contract or issued the purchase order	Signed document attached	Handoff fields for post-sale
Lost	Communicated another decision or stopped responding after the agreed cadence	Record of the last contact and of the decision	Loss reason for the stage the deal was in, competitor when there is one

Entry criterion, exit criterion and the deal that goes back

With the criterion written at the stage’s entry, the salesperson answers “can I move it yet?” on their own. Since buyers revisit decisions, the funnel has to allow moving back: the customer who validated the scope and then replaced the IT director returns to stage 3 until the new owner validates it. Log the regression with a reason, visible to the manager, so that no deal stays in stage 4 after losing the decision maker. Also write down when a deal may skip a stage, such as a referral purchase that enters straight into stage 3, always with the required fields of the earlier stages filled in.

How do you calculate stage probability from history?

The stage probability answers one question: of the deals that reached this point and already have an outcome, how many ended up won? To calculate it, take the deals that entered the stage within a twelve-month window and have already been closed, won or lost, and divide the wins by the total closed. Leave open deals out of the calculation, because counting them as lost drags the calculated probability down. The report has to start from the date each deal entered each stage, which the CRM keeps in the deal history, and not from the stage the deal was in when it closed.

In the illustrative example below, all 40 won deals went through every stage, and the last column shows the probability the team had configured in the CRM.

Stage	Entered and closed in 12 months	Won	Historical probability	Configured probability
1. Problem recognized	400	40	10%	10%
2. Diagnosis confirmed	200	40	20%	25%
3. Scope validated	120	40	33%	50%
4. Proposal with the decision maker	90	40	44%	70%
5. Formal approval	55	40	73%	90%

The gap at stage 4 weighs the most. With R$ 2 million in proposals, the configured probability projects R$ 1.4 million, and the historical one projects R$ 880 thousand. The method that uses these numbers, with commitment categories and a weekly ritual, is in sales forecasting. The number is only valid if the team applied the stage criterion the same way during the whole window.

Care with sample size and segment

With only a few dozen closed deals per stage, the rate swings between quarters even when nothing changed in the operation; extend the window to 18 or 24 months or group similar segments. When small and large accounts convert differently, calculate by deal size band, because the average of the two describes neither. Recalculate every quarter and record which version was in force in each period.

Time per stage and loss reason per stage

How long a deal stays in each stage

Measure the median number of days in each stage, separating won and lost deals. The median for won deals shows the normal buying pace in your market and, added stage by stage, approaches the sales cycle. The median for lost deals shows where deals die slowly: if lost deals stay in stage 4 three times longer than won deals before someone closes them, the team is holding on to proposals the customer has already abandoned. The “stuck beyond the cycle” threshold, one of the exception lists in the pipeline review, comes from this number.

A closed list of loss reasons for each stage

A single list of reasons mixes losses of different kinds: the deal lost at stage 1 for being outside the profile speaks to the lead criterion, and the one lost at stage 5 to a legal veto speaks to the contract. With a closed list per stage, each team knows which loss belongs to it. “No response” can be used at any stage, as long as the salesperson completed the agreed cadence, and if “other” goes above one in ten lost deals, the list needs a review.

Stage where the deal stopped	Loss reasons in a closed list	Who uses the reading
1. Problem recognized	Outside the ideal customer profile; problem not a priority for the company; contact unrelated to the area	Marketing, to review channel and lead criterion
2. Diagnosis confirmed	Impact too small for the investment; no budget this year; decision postponed by the customer	Sales and marketing, to review qualification
3. Scope validated	Requirement the solution does not meet; customer decided to solve it in-house; integration not feasible	Product and presales
4. Proposal with the decision maker	Chose a competitor; decision maker did not approve the investment; perceived value below price	Sales leadership and executives, for proposal and pricing
5. Formal approval	Legal or procurement veto; budget frozen; leadership change at the customer	Executives and legal, for contract and terms

How do you connect the marketing, sales and post-sale funnels?

In the same 2018 press release, Gartner noted that B2B buyers use sales reps and the supplier’s website equally, for discovery and validation, throughout their buying journey. For the operation, the link between funnels depends on information following the customer from one team to the next: the pain the lead described to the SDR reaches the pain field in stage 1, and the impact confirmed in stage 2 becomes the success criterion that post-sale tracks.

The design holds together when one funnel’s exit criterion is the next funnel’s entry criterion. The last step of the marketing funnel, the lead accepted by sales, corresponds to the entry into stage 1, with the deadlines and returns defined in the SLA between marketing, sales and post-sale. The won criterion, with the handoff fields filled in, is the entry into the post-sale funnel, which follows the same logic of customer milestones: kickoff completed with the plan approved by the customer, first success criterion reached, usage within what was agreed and renewal decided. That funnel, with its account rituals, is covered in B2B post-sale.

Marketing starts tracking conversion by source all the way to won, along with the losses at stages 1 and 2, which show whether the lead criterion is calibrated. Sales leadership reads conversion, time and regressions by salesperson and takes the pattern into coaching. Post-sale receives customers with pain, impact and committee on record and sends cancellation reasons back to marketing.

How do you migrate from the old funnel without losing history?

Changing stages worries leadership because the conversion reports of past years depend on the old stages. Depending on the tool, deleting a stage forces you to move the deals in it and scrambles those reports, so I prefer to create the new funnel next to the old one, close the old one to new deals and keep it for reference. Before that, export the stage history with entry dates and build a mapping: some stages merge, and the ones that were salesperson tasks become cadence activities.

Migrating open deals through the mapping automatically carries the old funnel’s ambiguity into the new one. Set aside a session with each salesperson and reclassify their deals by evidence: a deal with a logged meeting with the decision maker goes to stage 4, and one with only a proposal sent goes back to the stage the evidence supports. Warn leadership in advance, because the value in the late stages usually shrinks at cutover.

Roadmap for leadership

1. Take a snapshot of the current funnel. Compare each stage’s real conversion over the last twelve months with the configured probability and count how many lost deals have a generic reason.
2. Write the stages from the buyer column, with sales, marketing and post-sale leadership, defining evidence and required fields.
3. Test the text on closed deals. Pick twenty won and lost deals from last year and ask two people to classify each one by the new criterion, without talking. Rewrite the stages where they disagree.
4. Configure the CRM after the text is approved. Required fields per stage, loss reasons per stage and regression with a reason.
5. Run the cutover with reclassification, following the checklist above.
6. Review every quarter. Recalculate probability and time and read the loss reasons with the teams responsible for them.

Why does the funnel only predict revenue when the whole team applies the same criterion?

The model in this article fits in a two-page document, and the funnel only works when ten salespeople read “the decision maker received the proposal” the same way on the 28th, with the target still open. The salesperson who learned at a previous company that proposal sent is a stage will move the deal out of habit, and a written criterion does not change that habit on its own. The habit changes when the salesperson practices classification on their own deals, hears from peers which evidence is missing and meets, at the next meeting, a manager who asks for the evidence before accepting the move.

In the Revenue Operations program I lead, the group redesigns the funnel in the room, with sales, marketing and post-sale together. Each participant brings open and closed deals, the group classifies the cases by the new text and rewrites the criterion wherever it left doubt. The manager leaves knowing which evidence to ask for at each stage, and whoever administers the CRM leaves with a configuration approved by all three teams.

What HR and L&D need to assess

The signs of need show up in the CRM: the funnel uses the probability that came configured in the tool, stage conversion swings between quarters with no market explanation, deals jump two stages in closing week and “other” and “no response” dominate the loss reasons. The group brings together sales leadership, whoever owns RevOps or administers the CRM and the marketing and post-sale leaders, who define the handoffs between funnels. Salespeople and SDRs join the classification sessions, because they are the ones who move deals day to day. To measure, set the baseline first: the share of open deals with the stage evidence on record, in a monthly sample, the gap between historical probability and realized conversion and the share of losses with a specific reason. Compare after two quarters.

If your leadership discounts the forecast on its own and nobody can say what separates one stage from the next, that is the work of the in-company Revenue Operations training. I run the program on your operation’s real funnel and deals, with sales, marketing and post-sale in the same room, and the group leaves with the SLA between marketing, sales and CS, which sets the handoff criteria between the three funnels, and with the RevOps metrics matrix, where conversion, time and loss reason per stage get an owner and a review frequency.

Frequently asked questions

What is a B2B sales funnel?

It is the sequence of stages an opportunity goes through between the first qualified interest and a company’s purchase decision. In a funnel that predicts revenue, each stage records something the buyer did, with evidence in the CRM, and the conversion between stages comes from the operation’s own history.

How many stages should a sales funnel have?

For new B2B business, between five and seven open stages, plus won and lost. Short sales with one decision maker fit in five, and long sales with a committee, IT and legal need six or seven. Beyond that, the extra stages tend to be salesperson tasks.

What is the difference between a sales funnel and a sales pipeline?

Both use the same stages. The funnel is the conversion view, showing how many deals move from one stage to the next and where they stall. The pipeline is the portfolio of open deals, with value, stage and date, that leadership reviews every week.

How do you calculate the probability of each funnel stage?

Take the deals that entered the stage in the last twelve months and have already closed, and divide the wins by the total closed. Leave open deals out, recalculate every quarter and split by segment or deal size band when they convert differently.

How do you change funnel stages without losing CRM history?

Export the stage history with dates before changing anything, create the new funnel next to the old one and keep the old one closed for reference. Reclassify open deals by evidence with each salesperson, and use the old conversion as a provisional probability until the new history has enough volume.

Who should take part in sales funnel training?

Sales leadership, whoever owns RevOps or administers the CRM and the marketing and post-sale leaders, who define the handoffs between funnels. Salespeople and SDRs join the sessions where the group classifies real deals by the new criterion, because the funnel only predicts revenue when the people who move deals apply the same standard as the manager.