Top of Funnel

Back when marketing performance was managed as a single number, every activity got asked for the same thing. Ads, blog posts, and events were all judged on signups produced, so whoever ran them attached a signup request to every surface.

The mismatches repeated. Someone searching for how to organize meeting notes reads an article that ends with a button telling them to pay now. That person learned five minutes ago that document tools exist, and is being asked to buy one. The reverse happened too: someone two months into comparing three products and digging through pricing pages gets yet another introductory email announcing that this product exists. In both cases the sender worked hard and the recipient got nothing, and the report at the end of the month showed effort rather than movement.

The marketing funnel fixed this by cutting the path from stranger to paying customer into stages, and stating that each stage has its own message and its own metric. Top of funnel (TOFU) is the name of the highest one, with middle of funnel (MOFU) and bottom of funnel (BOFU) below it. The name comes from the shape: wide at the top, narrowing on the way down.

It holds the most people and moves the smallest share of them onward. Ten thousand people read an article in a month, four hundred leave contact details, sixty try the product — on raw numbers this stage always looks like a failure.

1. The search terms are problem names, not product names. How do I keep meeting notes organized is the question here; what does this specific product cost is a question from the bottom. People at this stage haven't even assembled a shortlist, which is why naming competitors to them lands as noise.
2. The metrics differ. Reach, the number of people who saw something, impressions, the number of times it appeared on a screen, and brand awareness, how well the name is known and recalled, are the readings. Measure revenue here and it comes back near zero every time, no matter how well the stage is actually doing its job.
3. The channels are fairly fixed. Search engine optimization (SEO), which tunes a site and its writing so it surfaces in search results, and content marketing, which brings people in with writing worth reading, carry most of the load.
4. Results arrive late. A single article takes four to six months to settle near the top of a search result. In exchange, once it settles it keeps bringing people in without being paid for, which advertising never does — stop the spend and paid traffic stops the same day.

That makes the stage hard to evaluate quarterly. Demand results in three months and whoever owns it starts writing about high-volume topics, which lifts pageviews and brings in people with no connection to the product. The metric improves and the pipeline does not.

There is one thing the judgment rests on: whether the person came back.

Three signals confirm it — contact details left, the same person reading a second article, or a visit to the pricing or feature pages. Someone who leaves contact details is a lead, and the device that trades something useful for a name and email is a lead magnet. A downloadable bundle of meeting-note templates gated behind an email address is the standard shape.

The frequent mistake here is setting the threshold too high. Ask someone who just recognized the problem to book a demo and most of them leave; it's asking for thirty minutes on a call from a person who doesn't yet know what to ask. Keep the threshold to something that takes five seconds, like one template download or a newsletter subscription, and save the demo request for people who already have a shortlist.

The matching principle is not dragging people down who haven't moved. Of ten thousand, ninety-six hundred do nothing this month, and a good share of them come back six months later. This stage's job ends at making sure a name is there to surface when they do, which is why the work reads as slow even when it is working.

"This article is for TOFU, so I left pricing out of it."

A judgment that the reader isn't at a stage where buying is on the table.

"TOFU pageviews tripled and leads are flat. I think the topics are off."

Raising the problem that the arrivals don't have the problem the product solves.

"This campaign is TOFU, so don't evaluate it on signups."

Asking for the stage's own metrics to be used.

At Pagedeck, Ravi filled this stage with eight articles about keeping meeting notes organized. Four months in, monthly visits reached twelve thousand, and a button under each article offering a bundle of meeting-note templates collected 380 email addresses a month.

The first thing Grace asked on seeing those numbers was about revenue. Six of the 380 paid that month, and the suggestion in the room was to shut the whole effort down. Ravi's counterargument was time: counting the paid conversion rate of people who arrived through articles six months earlier gave 1.5 percent in the first month and 9 percent cumulatively by month six. Since then this stage gets read cumulatively rather than on the month's payments.

The most common misreading is that a lot of visitors means the stage is working. It's where the numbers move most, so the impression is easy to form. But what needs to grow here isn't the count of people, it's the count of people who have the problem. Write about broad high-volume topics and visits climb to a hundred thousand while the number leaving contact details never moves. Put pageviews and lead counts side by side and the gap shows up immediately, usually as a lead rate falling while the traffic chart rises.

Demanding revenue from this stage is the other one. Every activity in the company gets judged on the same basis, so the request comes naturally. But evaluate an activity whose results land in four to six months on a quarterly cycle and it records a failure every time, and the budget usually gets cut around month six. Cutting it leaves the existing articles standing with nothing new accumulating, so a year later the company is back to zero traffic the moment paid advertising stops.