Mapping your funnel from your own data in an afternoon

4 days ago
3 min read

Updated: 3 days ago

Introduction

The instinct when starting is to find a model and fill it in. That produces a diagram matching an article and not matching your business, with stages you cannot count and boundaries nobody can define. It is then abandoned within a month, which is the correct response to a document that describes somebody else's company.

Building it from your own records takes longer to start and produces something usable. You already hold enquiry logs, quotes, invoices, diaries and email threads, and between them they describe how your last fifty customers actually arrived. Reading that history backwards is the whole method, and it requires no tools beyond a spreadsheet and a free afternoon.

1. Mapping your funnel from your own data starts at the end

Work backwards from customers.

Take fifty recent customers and trace each one back: how did they arrive, what happened next, how long did each step take. The common path across fifty cases is your funnel, and it will not match any template.

2. Use records rather than memory

Memory reconstructs a tidier story.

Emails, quote dates, diary entries, invoices, enquiry logs. The dates matter as much as the events, because elapsed time between steps is one of the most useful things you will find.

3. Note where the paths differ

More than one funnel, usually.

Referrals frequently skip several steps. Repeat customers skip more. Enquiries from advertising behave differently from those from a recommendation. If two paths differ substantially, they are two funnels and should be counted separately.

4. Define each stage as an event

The test of a usable stage.

A form submitted, a call answered, a survey booked, a quote sent, a deposit received. If you cannot say precisely when someone entered a stage, you cannot count it, and the map will produce arguments rather than numbers.

5. Count the failures as well as the customers

The other half of the data.

Take fifty enquiries that never became customers and trace those too. Where they stopped is the finding, and a map built only from successes shows a process with no leaks.

6. Record the time between stages

Frequently the biggest surprise.

Median days from enquiry to quote, from quote to decision, from decision to start. Long gaps are where deals cool, and most owners are wrong about their own numbers by a factor of two.

7. Keep it to five stages or fewer at first

Resolution can come later.

A five-stage map you can populate is worth more than a twelve-stage one you cannot. Detail should be added where the data shows a problem, not everywhere in advance.

8. Write down what you cannot measure

The gaps are informative.

If you cannot tell how many people considered you, or where enquiries came from, that is the recording problem to fix first. Noting it is more honest than estimating and then treating the estimate as fact.

9. Check the map with someone who does the work

The correction step.

The person answering the phone or attending the site knows about steps you have never seen: the second call, the informal approval, the thing customers always ask. Their version is usually more accurate than yours.

Expect the finished map to be untidy. Real funnels have loops, re-entries and paths that skip three stages, and a map that looks like a clean cone is usually one that has been simplified until it stopped describing anything.

Conclusion

Build it backwards from fifty real customers, because your own history is more accurate than any model.

Use dated records rather than memory, separate the paths that genuinely differ into their own funnels, define each stage as an event that either happened or did not, trace fifty enquiries that failed as well as the ones that succeeded, record the elapsed time between stages, keep it to five stages until the data justifies more, write down what you currently cannot measure instead of estimating it, and have someone who does the work correct your version.