Three ways in, none of them a funnel. We were counting like there was only one.

For weeks the weekly report for an outdoor-living retailer had one headline number on it: leads. Everyone knew what it meant, everyone quoted it, and it was wrong — not miscounted, but incomplete in a way that quietly misdirected the budget.

"Leads" meant price-list requests. Someone fills a short form, asks for the catalogue and the prices, and the sales team gets a name. That is a real signal of intent and worth counting. It is also one of three front doors, and the other two had been arriving all along, unreported.

Three doors, not three stages

The second door is the plain contact form. An open question, no product attached yet — "I have fifty-five square metres of terrace, what would you suggest". Rare, and substantial when it happens.

The third is the shop basket. Someone assembles a specific configuration and sends it as a quote request, with a value attached to it before anyone has spoken to them.

The temptation — and I watched myself reach for it — is to draw these as a funnel. Contact at the top, price list in the middle, quote request at the bottom, conversion rates between them. It is neat and it is fiction. Nobody walks them in order. Somebody can request the price list in March and raise a quote request in May without ever touching the contact form. Somebody else goes straight to the basket on their first visit. And plenty of people simply phone, which puts them in none of the three.

They stand beside each other, not behind each other. The moment you add them into a funnel you have invented a journey nobody took.

This is not a small-business quirk. It is the thing Google's research on the "messy middle" describes and McKinsey's decision-journey work described a decade before that: people loop between exploring and evaluating, leave, come back, and enter wherever they happen to be standing. The funnel is a reporting convenience that we then mistake for a description of behaviour.

What changed when we counted all three

Same spend, same week, same ads. Cost per enquiry fell by roughly a third — purely because the denominator finally included enquiries that had been landing in the business the whole time.

That is not a win and I was careful not to present it as one. It is a correction, and it is exactly the trap I described when a client's ROAS doubled without a single extra sale: a ratio is two numbers, and moving the one nobody is looking at moves the headline just as effectively as real performance does. So both figures stay on the cockpit — the old narrow one and the new complete one — until enough weeks have passed that the comparison is honest.

The genuinely useful finding was hiding in the daily split. The strongest day of the week was a Tuesday, and it was almost entirely quote requests from the basket — before the paid campaigns started running that week. The route with the highest apparent buying intent turned out to be the one least dependent on ads.

Count only the price-list door and you draw the opposite conclusion: that paid media produces the serious enquiries. You would then push budget toward the door the ads happen to touch, and starve the one already producing the ripest demand. That is not a measurement problem any more. That is money.

Say what the number leaves out

The last change was the one I would repeat everywhere. The data note under the report now lists, in plain words, what the figures do not contain: telephone enquiries, walk-in customers at two branches, and anything sent to a salesperson's personal address. Real demand is higher than any of these numbers. The report says so, on the report.

Dashboards rarely lie by stating something false. They lie by leaving things out and letting the reader assume the frame is the whole picture. A clinic counting new patients has the same blind spot in a different costume — it can count bookings and still not know how many people it actually treated.

An honest footnote costs one sentence and buys you the right to be believed about everything above it.

What I still can't answer

Which door actually turns into an order. That needs the feedback loop from the ERP, and it is the next piece of work. My guess is that the quote request looks the ripest and the trade price-list request is worth more over a lifetime — a landscaper who asks for the catalogue is not buying once. But it is a guess, and guesses are what you write down so you can be wrong on the record.

Three doors. Three different kinds of readiness. One report that finally admits it.

Sources & further reading

External
Think with Google: how people decide what to buy lies in the "messy middle"
McKinsey's consumer decision journey, summarised (Smart Insights)

Related posts
Our ROAS doubled. We hadn't sold a thing more.
Every clinic counts new patients. Almost none can count the ones who booked.
Your analytics dashboard is lying to you by leaving things out.
People spend weeks choosing a sofa. Most furniture brands sell like it's one visit.