Sales Funnel Calculator
Stage-by-stage conversion with the weakest step, the biggest loss of people, and what a relative improvement at each step is worth in money.
Stages 6
Overall conversion 0.096%
which is the product of every stage rate, not their average
Visitors 48,000
Signups 3,100 · 6.5% of the step, 44,900 lost · 6.46% of the top
Qualified 820 · 26.5% of the step, 2,280 lost · 1.71% of the top
Demo 310 · 37.8% of the step, 510 lost · 0.65% of the top
Proposal 140 · 45.2% of the step, 170 lost · 0.29% of the top
Closed 46 · 32.9% of the step, 94 lost · 0.1% of the top
The weakest step
which Visitors → Signups
rate 6.5%
people lost there 44,900
The biggest loss in people
which Visitors → Signups
people lost 44,900, 93.5% of everybody who entered
note the same step, which makes it the obvious place to start
What 10% more on each step would do
Visitors → Signups 117.2 closed, 71.2 more · $502,569.29
Signups → Qualified 63.4 closed, 17.4 more · $122,705.56
Qualified → Demo 58.2 closed, 12.2 more · $85,855.59
Demo → Proposal 56.2 closed, 10.2 more · $71,870.40
Proposal → Closed 60.0 closed, 14.0 more · $98,784.00
Value
deals closed 46
revenue $450,800.00
gross profit $324,576.00
value of one entrant $6.7620
which sets the most you can pay for one, before any other cost
Overall conversion is 0.096%, which is the product of the stage rates
rather than their average. Every stage multiplies everything after it,
which is why a funnel with five reasonable-looking stages can convert
under one percent.
The weakest step is Visitors to Signups at 6.5%, and the biggest loss in
people is Visitors to Signups. Those are usually different steps, and
which one to work on depends on the cost of the fix rather than on the
number: a bad rate late in the funnel affects few people, and a small
leak at the top affects everybody.
A relative improvement is worth the same wherever it is made, because
the stages multiply: ten percent more of any stage rate is ten percent
more closed deals. Percentage points are not: adding 10% to a 6.5% step
multiplies it by 2.55, which is why the table above is so lopsided
towards the weak steps.
Stage definitions have to be written down. What counts as a lead, an
opportunity or a qualified opportunity varies between teams and between
quarters, and a funnel whose definitions moved is measuring the
definitions rather than the business.
The funnel only narrows if everybody is counted once. Somebody who
returns a month later and buys is a new entrant to most analytics and
the same person to sales, and a funnel that mixes sessions with people
produces rates above 100 percent at some step.
Time is missing from this and matters enormously. A stage with a good
rate and a ninety-day lag ties up pipeline for a quarter, and two
funnels with identical rates behave differently if one takes a week and
the other a season.
One entrant is worth $6.7620 of gross profit. That is the ceiling on
what acquiring one may cost, before sales salaries, tooling or anything
else, and it is the number to hold against a cost per lead.
A funnel is a model, not the customer journey. Real buyers loop, leave
and return, talk to colleagues and arrive already decided. The model is
useful for finding the leak; it is not a description of how anybody
actually bought anything.
Output is valid and updates as you type.
Fix the highlighted fields to update the output.
Stage rates multiply. They do not average. Six stages at reasonable-looking rates give an overall conversion of 0.096 percent, and every stage is a multiplier on everything after it.
That has a consequence worth knowing before optimising anything: a relative improvement is worth the same wherever you make it, because the stages multiply, but percentage points are not. Adding ten points to a 6.5 percent step multiplies it by 2.55; adding ten points to a 45 percent step multiplies it by 1.22. Which is why the table below is so lopsided towards the weak steps.
How to use
- List the stages widest first, as
name : how many reached it. - Add the deal value so the improvements are priced in money rather than percentages.
- Compare the weakest rate against the biggest loss of people. They are usually different steps.
Example
Overall conversion 0.096%
which is the product of every stage rate, not their average
Visitors 48,000
Signups 3,100 · 6.5% of the step, 44,900 lost · 6.46% of the top
Qualified 820 · 26.5% of the step, 2,280 lost · 1.71% of the top
Demo 310 · 37.8% of the step, 510 lost · 0.65% of the top
Proposal 140 · 45.2% of the step, 170 lost · 0.29% of the top
Closed 46 · 32.9% of the step, 94 lost · 0.1% of the top
What 10% more on each step would do
Visitors → Signups 117.2 closed, 71.2 more · $502,569.29
Signups → Qualified 63.4 closed, 17.4 more · $122,705.56
Proposal → Closed 60.0 closed, 14.0 more · $98,784.00
Value
value of one entrant $6.7620
which sets the most you can pay for one, before any other costPitfalls
The weakest rate and the biggest loss are different questions. The weakest rate is where the process works least well; the biggest loss is where most people leave. Which to fix depends on the cost of the fix, and a bad rate at the bottom of a funnel affects very few people.
Write the stage definitions down. What counts as a lead, an opportunity or a qualified opportunity varies between teams and drifts between quarters. A funnel whose definitions moved is measuring the definitions.
Count people once. A visitor who returns next month is a new session to analytics and the same person to sales. Mixing sessions with people produces step rates above 100 percent and a funnel that widens, which this tool refuses rather than calculating.
Time is missing and matters. A stage with a good rate and a ninety-day lag ties up the pipeline for a quarter. Two funnels with identical rates behave completely differently if one takes a week.
Stage-by-stage optimisation has a ceiling. Past a point, the constraint is the audience rather than the process: a funnel converting well from the wrong traffic still fails, and no stage rate reveals that.
The value of an entrant is the ceiling on acquiring one. Not the target. Sales salaries, tooling and the leads that were never real all come out of the same margin.
A funnel is a model, not a journey. Real buyers loop, leave, return, ask colleagues and arrive already decided. The model is useful for finding leaks and it is not a description of how anybody bought anything.
Compatibility
Arithmetic in the browser: nothing is uploaded and nothing is stored.
The overall rate is computed as the last stage over the first and checked in the test suite against the product of the individual rates, which is the identity that makes the whole analysis work. A funnel where a later stage is larger than an earlier one is refused with an explanation, because that means the two stages count different things.
The improvement table adds the same number of percentage points to each stage in turn and propagates the change through every later stage, which is how a funnel actually responds. The money figures use the gross margin, so they are comparable against the cost of making the change.
Up to twelve stages, which is more than any funnel benefits from. Past five or six, the stages usually describe the sales process rather than the customer.