Marketing Funnel

When a team treats promotion as one lump of work, the results arrive as one lump too. We spent this much last month and signups went up by this much — and between those two numbers, nothing. Signups rose, so the month was good. They fell, so it was bad. Nobody could say why either happened, which meant nobody could say what to repeat and what to stop.

Meetings kept stalling in the same spot. The founder asks why paid spend doubled while paying customers stayed flat. The marketer answers that impressions tripled. Both statements are true and they describe different points on a path, so the conversation goes nowhere. Nobody had counted how many doors sit between an ad appearing on a screen and a card number being entered, or which door people turn around at.

The funnel is the idea of cutting that stretch into segments and counting the people standing in each one. Fewer people remain at every step down, the shape resembles a funnel, and the name comes from there. Say 10,000 people see an ad, 400 land on the site, 60 create an account, and 6 pay. The moment those four numbers sit in a row, the single complaint "this isn't working" splits into three different diagnoses. If the drop from 10,000 to 400 is unusually steep, the ad promises something the first screen doesn't repeat. If 400 to 60 is the thin part, signup asks for too much. If 60 to 6 is where people stop, they tried the product and never found a reason to pay. The fixes have nothing in common: rewrite the ad, cut signup steps, or change the product or its price.

Splitting the path surfaces a second thing. The people in each segment are in different states of mind. Show a discount to someone hearing the product name for the first time and they close the tab; send "here's a tool you might like" to someone who has been comparing options for two months and both sides waste an afternoon. Treated as one lump, those two people get the same message.

The vocabulary under the marketing funnel splits six ways: how many segments to cut the path into (segments), what to name those segments (models), what to call the people standing in them (lead types), how to sort and grow those people (operations), how to create demand in the first place (activity), and which direction you meet people from (approach).

Segments are the basic cut. The funnel gets sliced into a top, a middle, and a bottom.

1. Top of funnel (TOFU) — the wide opening. People here don't know the product name, and most of them don't know their problem has a name either. Someone typing "can never find last week's meeting notes" into a search box is standing here. That's why the top of the funnel talks about the reader's problem rather than about the product.
2. Middle of funnel (MOFU) — the comparison stretch. These people now know solutions exist and they're lining up candidates. They open feature comparisons, check pricing, and hunt for reviews from people already using the tool. What they need here is a reason you differ from the other tab they have open.
3. Bottom of funnel (BOFU) — the decision point. The list is down to two or three and someone is doing a final check. Case studies, trial accounts, quotes, and security documentation all get used in this segment.

Models decide how many segments there are and what they're called. AIDA — attention, interest, desire, action — is the oldest of them, and it still gets used to order the sentences inside a single ad or email. AARRR splits the path into acquisition, activation, retention, referral, and revenue. The difference is that it reserves two stages for what happens after the purchase. For a subscription product that bills every month, payment is a starting line rather than a finish line, so whether people keep using the product and whether they bring others in belong in the same picture. The five initials spell AARRR, which sounds like a pirate, and the model picked up the nickname from that.

Lead types are the names for people inside the funnel. It's the same person all the way down, but the label changes as they move.

1. Lead — someone who left contact details. The moment a name and an email address arrive in exchange for a document, that person is on the list. The line between an anonymous visitor and a lead is whether you can reach them.
2. Marketing qualified lead (MQL) — a lead marketing believes has a real chance. Three visits to the pricing page, or a document downloaded with a company email address, moves someone up to this label.
3. Sales qualified lead (SQL) — a lead sales has confirmed is worth a call. Sales takes what marketing hands over and filters it once more, which is also where the two teams argue most often, because marketing counts what it passed and sales counts what was worth meeting.
4. Product qualified lead (PQL) — a lead confirmed by product usage instead of by conversation. Someone on the free plan who created 50 documents and invited 8 teammates is ready to buy without being asked. Companies that let people use the product before talking to anyone lean on this label.

Operations is the hand that works the list. Names piling up are just a list, and two jobs move them downward. Lead scoring ranks leads with points: 10 for a pricing page visit, 20 for a company email domain, 30 for a company above 50 people. When a total crosses the line, the lead goes to sales. Without a rule, sales calls the thirty leads that arrived today from the top down and runs out of hours halfway. Lead nurturing keeps the low-scoring people instead of discarding them. It means sending something worth reading every few weeks to someone with no intention of buying today, so the product is still in mind when their evaluation actually starts. Business buyers run on budget cycles, so a lead with no interest in March buying in July is ordinary rather than surprising.

Activity and approach are how people enter the funnel at all. Demand generation goes a step past catching people who are already searching and creates the want itself — naming a problem nobody is looking up yet so that readers recognize it as a problem. That's why its results land months later rather than this week. Approach is a question of direction. Inbound marketing lays down articles and search results and lets people walk in on their own; it takes time and then keeps working after the spending stops. Outbound marketing knocks first, through email, calls, and conference badges. It produces replies the day you start and stops producing them the day you pause.

The six branches run as one chain rather than separately. Approach and activity bring people in, those people land somewhere along the segments, the model gives each segment a name to count under, the lead label describes where a given person currently stands, and operations moves them one segment down. Working on a single branch does move a number: pour effort into the top and lead counts rise, tighten operations and handoffs rise. Whether any of that reaches a payment is only visible when all six are locked together.

"MQLs are up 40 percent from last month, but SQLs are flat."

The upper number inflated on its own. Either the bar for passing leads along got looser, or the people arriving at the top aren't our buyers.

"This campaign is meant to fill the top of the funnel, so it won't show up in this quarter's revenue."

The people arriving now take months to reach a payment, so please don't judge the campaign on this quarter's number.

"People get all the way to the signup page and half of them leave there. We should fix the middle."

Instead of buying more traffic, fix the one segment where the people we already have are falling out.

At Pagedeck, Ravi opens the funnel every Monday morning. Ad impressions, site visits, free signups, and paid conversions sit in four rows, each with the change from the week before beside it. Last month visits climbed 30 percent while free signups stayed put, so Ravi rewrote the first screen instead of adding budget to the ads.

Business inquiries branch off in the middle. When a form arrives with a company name and headcount, Ravi passes it to Sofia, and Sofia reports back how many of those turned into actual meetings. Ten handed over and one meeting means the two of them re-tune the bar for passing leads. In the quarterly review, what Grace reads is these per-segment numbers rather than total traffic. A different slice goes to Mia: when half of new free accounts create no document in their first week, that isn't something an ad can fix, so it lands with the product team.

The most common mistake is reading the funnel as the route customers actually walk. The diagram is an arrow pointing down, so it looks like people pass through in order. Real people don't. A recommendation from a friend skips the comparison stretch and goes straight to payment; someone reaches the pricing page, disappears for six months, and comes back to start over from the beginning. Believing the funnel is a route leads teams to conclude their tracking is broken when the data arrives out of order, and sometimes to enforce the order by blocking later screens from anyone who skipped an earlier one. The funnel isn't a path people walk. It's a set of lines drawn so people can be counted, and the lines are what make a drop visible.

The second mistake is assuming a wider top produces a proportionally wider bottom. The picture invites it — a bigger opening should mean more coming out. That holds only if the rate between segments stays the same. In a funnel where 400 visits produce 60 signups and 6 payments, doubling ad spend to 800 visits should produce 12 payments. In practice, spending more reaches people who fit less well, the rates fall along with it, and 800 visits produce 9. Cost per payment goes up rather than down. Spend the same money lifting visit-to-signup from 15 percent to 20 percent and payments reach 8 on the original 400 visits. Which of those is cheaper is a question only the per-segment numbers can answer, and without them the only move a team knows is a bigger budget.