TR Start free
HomeBlog › Guide

How to set up sales pipeline stages: how many, what has to be true to leave each one, and what to call them

O Ohana360 Team • September 20, 2026 • 12 min read
Illustration of five sales stages with their exit criteria, beside two panels comparing action-style and state-style stage names

Nine stages. Ninety-four open deals. Eleven columns on the kanban board. And one question nobody could answer on a Monday morning: of the thirty-three deals sitting in "Quote Sent", how many are still alive?

Larkfield Metering in Leeds sells water metering hardware to facilities teams and housing associations. Twenty-six people, four sales reps. When the head of sales counted, eleven of those thirty-three deals were older than six months. Six of the nine stage names described something her team had done: Quote Sent, Sample Shipped, Demo Delivered, Price Given, Following Up. None of them were wrong. Each had been true once, and none of them would ever become false again. That is why the pipeline kept swelling. (Demo data.)

This guide is about the stages themselves: how many to run, the exit criterion each one needs, how to name them, what the probability percentage is for, and what you can automate the moment a stage changes. If the opportunity record and its fields are new ground, start with the sales tracking software guide instead.

Start with your own numbers

Before arguing about stages, count. This is the audit Larkfield ran on day one, and it is the whole brief for rewriting a stage list:

MeasureLarkfield, day oneWhat it tells you
Open stages9Eleven kanban columns: the board scrolls and nobody sees the end
Deals in one stage33 of 94A third in one bucket: that stage is really a waiting room
Untouched over 30 days29 dealsAlmost a third of the pipeline is a ghost, and the forecast follows it
Stages named after an action6 of 9The name can never become false, so records never move themselves
Stage pairs that change together2 pairsEach pair collapses into one: nine stages become seven

That last row does more work than it looks. If two stages nearly always change on the same day, there is no real decision between them, and they are one stage wearing two labels.

How many stages?

The practical answer is four to six open stages. The better question is not how many stages but how many decisions. A stage is the name of a genuine threshold: crossing it requires something to be proven, and once it is crossed the deal is worth a different amount. Any step in between that proves nothing is just another click.

A crowded list bills you straight away in Ohana360. The kanban view gives every stage its own column and adds Closed Won and Closed Lost next to the open ones, so six open stages means eight columns. The stage bar across the top of a record page starts scrolling sideways as the list grows, which means people decide without seeing the last steps. Each column header carries the total amount of the deals underneath it, and a total spread across eleven columns is a total nobody remembers.

Three questions that remove a stage. To keep a stage on the list, all three must be yes. Does leaving this step require something verifiable to be true? Does the chance of winning measurably change here? Is the work you would do for a deal stuck in this step different from the work in the step before it? One no, and the stage merges into its neighbour.

Every stage needs an exit criterion

This is the part teams skip. A stage is not defined by what happens inside it but by what has to be true before a deal can leave it. A stage without a written exit criterion is a box each rep fills to their own taste, and two months later the number in that box means nothing.

A good criterion has three properties. You check it by looking at a field or a file on the record. It rests on a fact, not on how the call felt. And it describes something the buyer did, not something you did. Here is where Larkfield landed after cutting nine stages down to five:

Five stages, five exit criteria Larkfield Metering, Leeds, water metering hardware (demo data) 5 open stages 1 Budget Confirmed Exit: decision maker and budget period on record 10% 2 Requirements Agreed Exit: spec list written, close date confirmed 25% 3 Quote Under Review Exit: buyer has the quote, review date is known 50% 4 Terms In Negotiation Exit: price and delivery accepted in writing 75% 5 Contract With Legal Exit: signed contract sits in the record's Files 90% Closed Won and Closed Lost are system stages: they always come last, sit at 100% and 0%, and cannot be edited.

Notice that every criterion maps to a trace on the record. "Decision maker on record" means a related contact or a filled field. "Signed contract in Files" means there is a document on the record's Files card. Written that way, a stage becomes auditable: in the Monday review, the answer to "why is this still in negotiation?" stops being a debate and becomes an empty box on a screen.

Enforcing the criterion in Ohana360

Writing a criterion is one thing; making it stick is another. Two ways, neither of which needs code.

If the threshold really needs a person's sign-off, say a discount above a certain level, an approval process is the right tool rather than a stage rule; the approval workflow guide draws that line. The content of the quote itself is covered in the business quote guide.

Stage names: a state, not an action

Larkfield's real problem was never the count. It was the names. "Quote Sent" describes your own work. It becomes true the second you press send and it never becomes false. The buyer can decide, go to a competitor or go out of business, and the deal still sits in the same column, because nothing about that sentence ever needs revising. "Quote Under Review" describes the buyer, and the day a decision lands it is simply wrong, so somebody has to move the record. That is how a pipeline cleans itself.

What should a stage name say: what you did, or where the deal is? Two naming lists for the same pipeline (demo data) ACTION NAMES: WHAT YOUR TEAM DID Quote Sent You sent it. What the buyer did is unknown Demo Delivered A calendar entry, not a position in the deal Sample Shipped A courier update turned into a stage Price Given Accepted, refused or ignored? Unclear Following Up Every rep reads this label differently STATE NAMES: WHERE THE DEAL STANDS Budget Confirmed Decision maker and budget period on record Quote Under Review Buyer holds the quote, review date known Terms In Negotiation Price and delivery are being agreed Contract With Legal The draft sits with the buyer's lawyers Closed Won Signed, and an order can be raised The test: can you put the stage name at the end of the sentence "This deal is currently ..."? If you can, the name describes a state. If you cannot, what you have is an activity log, not a stage.

Teach the team one test: put the name at the end of "This deal is currently ...". "This deal is currently under review" is a sentence. "This deal is currently quote sent" is not. When the sentence refuses to form, what you are holding is an activity: sending the quote is an email or a task on the record, not a position in the pipeline.

Two practical notes. Names like Following Up come off the list, because every rep reads them differently and deals that land there never leave. And Closed Lost stays a single stage while the reason becomes a field: do not open three losing stages for No Budget, Lost To Competitor and Wrong Timing. Add one picklist field and make it required in that stage only, using the conditional field trick above.

Probability and forecasting: what the percentage does and does not do

In Ohana360 every open stage carries a probability, and the number lives on the stage rather than on the deal. The mechanics:

Do not invent the percentages. Read them off your own history: of the deals that reached negotiation last year, what share was won? If the answer is sixty per cent, negotiation is sixty, not seventy-five. The number on its own proves nothing; what makes it useful is that the whole team means the same thing by the same stage. If win rates differ wildly between customer types, the right fix is usually to split the segment rather than tune the percentage, which is the subject of the customer segmentation guide.

On the reporting side, Ohana360 ships a ready-made Pipeline by Stage report: opportunities grouped by stage with the total amount per group drawn as a bar chart. In the report builder you can copy it with your own filters, or group by owner or by close month instead.

The limit: there is no weighted forecast that multiplies amount by probability and totals it. The report builder only offers number fields as a sum metric, and a formula field multiplying amount by probability is not selectable as that metric. If you want the weighted figure, export the report and do the multiplication in a spreadsheet.

Seeing the deals that are stuck

The real job of a stage list is to make a stalled deal visible. Two places do that work.

The Deals Needing Attention list on the Sales360 home page. It scans open deals and ranks anything matching one of four reasons: the close date has passed, it closes within seven days, it has had no activity for thirty days, or the amount is large. The reason is printed on the row, so you read "close date 12 days ago" or "no activity for 41 days" rather than guessing. Treat it as a health check on the stage list itself: if the same stage keeps topping that list, its exit criterion is written wrong.

A scheduled flow. One ships switched on: every morning, owners are notified about open deals closing within the next seven days that are neither won nor lost. Among the ready templates under Setup > Flows there is also one that fires three days before the close date with a push notification and a call task for the owner. To build the whole reminder chain deliberately, the automated follow-up reminders guide shows which reminder comes from which record.

If you want to filter for inactivity yourself, know the boundary: saved list views on Opportunities filter on stage, owner, amount, close date and name. You can add Last Modified to the list as a column, but you cannot build a filter on it and save that as a view. The thirty-day rule is computed by the home page panel.

What a stage change can trigger

Stage is the most valuable trigger on the record: it changes often enough to be useful and rarely enough to mean something. In the flow canvas, pick the Opportunity object with the update event and two extra operators appear in the condition list: changed and changed to. The second is what you want, because the rule runs once at the moment the stage lands on that value rather than on every later save.

Stage changeWhat you can automateWhere
Moves to Closed WonReady template: a celebration post on the record's timeline, a notification for the team, and the linked account's Type set to CustomerSetup > Flows > Templates
Moves to Closed WonThe shipped validation rule blocks a zero amount; a Create Order button appears on the record and carries the line items acrossOpportunity record page
Moves to negotiationA push notification for the sales manager; flow actions include Push Notification, Send Email and Create TaskFlow canvas
Moves to Closed LostA Loss Reason field becomes visible and required by condition; the Lost button on the record moves a deal there in one clickEdit Page
Any stage changeWith the Field History add-on enabled, who moved it, when, and from which stage to which is kept on the recordRecord page > Field History

The canvas also gives you a not met branch beside the one that runs when the condition is satisfied, so you can act when a condition newly breaks, for example when a deal written off as lost is reopened.

Where stages are configured in Ohana360

The same work happens on a phone: in the mobile app the Opportunities list switches to the kanban view and a card is dragged between stages with a finger. The details are in the mobile CRM app guide.

What it does not do

The tour walks the whole path, from a lead to an opportunity, across the kanban board and into pipeline by stage:

Week one: from nine stages to five

DayTaskTime
1Run the count: how many stages, how many deals in each, how many untouched for over 30 days, and which names describe an action30 minutes
2Take the last ten deals you won and write down the steps they actually passed through, described by what the buyer did rather than what you did45 minutes
3Agree on five stages and write one sentence of exit criterion for each; every criterion must be checkable against a field or a file40 minutes
4Build them under App Settings > Opportunity Stages, order them with the arrows, and set probabilities from your own historical win rates20 minutes
5Move the open deals: select them in the Opportunities list, apply Change Stage in bulk, then delete the stages you emptied45 minutes
6Make the criteria bite: one validation rule on the stage that clogs most, and a conditional required Loss Reason field for Closed Lost25 minutes
7Wire the automation: enable the Closed Won template, save the Pipeline by Stage report, and read the attention list on the home page20 minutes

At the end of week two, repeat the day-one count. Larkfield's target was a single line: bring deals untouched for more than thirty days from 29 down to under 10. A stage list changed without measuring drifts back to nine within a few months, because opening one more column is always the easiest way to describe an awkward deal. To look at the sales module itself, see Sales360.

Frequently asked questions

How many stages should a sales pipeline have?
Four to six open stages are enough for most teams. What matters is not the count but the number of real decisions: moving from one stage to the next should mean a genuine threshold was crossed. If two stages almost always change on the same day, they are one stage. A crowded list costs you immediately: the kanban view in Ohana360 gives every stage a column and adds Closed Won and Closed Lost beside the open ones, so six open stages means eight columns. The stage bar on the record page starts scrolling sideways too, and a team decides without seeing the last steps.
What makes a good exit criterion for a stage?
A fact you can write in one sentence and verify by looking. "The customer seems interested" is not a criterion; "the decision maker's name and the budget period are on the record" is. A good criterion has three properties: it is checked against a field or a file on the record, it does not depend on the rep's mood, and it describes something the buyer did. Ohana360 lets you enforce it in two places: a validation rule that blocks the save, or a field made visible and required only in that stage.
Why should stage names describe a state rather than an action?
"Quote Sent" describes your own work. It becomes true the moment you press send and it stays true six months later, so a dead deal keeps sitting in it: nothing about the name ever has to change again. "Quote Under Review" describes the buyer instead, and it turns false the day a decision lands, which forces someone to move the record. That is what makes a pipeline clean itself. The quick test: put the name at the end of the sentence "This deal is currently ...".
Can I rename a stage later in Ohana360?
No, the stage list has no rename. The setup screen lets you add a stage, delete one, reorder the list and edit each probability; the name cell is plain text. To change a name, add a new stage with the wording you want, select the records in the Opportunities list and move them with the Change Stage bulk action, then delete the old one. A stage that still holds open opportunities cannot be deleted, and at least one open stage must always remain.
What is the probability for, and is the forecast calculated automatically?
Every open stage carries a probability. When an opportunity is created it takes the first stage's probability, and whenever the stage changes the probability is rewritten from the new stage. It shows on the record page and on the kanban card, and it can be typed over by hand, although the next stage change resets it. Changing a stage's probability in setup does not touch existing records: it applies to new stage assignments only. To be straight about the limit: there is no weighted forecast screen that multiplies amount by probability. The report builder totals amounts by stage, not weighted amounts.
How do I find deals that are stuck in a stage?
The Deals Needing Attention list on the Sales360 home page works this out: open deals whose close date has passed, that close within seven days, that have had no activity for thirty days or that carry a large amount, sorted by urgency with the reason printed on the row. A ready-made scheduled flow also notifies owners each morning about open deals closing within seven days. The limit: inactivity cannot be saved as a list view, because the Opportunities list filters are stage, owner, amount, close date and name.

Let the stage say where the deal is

Five stages, five exit criteria and a pipeline that cleans itself. Stalled deals surface on the home page, and a stage change starts the next piece of work on its own.

Read next