The Trial Expiry Email Sequence

The trial expiry sequence is the most influential email a SaaS company sends, and most teams write it in an afternoon and never touch it again. It is also the one place where the timing is genuinely fixed, because expiry day exists whether you email or not. That makes it the easiest sequence to build correctly and the most expensive one to get wrong.

Everything below counts backwards from expiry day. Build it that way and the same sequence works for a 14 day trial, a 30 day trial and the 21 day trial your sales team keeps granting as an exception.

The seven emails, counted backwards from expiry day

Four before, three after. The day numbers below assume a 14 day trial, but the offsets from expiry are what you actually configure, so a 30 day trial keeps identical timing on the last four sends.

Email three is the only one that needs to be blunt. Say the date, say what stops working, say what happens to the data, and say how to keep it. Anything softer than that gets skimmed, and users who miss the deadline blame you rather than themselves.

What the usage recap email actually needs from your product

Three categories of number, pulled live at send time: actions taken, people involved, and value created. Everything else is filler. A recap that says you logged in four times is worse than no recap, because it proves you measured the wrong thing.

The outcome number is the one worth engineering for. Loom can tell you how many people watched your videos. Calendly can tell you how many meetings got booked. If your product cannot produce a sentence of that shape, the recap email will be a list of clicks and it will convert accordingly. Fixing that is a product problem, and it is usually the same problem described in activation email sequences.

For accounts where the recap is embarrassing because almost nothing happened, do not send it. Route those users to the stalled branch and send the restart framing instead. Reflecting a failed trial back at someone is a reliable way to end the relationship.

Opt in and opt out trials need different sequences

The card decision changes the entire job of the sequence, and any benchmark quoted without saying which model it measured is unusable. Here is the spread as we would plan against it.

That ChartMogul median of 8.9 percent across roughly 200 products gets quoted constantly with no note about the mix underneath it. If you run an opt out trial and someone benchmarks you against 8.9 percent, they are comparing you to a population that is mostly opt in.

For opt out, the sequence is a cancellation prevention flow wearing a conversion flow’s clothes. Email three should tell the user precisely what they will be charged and when, because surprise charges generate chargebacks, and chargebacks cost more than the conversions they create. Stripe’s own guidance on this has been consistent for years and the compliance direction in most markets is toward more disclosure, not less.

For opt in, the sequence has to build the decision and collect a card in the same three weeks, which is why the value proof email at expiry minus seven matters so much more in that model. We break the numbers down further by ACV band and motion in the trial conversion and email data set.

Two sequences in one: the activated user and the one who never started

This is the branch that earns its complexity. Run the activation check at expiry minus seven and route accordingly, then re-check at expiry day in case someone woke up.

The activated user has created things, invited someone, maybe connected an integration. Their objection is price, procurement or timing. Send the usage recap, name what they will lose, make the upgrade two clicks, and offer to talk to someone about annual pricing or seat counts. Do not offer them a restart, because their clock is not the problem.

The user who never activated has an empty account. Their objection is that they never found out whether it works. Send a restart offer: a fresh window starting whenever they are ready, plus one specific first thing to do. Do not send them to a sales call. A rep spending 30 minutes with someone who has never seen the product produces a bad meeting for both parties and a false pipeline entry for the forecast.

There is a third population worth naming: users who activated, then went quiet for the second week. Treat them as activated for messaging but add the one question from the onboarding playbook, which is the human reply ask. The teardown of how Notion, Figma and Loom handle exactly this window sits in the onboarding email teardowns.

The grace period beats the hard cut off

Turn the account read only for 14 to 30 days instead of locking it. The user keeps seeing their own work, which is the strongest argument your product has, and you keep a live upgrade path instead of a support ticket.

The deletion warning is worth calling out. It reliably outperforms every promotional email in the sequence, and the reason is uncomfortable: loss aversion works. Use it once, honestly, with a real deletion date you will actually honour. Bluffing about deletion and then not deleting is how a team loses the one message that still gets opened.

Extension or discount, and the rule for choosing

Extension, in almost every case. The objection at expiry is usually that the buyer has not finished evaluating, and time fixes that while money does not.

Discounting at expiry does three things, and only one of them is good. It converts a small number of price sensitive accounts, it trains every future buyer to wait for the offer, and it permanently lowers the anchor for the renewal conversation twelve months out. That third one is the expensive one and it never shows up in the campaign report.

Notice that only one row involves money, and it pays a switching cost rather than discounting the product. If your sequence has a percentage off in it as standard, pull it for a quarter and watch what happens. In most accounts we have seen, conversion drops by less than the margin recovered, and the effect on renewal pricing shows up later as pure gain. Model your own version in the trial email conversion calculator before you argue about it in a meeting.

Timing, suppression and the sales handoff

Never send two of these within 36 hours except in the final 48, where email three and email four are deliberately close together. Suppress the entire sequence for any account with an open opportunity and an assigned rep, and suppress emails four through seven for anyone who has already upgraded, which sounds obvious and fails constantly because upgrade events and email jobs run on different schedules.

Three more rules that save trouble. Send in the recipient’s local morning, not yours, because expiry day emails landing at 11pm lose a full day. Keep every one of these in plain text from a named person. And write subject lines that name the date rather than the emotion, which is the pattern that holds up across the SaaS email subject line library.

Billing failures are a separate system. A card that declines on the first charge after an opt out trial belongs in a dunning flow, not this one, and the two should never run at the same time. The mechanics of that are in dunning email sequences.

Measuring it without fooling yourself

Report conversion by cohort of trial start, not by month of conversion. Otherwise a good week in the grace period of an old cohort makes this month look better than it was, and you will optimise against noise.

Three numbers to put on one slide: trial to paid rate split by opt in and opt out, conversions occurring after expiry day as a share of the total, and revenue per trial started. That last one is the honest measure, because it catches the case where an aggressive discount raised conversion and lowered revenue at the same time. Wider engagement and revenue benchmarks per lifecycle stage sit in the SaaS email benchmarks set.

The failure mode to guard against is attributing the whole trial conversion rate to this sequence. A trial converts because the product worked during the trial. The emails move the margin, and the margin is worth real money, but a sequence cannot rescue a product a user never got working. If your never activated share is above half, the sequence is not your problem and the onboarding sequence upstream of it is where the work belongs.

Build it in this order

Ship email three and email four first, the 48 hour warning and expiry day. They take an afternoon and they carry most of the conversion. Add the usage recap next, once you know which outcome number your product can honestly report. Then the grace period path, which is engineering work rather than marketing work and is worth scheduling properly. Restart and win back come last.

If you run a free tier alongside the trial, the upgrade triggers work differently and are covered in freemium to paid upgrade emails. For how this sequence connects to everything either side of it, start at the SaaS email marketing hub and work outwards.