Most trial email sequences are built on the calendar. Day 1 welcome, day 3 tips, day 5 case study, day 7 last chance. Everyone gets the same seven emails in the same order, and the only thing determining which one you receive is how long ago you signed up.
That’s a sequence built around the vendor’s clock. It has no idea what the user has actually done.
So the day-5 “here’s what our power users build” email lands identically in two inboxes. One belongs to someone who has never logged in — being asked to admire a house they’ve never walked into. The other belongs to someone who burned through 80% of their credits on Tuesday, being taught something they already know while their account quietly approaches a wall nobody warned them about. Same email, two kinds of wrong.
The signal that should drive the message isn’t the date. It’s the user’s position relative to the Golden Path — the one action that separates the users who convert from the users who don’t. Sorted that way, every trial user in your database is one of four people.
01 — The Ghost
Signal: account created, no login activity. Goal: first login.
They signed up, then vanished. This is the largest silent bucket in most trials and it gets the least useful mail, because the standard sequence assumes they’re inside the product looking at things. They aren’t.
A Ghost doesn’t need your tutorial. They had genuine intent at signup — something in the next 48 hours buried it. What they need is a reason to come back, not instructions for once they arrive: inspiration over instruction, and one low-friction invitation to do one thing. The blank canvas is the enemy here, and every extra option is another reason to close the tab.
02 — The Explorer
Signal: has logged in, has not completed the Golden Path action. Goal: first Golden Path action.
They’re inside. They poked around. They didn’t commit to anything.
Explorers have crossed the threshold of interest but not the threshold of trust. They want to see the product work before they invest real effort in it, so proof comes before instruction: use cases that match the reason they signed up, escalating to education as the clock runs down.
This group is usually much bigger than teams expect. A B2B SaaS client I’m working with now sees roughly one percent of monthly signups produce any activation signal at all. Ninety-nine percent of their trial list is Ghosts and Explorers — people being emailed about advanced workflows they have never once triggered, when the real work is getting them to one recognizable result in the first session.
03 — The Builder
Signal: has completed at least one Golden Path action. Goal: deeper usage, and the upgrade positioned as the obvious next step.
They did the thing. The sequence should now change completely.
Motivation is finished work for a Builder — they’ve already proven they’ll do the action. Stop selling the idea and start expanding the footprint: show them what to build next, and let their own usage ceiling do the selling. The upgrade should read as the continuation of momentum they built themselves, not a pitch.
04 — The Power User
Signal: high usage, at or approaching the credit or feature ceiling. Goal: a frictionless upgrade.
They’re burning through the product. Drop inspiration and education entirely — every “here’s a tip” email you send this person costs you credibility.
The only job left is operational: make sure their workflow doesn’t stop. Confirm the value they’re getting, show them where they stand against the ceiling, and remove every step between them and a paid plan. Loss aversion is the lever, and it’s the strongest one in the set, because this user has already experienced the value they’d be losing.
Like this? I send one idea like it most weeks. One PLG idea a week → subscribe
These are states, not segments
The distinction matters more than the four names do.
A segment is a list you build on Monday and mail on Thursday. A state is where the user is right now. In the 4-Profile System every user is in exactly one state, and product events move them between states in real time. A Ghost becomes an Explorer the moment first login fires; an Explorer becomes a Builder the moment the Golden Path action fires. When the transition happens, the old sequence stops immediately and the new one starts.
That last clause is the whole thing. The most common failure I find in trial email isn’t bad copy — it’s a user sitting inside three sequences at once, getting a “still haven’t tried it?” nudge on Tuesday and an upgrade prompt on Wednesday. Segments collide by default. A state machine can’t, because there’s only ever one state to be in.
The part nobody builds: compression
Here’s the detail that separates a working system from a diagram.
Someone logs in for the first time on day 5 of a 7-day trial. They just became an Explorer. They cannot receive a seven-email Explorer sequence — there are two days left. So the system calculates days remaining and enters them at the right compressed point: log in by day 2 and you get the full sequence; day 3 or 4 and you skip the identity and problem-framing emails; day 5 or 6 and you get urgency only; day 7 and you get the final ask, alone.
Without compression, late transitions either flood a brand-new user’s inbox with five emails in 48 hours or deliver the “your trial ends tomorrow” email three days after it ended. Compression is what makes the design survive real user behavior, and it’s the first thing cut when someone builds this from a slide.
What it actually costs to run
Eleven product events, and none of them are exotic: account created, first login, the Golden Path action fired and completed, credit thresholds at 50/80/90/100 percent, a 48-hour inactivity marker, and calendar-day markers for the trial window.
Every one has to be firing and verified before you wire a single trigger. That’s where these projects stall — not at the strategy, at the instrumentation. Teams argue about email copy for a week when the honest blocker is that nothing in the product tells them whether a user reached the First Value Moment at all.
Two client sessions this month, in unrelated products, arrived at the same place from opposite directions. One is pushing milestone flags — data ingested yes/no, teammate invited yes/no — into their CRM so the nurture can branch on them. The other is scoring in-product usage so the upgrade prompt fires on the click right after the value moment instead of on a date. Neither team called what they were doing segmentation. Both were building the same spine.
And none of this is personalization. Personalization is a first name in a subject line. This is changing what you say based on what the person did.
Where this sits in the system
The 4-Profile System is downstream of the Golden Path, and it can’t be built before it. Every state boundary in the model is defined by that one action, so a company that can’t name its Golden Path can’t segment by anything that matters — it can only segment by date, which is where we started.
It’s upstream of nearly everything else, though. Product-qualified lead scoring is just the Builder and Power User states with a sales motion attached. Lifecycle email, in-product prompts, and the timing of the upgrade ask all read from the same state. In the 10 Laws of Conversion this is the layer that turns a measured funnel into a system that responds to it, and it’s one of the first things I build in an engagement.
The profiles also do something quieter that I didn’t expect the first time I installed one: they give a client’s whole team a shared vocabulary. When someone says “that’s a Ghost problem” in a standup, everyone knows which users they mean and what those users need next.
So pull last month’s trial list and sort it into four buckets. If you can’t — if you can’t tell a Ghost from an Explorer with the data you have today — that’s the project. The emails are the easy part. And if you can’t yet say what share of your signups reach value at all, start one step earlier — that one number sets almost every other number in your funnel.
Running a trial you’re not happy with? Book a call and I’ll walk your funnel live.
