Your sales funnel is a production line

I spend my days modeling tissue-culture production. A few clean cultures grow into a shippable crop, and some material is lost at every stage along the way. Lately I've been building a company around that software, and I keep noticing something. The sales funnel I'm learning to run is the same thing I already know how to make better. It leaks at every stage. It settles into a balance. There is one real difference, and it turns out to be the whole game.

It's one model in two skins. Flip the toggle below and watch where the growth happens: in a subscription funnel it loops back to the top, and in a tissue-culture line it bulges in the middle. Then read on for what you're looking at.

Draw a normal sales funnel and it only narrows. Leads become sign-ups, sign-ups become active users, and active users become paying customers, with some falling away at every step. Draw a tissue-culture line and something different happens in the middle. A few explants multiply into a large group of plants. Then rooting, acclimatization, and loss narrow them back down to salable plants. The production line bulges into a diamond. The funnel stays a triangle. The shapes are different because the growth happens in different places.

In the lab, the product copies itself. The plant literally makes more of itself inside the pipeline. In software, the product doesn't multiply. The customers do, from outside the pipe, when a happy subscriber sends a new lead back to the top. Referral and expansion aren't a stage in the middle of the funnel. They're a loop.

Turn Multiplication down on the tissue-culture skin and the diamond collapses. It gives out before 1×, not at it. Multiplication first has to replace what contamination takes from your stock, so break-even is 1× plus your loss rate. Below that, the working stock shrinks and there's no crop. Only at zero loss does a flat 1× hold steady. On either skin, if you push loss past what inflow can replace, the pool itself drains too. That's a leaky bucket. Keeping the pool full is the whole game, and it works through a different lever on each side.

The rest of the pipeline lines up too, once you split it into the stages a unit passes through and the two forces that cut across all of them. What a lab calls contamination, a founder calls churn. What a lab calls acclimatization, a founder calls the moment a card finally clears. Loss isn't only at the bottom. It comes off at every stage, which is the honest part most funnel diagrams leave out.

A subscription business at rest is in balance. New sign-ups have to replace lost customers just to keep monthly revenue flat. A production line does the same thing: it keeps starting new cultures to replace what contamination takes, so shippable output stays flat. The prize in both is the same. Get the growth engine to outrun the loss, and the pool stops just holding and starts compounding. In software, the name for that is net revenue retention above 100%, or negative churn. It is the same idea as a multiplication rate that keeps itself going. The multiplication just moved from the product to the people.

Because both systems settle into a balance, both have a ceiling you can work out long before you reach it. In Software as a Science, Dan Martell, Matt Verlaque, and Johnny Page call the SaaS version the Growth Ceiling. It is the point where you lose as many customers as you add, and growth quietly goes flat. It comes straight out of the funnel above. Your funnel sets how many new customers you win each month: leads times the conversion at every stage. Churn sets the fraction you lose. The ceiling is where those two balance. This chart shows whichever skin the funnel above is set to, so flip the funnel's toggle and the ceiling switches with it.

Put numbers on it. Say you win 40 new customers a month, which is your leads times the conversion at every stage. If you lose 3% of your base each month, the line stops climbing at 40 ÷ 0.03, or about 1,300 customers. Turn on expansion and it climbs higher. Net churn becomes churn minus expansion, so one point of expansion against three of churn lifts the ceiling to 40 ÷ 0.02, or 2,000. Push expansion past churn and the churn ceiling disappears altogether, but you don't grow forever. You grow until you run out of market. Your total addressable market is the real cap, and the curve bends into an S that flattens there instead of racing off the top. So the ceiling is whichever comes first, churn or the market. It is the same role that bench capacity plays in the lab: a hard limit the growth engine can't argue with. The chart above reads all of this straight from the funnel. Drag the funnel's leads, churn, expansion, or market size and watch the plateau move.

One honest caveat. Customer count is what runs out of market. Revenue can climb a while longer, because each customer can pay more (bigger plans, more seats), so a business with net expansion keeps growing revenue after its customer count has leveled off. The ceiling here is customers. The dollars ceiling sits higher, bounded by the market's total willingness to spend rather than its head count.

Flip the chart to Tissue culture and the same ceiling shows up, this time in plants, and on the pool you hold: your working stock. Multiplication is the growth engine, contamination is the leak, and the working stock settles where they balance, at inflow over loss. Your shippable output is the throughput you draw off that stock each cycle, and once plants ship they leave the lab and the leak can't touch them. It reads straight from the funnel, just like the SaaS view. Drag the multiplication, survival, or loss and the plateau moves.

That plateau assumes you can subculture as much as you want each cycle, but real labs can't. Bench capacity is the harder limit. You can only subculture so many plants a week, so multiplication, which wants to grow fast, gets held back by labor. Contamination taxes the standing stock every week too. The most a lab can hold then settles at capacity × net multiplication ÷ contamination, which is inflow over loss again, the same shape. The most it can ship by a deadline is whatever it has grown to by then. That's exactly the question my production planner is built to answer, using a lab's own measured rates at each stage.

Here's the leverage they share, and one place they part ways. Both run on a pool you hold and keep alive, and that pool leaks. In software it's your subscriber base, draining at churn. In the lab it's your working stock, the plants you keep and multiply from, draining at contamination. That held pool settles at inflow divided by loss, so it grows as one over loss: cut your churn in half and you hold twice the subscribers, cut your contamination in half and you hold twice the working stock. Where they differ is the finished product. A subscriber stays and keeps paying, so churn keeps biting it. A finished plant ships out and is gone, booked and safe, no longer exposed to loss. So in the lab the leak acts on the stock you hold, not on the plants you've already shipped, and your output rides on that held stock. Either way, protecting the pool is what compounds: the top of the funnel adds a little at a time, but the leak divides. That is why cutting loss almost always beats pushing more in. The book's three growth levers, acquisition, retention, and expansion, are the same three knobs I turn on a production line: initiation throughput, contamination control, and multiplication rate.

Here's where I'll be honest about my own funnel. I understand retention. It is the same steady-state math I've worked on for years, and I'm comfortable there. The top of my funnel is where I'm weakest: getting leads to activate, to reach the moment the product earns the next step. In production terms, my rooting stage has the worst conversion rate in the line, and no amount of multiplication later fixes a stage that leaks near the top.

So that's the question I'm taking to people who've done this before. How did you improve the stretch from acquisition to activation in your software funnel, the part between "signed up" and "actually got value"? If you've moved that number, I'd really like to hear what worked. And if this way of seeing the funnel is useful to you, leave your email below and I'll send the next one.