Two hundred people started a trial this month. Six of them paid. That single ratio is the number founders stare at, and it is also the least useful number in the whole funnel, because it collapses two completely different failures into one digit. Fix the wrong one and the ratio does not move, no matter how much effort you put into it.
The number hides two questions, not one
The first question is whether people ever reached the point where your product actually did something for them. Call that activation: whatever the specific action is that makes the value obvious, connecting an account, importing a first file, sending a first message. The second question is whether the people who got there decided it was worth paying for. Those are not the same failure, and the fixes for them do not overlap at all.
Say 200 people signed up and only 30 ever completed the core setup step. Your trial-to-paid rate looks like 3 percent. Your activation-to-paid rate is 6 out of 30, or 20 percent. The second number is the true measure of whether your product converts once someone has actually used it, and it is a far healthier number than the first one implies. You cannot see this split unless you are tracking the middle step. If you are only counting signups and payments, you are diagnosing blind.
Pick one event, not a checklist
Founders who do try to measure the middle step often pick the wrong thing to measure. A checklist of five onboarding steps completed is not activation. Activation is the one specific action that, historically, separates the people who end up paying from the people who don't. Sometimes it is obvious: a payments product's activation event is a real transaction processed, not a settings page filled in. Sometimes it takes a little digging through your own early customers to find it, by asking what they had actually done, concretely, in the days right before they decided to pay. Pick that one action, track it separately from every other click in the product, and you have turned a vague sense of engagement into a number you can actually diagnose against.
Once that number exists, watch what it does before you touch anything else. A rising signup count with a flat activation count is the top of the funnel working and the middle of it broken. A flat signup count with activation climbing is the opposite, and it means your outreach, if you are running any, is the actual constraint. Conflating the two is how founders spend a quarter improving the wrong half of the business and conclude, wrongly, that nothing they try moves the number.
Why sending more people through a broken pipe makes it worse
The instinct when a conversion number looks bad is to feed the top of it. Run more outreach, buy more traffic, widen the targeting so more people find the signup page. If the actual problem is that people never get past setup, this does not just fail to help. It actively hides the diagnosis, because a bigger denominator with the same broken middle produces the same discouraging ratio, and now you have also spent money proving nothing changed.
Before touching the top of the funnel, pull the number that matters: of everyone who signed up, what share did the one thing your product needs them to do to see value. If that number is low, you have found the actual leak, and it is nowhere near the acquisition side.
When it is a product problem, say so plainly
An activation failure is fixed by shortening the distance between signup and the first moment of visible value. That usually means removing steps, not adding messaging: a setup flow with fewer fields, a default that works without configuration, one thing to do first instead of five things presented at once. No campaign, no better subject line, and no amount of outreach touches this, because the person already arrived. They tried the thing and the thing did not get out of its own way fast enough. That is a build problem, and it gets solved in the product, by cutting the setup down, not by writing to more people about how good the product is once you get through it.
A value failure looks different. Someone activates, uses the product for real, and still does not pay. Ask the handful of people this happened to what they were expecting that they did not get. Usually it is a specific missing capability that shows up in more than one conversation, or a price that does not match what they were willing to spend for what they saw. Both are answered by changing the product or the packaging, not by chasing them harder.
The one slice of this that is genuinely a pipeline problem
There is a group inside a stalled trial that outreach can legitimately reach, and it is a narrow one: people who activated, who used the product enough to see what it does, and then went quiet without a decision either way. They are not strangers who need convincing that a category exists. They are known people who saw it work and got pulled away by something else before they acted. A personal note that references what they actually did, sent from a real inbox rather than a drip sequence, catches some of that group before day fourteen turns into never.
This is the part of the trial problem Rocketship can actually help with, and it is worth being precise about the size of that part. The same worker that would otherwise go find strangers who match a plain-English description of your buyer can be pointed instead at your own list of activated, non-converted trial accounts, write to each one personally from your Gmail, and read what comes back so a reply at eleven at night gets an answer instead of sitting for a week. If a confused user picks up the phone rather than dig through a help page, your business line can be answered day and night too, for a five-dollar-a-month add-on on top of whichever plan you're on, so that call does not roll to voicemail. None of that manufactures activation that never happened, and none of it invents value the product did not deliver. It closes the gap for the specific person who already got the value and simply drifted, which is a real and recoverable loss, unlike the two failures above it.
Run the split before you run anything else
Before buying anything meant to fix a trial-to-paid number, get the middle step measured. If the drop happens before activation, spend the next month on the setup flow, not on a sales tool. If it happens after activation, talk to the people it happened to and find out what they actually wanted. Only the last, smallest slice, people who got real value and then simply went silent, is a list worth writing to. Treating the whole 200-to-6 ratio as one problem is how a founder ends up buying outreach software to fix an onboarding screen, and paying for it every month without the number ever moving.