Sales Pipeline Stages: What Each One Is and What It Costs You

A sales pipeline has 17 measurable stages, not the five a CRM ships with. Each one has a single metric and a single typical failure. The arithmetic is unforgiving: the twelve sequential stages converting at 80% each pass just 6.9% of clicks end to end, which is why a funnel almost never leaks in one obvious place.

• A stage is a countable ratio — a population enters, a smaller population leaves, and the ratio has a name. If you cannot count both numbers, it is a label, not a stage.
• The 17 stages group into four phases — acquisition (2 stages), conversation (5), decision (5), and recovery and reuse (5).
• Five of the 17 have no acquisition cost attached. No-show recovery, cancellation recovery, long-term follow-up, database reactivation and referral or repeat purchase all run on traffic you have already paid for. They are the cheapest deals in the pipeline and the least likely to be measured.
• Fix the stage with the largest relative gain available, not the lowest rate. Taking a 55% contact rate to 75% is worth more than taking a 30% close rate to 36%, because the stages multiply.
• Measure fewer stages at low volume. At 50 observations, a rate near 20% carries a standard error of about 5.7 percentage points, so a move from 20% to 25% is inside the noise.

What counts as a sales pipeline stage — and what does not

A sales pipeline stage is a point in the buying process where you can count the population that arrives, count the population that moves on, and name the ratio between them. That definition is doing real work, because it excludes most of what CRMs call stages. “Negotiation” is not a stage if nothing enters or leaves it on a schedule you can measure. “Nurture” is not a stage; it is a place deals go to stop being counted. An activity — sending a proposal, running a demo — is not a stage either, though the ratio of opportunities that receive one is.

If a stage has no denominator, it is a label on a card, not a stage in a pipeline. The test is simple: can you say, for last month, both numbers and the rate? If not, that part of your process is unmeasured, and unmeasured stages are where the money goes.

The reason the standard five-stage or seven-stage model persists is that it was designed for forecasting, not for diagnosis. Five stages are enough to tell a board what might close this quarter. They are nowhere near enough to tell a sales lead which repair to make on Monday, because four of the five sit after the point where most volume has already been lost.

How it works

How to find the pipeline stage that is costing you most

01

Count all 17 stages

For one full month, write down the population entering and the population leaving each of the 17 stages. A stage you cannot count is a label, not a stage.

02

Turn each into a rate

Divide leavers by enterers. You now have 17 numbers. The product of the twelve sequential stages is your end-to-end click-to-deal rate; the five recovery stages feed people back in rather than multiplying through.

03

Rank by relative gain

For each stage, divide the rate you could realistically reach by the rate you have today. The largest ratio is the stage to fix, not the lowest rate.

04

Fix one, re-read all

Change a single stage, hold the rest, and re-read all 17 after 30 days. Neighbouring stages move together, so attribute the result cautiously.

The 17 sales pipeline stages, their metrics, and what each leak costs

This is the full breakdown. Every row names one stage, the metric that measures it with its denominator stated, the failure we see most often, and what the leak actually costs. Where a figure comes from published research it is linked in the row; where it is arithmetic, you can check it.

#	Stage	The metric (and its denominator)	Typical failure	What the leak costs
1	Ad click to lead	Click-to-lead rate — leads created ÷ ad clicks	The landing page asks for more than the ad promised	You pay the click and keep no record. At $6 a click, a fall from 8% to 6% moves cost per lead from $75 to $100.
2	Lead capture	Form completion rate — forms submitted ÷ forms started	Field count, and a required phone number on a first-touch form	Completion falling from 70% to 55% is a 21% cut to every number downstream of it.
3	Speed to lead	First response time — minutes from enquiry to first genuine contact attempt	Enquiries arrive after hours, on weekends, and faster than a roster can answer	In the Harvard Business Review audit of 2,241 US companies (2011), 23% never responded at all and the average response, among companies that responded within 30 days, was 42 hours.
4	Contact	Contact rate — leads reached ÷ leads worked	One attempt, one channel, one time of day	A straight multiplier on everything after it: 55% to 75% lifts every downstream count by 36%.
5	Qualification	Qualification rate — leads meeting criteria ÷ leads contacted	The criteria live in the rep’s head rather than in writing	Under-qualifying inflates set rate and deflates close rate at the same time, so the two errors hide each other in a blended number.
6	Appointment set	Set rate — appointments booked ÷ qualified conversations	Booking is attempted after the conversation instead of during it	Every qualified conversation that ends without a date on a calendar re-enters the pipeline at stage 4 and pays the contact-rate tax again.
7	Show	Show rate — appointments attended ÷ appointments booked	Booked too far out, with no confirmation sequence in between	Show rate falling from 75% to 60% removes a fifth of all held calls while every cost above it is already spent.
8	Discovery call	Discovery conversion rate — calls with a defined next step ÷ calls held	The call ends on “I’ll send something through”	Ebsta and Pavilion’s 2024 analysis of 4.2 million opportunities found top performers were 412% more likely to have a next step or meeting defined.
9	Proposal	Proposal acceptance rate — proposals accepted ÷ proposals issued	Priced before the decision process is known, so the proposal lands with someone who cannot say yes	In the same Ebsta and Pavilion dataset, 77% of every slipped opportunity featured key objections being raised early in the process.
10	Between-call follow-up	Follow-up rate — open opportunities contacted this week ÷ open opportunities	Nothing happens between call one and call two	Ebsta and Pavilion recorded that more than 7 days of activity with no future activity scheduled reduces win rates by 65%.
11	Second close call	Second call rate — second calls held ÷ opportunities that did not close on call one	“I’ll think about it” is treated as an ending rather than as a stage with its own rate	Most one-call-close businesses have no denominator here at all, which means the largest single population in the pipeline is invisible.
12	Close	Close rate — deals won ÷ qualified opportunities	The denominator is undefined, so the number is not comparable month to month	A close rate quoted over leads rather than over qualified opportunities moves whenever lead quality moves, and tells you nothing about selling.
13	No-show recovery	No-show recovery rate — no-shows rebooked and held ÷ no-shows	Nobody owns the no-show list, so it is worked when someone remembers	The acquisition cost is already sunk. A recovered no-show is the cheapest held call in the pipeline.
14	Cancellation recovery	Cancellation recovery rate — cancellations rebooked and held ÷ cancellations	A cancellation is recorded as a lost deal rather than as a rescheduling task	Ebsta and Pavilion found one cancelled meeting reduces stage progression by 18%, and two cancelled meetings reduce it by 58%.
15	Long-term follow-up	Long-term follow-up conversion rate — deals won from leads older than one sales cycle ÷ those leads	The CRM marks them “lost” and the reason code is blank	In Ebsta’s 2025 GTM Benchmark Report (655,000 opportunities) (figures from the report’s sales-efficiency digest), win rate fell from 18% for deals slipped one week to 3% for deals slipped beyond six months.
16	Database reactivation	Reactivation conversion rate — deals won ÷ dormant records contacted	The dormant list is treated as dead data rather than as an owned, pre-paid audience	Across our own Colliers-era reactivation campaigns we averaged 4.4% conversion on dormant records, with a peak campaign at 8.9% — our record, not an industry benchmark.
17	Referral and repeat purchase	Referral and repeat rate — second purchases or referred deals ÷ customers	There is no stage for it, so it happens accidentally or not at all	The only stage in the list with no acquisition cost and no qualification cost, and usually the only one with no owner.

Every row in that table is a fraction you can calculate from last month’s data. Filling in all 17 takes about an hour, and the gaps you cannot fill are the answer.

The four phases, and the five stages almost nobody instruments

Stages 1–2 are acquisition: turning attention you paid for into a contactable record. Stages 3–7 are conversation: turning a record into a held appointment. Stages 8–12 are decision: turning a held appointment into a signed deal. Stages 13–17 are recovery and reuse, and they are different in kind from the other twelve.

Five of the seventeen stages have no acquisition cost attached to them at all. No-show recovery, cancellation recovery, long-term follow-up, database reactivation and referral or repeat purchase all operate on people you have already paid to reach. The click is spent, the form is filled, the qualification is done. Everything that happens in phase four is margin, and it is the phase that most often has no owner, no field in the CRM and no number in the weekly report.

That is why the recovery phase is where we start on most accounts. The 4.4% average and 8.9% peak from our Colliers-era database reactivation campaigns are what stage 16 produced once it was given a process instead of an intention — our own record, on records the client had already paid to acquire.

Which pipeline stage should you fix first? The relative gain rule

Because the stages multiply, the value of fixing a stage is set entirely by the relative improvement available in it — not by how low its rate looks. That is the whole rule, and it reverses the instinct most teams have.

The relative gain rule: rank each stage by the rate you could realistically reach divided by the rate you have today. The largest ratio is the stage to fix, regardless of which stage has the lowest number.

Here is the calculation end to end, with illustrative inputs you should replace with your own. The 17 stages are collapsed to seven links for legibility; the rule does not change when you expand them.

Link	Rate	Remaining
Ad clicks in the month	—	2,000
Click to lead	8%	160
Contact	55%	88
Qualification	60%	52.8
Appointment set	50%	26.4
Show	65%	17.2
Discovery to proposal	60%	10.3
Proposal accepted	30%	3.1 deals

At an average deal value of $8,000 that is about $24,700 a month, and an end-to-end click-to-deal rate of 0.15%. Now compare two repairs:

• Fix the worst-looking rate. Proposal acceptance is the lowest number on the page at 30%. Lift it to 36% — a genuinely good quarter’s work — and you have made a 20% relative gain. Deals go from 3.09 to 3.71: +0.6 deals, about $4,900 a month.
• Fix the largest relative gain. Contact rate is 55%, which does not look broken. Lift it to 75% by adding attempts, channels and after-hours coverage and you have made a 36% relative gain. Deals go from 3.09 to 4.21: +1.1 deals, about $9,000 a month.

The stage that was not the problem was worth roughly twice the stage that was. It is also why a stage at 95% is almost never the right first move — taking it to 99% is a 4% relative gain — and why a stage at 5% that can reach 10% doubles the whole pipeline on its own.

How many of the 17 stages should you actually measure?

Fewer than 17, unless you have the volume to read them. A rate measured on a small population moves around on its own, and chasing that movement is worse than not measuring it, because it produces confident decisions made on noise.

The arithmetic is standard: for a rate near 20%, the standard error is the square root of 0.2 × 0.8 ÷ n. At 30 observations that is 7.3 percentage points, at 50 it is 5.7, at 200 it is 2.8, and at 1,000 it is 1.3. So a stage needs roughly 200 observations a month before a five-point move is worth investigating.

Monthly leads entering the pipeline	Stages to measure separately	What to do with the rest
Under 50	4 — one rate per phase	Roll them up. At this volume a stage-level rate has a standard error wider than most improvements you could make.
50–200	7–8 — the acquisition and conversation stages, plus close	Track the recovery stages as counts, not rates. “We recovered 4 of 11 no-shows” is a usable fact; 36% is not yet.
200–1,000	All 17	Read them monthly. A five-point change is now worth investigating for the busier stages.
Over 1,000	All 17, segmented by source and by offer	The blended rate stops being informative here, because a change in traffic mix looks identical to a change in performance.

What it costs to instrument 17 stages by hand

Seventeen stages means seventeen denominators, and this is the part that is usually left out of advice like this. In practice it is: a timestamp or status field per stage in your CRM — you will already have some, and you will not have most; one saved report per stage; and one weekly review that reads all seventeen and changes nothing unless a number actually moved.

Building it runs to roughly a day per phase if the underlying events are already recorded, and closer to a week per phase if they are not — because the work is never the report. The work is getting the event recorded consistently by the person who creates it, every time, including on a Friday afternoon. Running it is then a couple of hours a week, most of it chasing opportunities closed with a blank reason code. Those are our own estimates from client accounts, not a published benchmark.

The stages that are hardest to run by hand are not the complicated ones; they are the ones that require someone to act within minutes, or on a date four months from now. Speed to lead, no-show recovery, cancellation recovery and long-term follow-up all fail for the same structural reason: they demand attention at a moment when a human is busy doing something else. Those four are exactly what an AI appointment setting service is built to run, and they are also the four that quietly stop happening in a busy week. Our own speed-to-lead response research summary and show-rate method cover two of them in detail. Whether you run them yourself or hand them over is an arithmetic question: the hours above, against the deals the relative gain rule says those stages are worth.

Why fixing three stages does not multiply into a 12x lift

In our own client work we typically see around a 3x lift in conversion when a business still running 2020-era sales operations — lead lists pulled once a day, follow-up inside business hours, no recovery stages at all — rebuilds these stages for how buying actually works now. Our own component observations run roughly: speed to lead alone about 3x, doubling contact rate about 2x, doubling set rate about 2x.

Those numbers do not multiply, and we would rather say so than publish figures that do not reconcile. Three times two times two is twelve, and we do not see twelve. They overlap heavily: answering in seconds is part of how contact rate doubles, and a higher contact rate is part of how set rate doubles. They are three views of the same repair, not three independent repairs. Treat the components as descriptions of one mechanism, and the headline as what the mechanism produces together.

These are first-person operator observations from client work. There is no published sample size or measurement window behind them, which is why we state them in the first person and never as a guarantee. Keep them separate from the independent research: the response-time finding is genuinely well documented, and Harvard Business Review’s 2011 study of 1.25 million leads across 29 B2C and 13 B2B companies found firms contacting a prospect within an hour were nearly 7 times as likely to qualify the lead as those trying an hour later, and more than 60 times as likely as those who waited 24 hours. The number we publish with a stated method is the 7x average sales lift on our methodology page, which also discloses that the median is closer to 4x.

When does the 17-stage list change for your business?

The list is fixed in structure and variable in which stages carry weight. Four conditions change it materially.

• One-call close businesses compress stages 8 to 12 into a single event. Stage 11 — the second close call — then becomes the largest untracked population in the pipeline rather than a minor one, because every “think about it” lands there with no denominator.
• Outbound-led pipelines have no stage 1 or 2. They gain a list-quality stage in front of stage 3, and their contact rate carries far more of the total variance than an inbound pipeline’s does.
• Long sales cycles (three months and up) shift the weight to stages 10 and 15. On the Ebsta 2025 data above, a deal slipped past six months converts at 3% against 18% for a deal slipped a week, so the follow-up stages stop being housekeeping and become the main event.
• Low-volume, high-value pipelines should not use rates at all below about 50 opportunities a month. Use counts and named deals; the rate will lie to you.

What does not change is the definition. A stage is a population, a smaller population, and a named ratio — and any part of your process that cannot produce those three things is where you should look first, because it is the only part nobody is watching.

Frequently asked questions

What are the 7 stages of the sales pipeline?

The conventional seven are prospecting, lead qualification, initial contact or demo, needs assessment, proposal, negotiation and closing. They are a forecasting model rather than a diagnostic one: five of the seven sit after the point where most volume has already been lost, and none of them has a denominator attached in a default CRM setup. The 17-stage breakdown on this page splits those seven into countable ratios and adds the five recovery stages the conventional model has no room for.

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

A sales funnel describes the buyer’s journey from the buyer’s side and is usually drawn as awareness, interest, decision and action. A sales pipeline describes the seller’s process from the seller’s side, and each of its stages is something a person in your business does or fails to do. The practical difference is measurement: funnel stages are states of mind and pipeline stages are ratios, which is why a pipeline can be audited and a funnel cannot.

Which sales pipeline stage do most deals leak at?

It varies by business, but the stages that most often leak without anyone noticing are the response and follow-up ones, because their failure produces no record. Harvard Business Review’s audit of 2,241 US companies found 23% never responded to a web enquiry at all and the average response among companies that responded within 30 days was 42 hours. That was 2011, so treat it as evidence of a structural problem rather than a current benchmark.

How do you measure a sales pipeline stage?

Count the population that entered the stage during the period, count the population that left it in the desired direction, and divide. Fix the period before you look at the numbers, and use the same denominator every month — most reporting disputes are two people using different denominators for the same word. A stage where you cannot state both counts is not measured, no matter what the CRM dashboard displays.

How many stages should a small business pipeline have?

Four, if you are under about 50 leads a month — one rate for each phase. At that volume a stage-level rate near 20% carries a standard error of roughly 5.7 percentage points, so most of the movement you would react to is noise. Add stages as volume arrives: seven or eight from about 50 leads a month, and all 17 once a typical stage sees 200 or more observations in a month.

What happens to a deal that slips between stages?

Its odds fall sharply and keep falling. Ebsta’s 2025 GTM Benchmark Report, built on 655,000 opportunities, recorded (in the report’s sales-efficiency digest) win rates of 18% for deals slipped by one week, 13% at one month, 8% at three months, 5% at six months and 3% beyond six months. That decay curve is the argument for treating between-call follow-up and long-term follow-up as stages with their own rates rather than as admin.

Every stage, and the page that covers it

Each stage below has its own metric, its own denominator and its own failure mode. These are the pages that cover them.

Fix a stage

• How to Increase Your Conversion Rate: The 11 Levers Between a Click and a Closed Deal
• How to Increase Sales Funnel Conversion Rate: The Conversion Chain
• How to Increase Speed to Lead Conversion Rate
• How to Increase Contact Rate on Inbound and Outbound Leads
• How to Increase Your Appointment Set Rate: The Three Levers
• How to Increase Sales Close Rate: Name the Denominator First
• How to Increase Your Sales Team’s Close Rate — and the Test That Tells You If It’s the Reps
• How to Increase Second Call Close Rate
• How to Increase Follow Up Rate Between Sales Calls
• How to Increase Long-Term Lead Follow-Up Conversion Rate
• How to Increase Proposal Acceptance Rate: The 7 Levers That Move It
• How to Increase No Show Recovery Rate: The 5-1-1-1 Window
• How to Increase Cancelled Appointment Recovery Rate
• How to Increase Database Reactivation Conversion Rate
• How to Increase Conversion Rate From Facebook Ads: The Seven Places a Paid Click Dies
• How to Track Sales Objections at Scale and Use What You Find

Diagnose a bad number

• Why Your Contact Rate Is Low and How to Tell Which Cause It Is
• Why is my appointment set rate low? How to tell which cause it is
• Why Your Sales Close Rate Is Low, and How to Tell Which Cause It Is

Benchmark yourself

• What Is a Good Speed to Lead Conversion Rate? Benchmarks by Industry
• What Is a Good Appointment Set Rate? Benchmarks by Industry
• What is a good sales close rate? Benchmarks by industry
• What Is a Good Database Reactivation Conversion Rate? Benchmarks by Industry

Calculate it correctly

• How to Calculate Sales Close Rate and the Mistake Most Teams Make
• How to calculate appointment set rate and the mistake most teams make
• Compounding Conversion Rate Math: Why Small Gains Compound Into Large Ones
• Speed to Lead Conversion Rate: Manual Follow-Up vs AI-Driven
• Appointment set rate: manual follow-up versus AI-driven
• Old vs Modern Sales Process: 2020 Operations vs 2026, Stage by Stage