Channel Fit

For a long time the only number used to pick a channel was arrivals. Teams ranked their channels by how many people showed up and handed budget down the list from the top. The number appeared immediately, comparison was trivial, and meetings reached a decision without an argument.

The trouble is that the ranking stops at the moment someone arrives. Ravi once ran short video ads for Pagedeck and brought in 800 people a day. That put the channel at the top of the table and the budget followed. Three months later Grace pulled up the revenue chart and the line hadn't moved. Signups were up; the number of people who invited a teammate was three or four a day. Everyone else made one document and never came back. A collaborative document tool is half a product when you use it alone, so people who arrive alone usually stop right there.

Making arrivals and making customers were two different jobs, and the table treated them as one. Channel fit is the name for promoting that distinction into a decision rule. Instead of counting who showed up, it follows how far those people get, weighs what they cost against what they pay, and uses that to decide whether the channel is worth working.

The judgment starts by separating out the people who came through one channel. Blend them into the company average and nobody can say which path brought what.

1. Count how many of them reached the point where the product's value first lands. At Pagedeck that point is creating a document and inviting at least one teammate. If three of 800 daily arrivals get there, that's 0.4 percent.
2. Work out what one of those people costs. Divide the channel's spend by that smaller number, not by arrivals. The cost of bringing in one customer is called customer acquisition cost (CAC), and a channel that looks like $8 per arrival can come to several hundred dollars per person who actually reaches the value point.
3. Estimate what that person will pay before they leave. This is lifetime value (LTV) — $15 a month for an average of ten months is $150.
4. Compare the two. Lifetime value comfortably above acquisition cost means the channel fits; roughly equal or below means every extra arrival deepens the loss. A ratio around three to one is the common rule of thumb for safety.

Those four steps run per channel, never once for the company. A blended acquisition cost averages a channel that fits with one that doesn't and explains neither. Paths where money moves directly, like paid acquisition, are easy to divide up; for search and word of mouth the cost side is the hours people spent and what production was billed at.

The value point in step 1 is worth choosing carefully, because everything downstream hangs on it. Pick something too early, like completing signup, and almost every channel looks fine. Pick something too late, like renewing a second year, and no channel can be judged inside a planning cycle. The useful marker is the earliest action that reliably predicts sticking around.

What makes this hard is the delay. Whether today's arrivals pay anything takes months to learn, and the ad spend leaves today. So teams substitute early signals — the share of people who return three times in the first week, the share who convert to paid in the first month — numbers chosen because they move together with the later result. The final call waits; the budget moves on the signal.

The first place is choosing channels at all. With ten candidates it's tempting to try a little of each, and then no channel accumulates enough data to judge. Picking two or three, giving each a quarter, and running the four steps above works better.

The second is deciding whether to spend more. Once a channel looks like a fit, the natural move is to push more money through it, and this is where fit and channel saturation travel together. Even a channel that fits gets more expensive per person past a certain volume, so fit has to be re-measured at each new spend level rather than settled once.

The third is after the product changes, which is the case teams skip most often. A note-taking tool for individuals that grows into a team tool changes who has to say yes. The path that reached a solo user doesn't reach the manager holding the budget, and last year's best channel can become this year's most expensive one. A price increase or a shift in target customer calls for the same review.

"That channel has the most arrivals and the worst fit."

It brings in plenty of people, but so few become customers that spending more loses money.

"We're capping that channel at $4,000 a month until fit is confirmed."

Drawing a line: no scaling before the judgment is in.

"We raised prices, so channel fit needs another look."

The old judgment doesn't survive a change in the product or its price.

Every quarter Ravi rebuilds the channel table at Pagedeck. Next to the arrivals column he puts the share who reach a teammate invite, the cost per one of those, and the share still around six months later. One or two channels always rank differently across those columns than they do on arrivals, and those are the ones the meeting is about.

Mia reads the same table. When people from one channel consistently stall before inviting anyone, the problem may be the product rather than the channel, so she looks at the first screen those particular people land on.

Grace doesn't open with the arrivals number. She asks whether customers from a channel are still there six months later, and if there's no answer yet she files the channel as unjudged. Unjudged channels get spending only up to a fixed cap.

The most common mistake is reading high arrivals as a good match. Arrivals land the same day while every other number takes months, so the visible one wins trust. But arrival counts are closer to a measure of how many people saw the ad than of how many needed the product. Leave that assumption in place and the team pushes budget toward whichever channel delivers the most people most cheaply, arriving six months later at triple the signups and flat revenue.

Treating fit as a one-time verdict is the other one. The relationship between a product and a channel feels like a fixed property. In practice both sides move. Ad prices rise, a competitor bids for the same placement, and last quarter's arithmetic collapses. Left alone, spending continues on the strength of an old judgment, and the reversal gets noticed two quarters after it happened.