Lifecycle emails that aren't a newsletter

What makes an email lifecycle and not a newsletter?

The cause. A newsletter is sent because it is Thursday. A lifecycle email is sent because a person did something, or conspicuously did not. That one difference changes the copy, the timing, the measurement and the result.

It also shows up in public data. Klaviyo's benchmarks, across more than 183,000 of its customers, report triggered flows at a 5.58% click rate against 1.69% for campaigns, with roughly 41% of all email revenue arriving from 5.3% of sends. The flows are a small share of volume and a large share of outcome. The lesson is not to send more email. It is that an email caused by a person's own behavior gets read because it is about them.

What are the five decisions?

Write down what a new signup is deciding in the days after they arrive. For most software bought by a team, the list comes out close to this.

1. Is this the thing I thought it was? They signed up from a page that promised something. The first email either keeps that promise in one line and sends them to the one screen that proves it, or it welcomes them to the family.

2. Can I get it working with my own data? The step where almost everyone stalls. This email is not encouragement. It is the shortest path through the single setup step your data says people get stuck on, with the option of someone doing it for them.

3. Is it worth getting someone else involved? Either a colleague whose data or permission they need, or a manager who holds the budget. This email hands them the words to use with that person, not a feature list.

4. Now that it works, what do I do with it? The second and third uses. People activate once and then forget, and this is the email that creates the habit rather than the first impression.

5. Is this worth paying for? Sent against what they actually used, in their own numbers, before the trial ends rather than on the last day.

For a consumer subscription the list is shorter and the decisions sit closer together. For a considered purchase with a procurement step, decision 3 splits in two. The count is not the point; the mapping is. If you cannot name the decision an email answers, you have written a newsletter with a different subject line.

What triggers each one?

The rule that fixes most sequences: trigger on the absence of an action, not on the day. "Day 3 after signup" sends your setup email to people who finished setup on day 1, which teaches them to ignore you. "72 hours after signup with the data source still not connected" sends it to exactly the people who are stuck.

In practice each email needs four lines written down before anyone opens the email tool.

• The entry condition. The event, or the absence of an event within a window.

• The exit condition. The event that removes somebody from the sequence immediately. Missing exits are the most common cause of a sequence that annoys customers.

• The wait. Short enough that the decision is still live. For a product used daily, 24 to 72 hours. For one used weekly, a week.

• The one action the email asks for. A single link. If there are two, the email has no measurable result.

What order should you build them in?

Not one to five. Build decision 2 first, the one standing between a signup and the product's first value, because it is closest to the money and because the data on where people stall already exists. Then build 5, the trial-to-paid message, for the same reason. Then 1, the welcome, which is easy and which everybody builds first for exactly that reason. Then 3, then 4.

Each one takes a day or two. Finishing one and measuring it beats having five drafts nobody has shipped, which is where most sequences die.

What goes wrong at each step?

• Decision 1. The welcome comes from a no-reply address and arrives 20 minutes late because it sits in a daily batch. Send it within a minute, from a human address, and read the replies. The replies to the first email are the cheapest customer research in the company.

• Decision 2. No suppression, so people who already connected their data get told to connect their data. Build the exit condition before you write the copy.

• Decision 3. It becomes a feature tour. The thing a person needs to involve a colleague is one sentence they can forward, in their colleague's language, not in yours.

• Decision 4. It is sent too late. By the time somebody has not opened the product in two weeks, the habit email is a win-back email and needs a different argument.

• Decision 5. It becomes a discount. A discount at the end of a trial teaches every future cohort to wait, and it converts people who were going to convert anyway. Lead with what they built or used during the trial, in their numbers.

How do you know whether any of it worked?

Hold out 10% of each new signup cohort from the sequence entirely for 60 days, then compare the two groups on the number the sequence is meant to move. Not opens. Not clicks. Activation rate, trial-to-paid rate, or 90-day retention.

A holdout feels wasteful for about a week and then it is the only thing anybody in the room trusts. Without it, you are comparing people who opened an email to people who did not, which mostly measures who was already interested.

Stop reporting open rates as a result. Apple's Mail Privacy Protection, introduced with iOS 15 in September 2021, stops senders seeing whether a message was opened for everyone who turns it on, so your reported opens now cover an unknown share of your list and misstate the rest. Opens are still useful as a relative signal between two subject lines sent to the same list in the same week. They are not a result.

The number to watch per email is narrow: of the people who entered this email's step, what share took the one action it asked for within 72 hours, and did the holdout take it anyway?