The welcome email sequence: the 7-day shape that actually moves new signups

The first email a new signup gets is the highest-engagement message you will ever send that person. Open rates above 50% are routine — meaning more than half the audience actually opens the email — because the brand is fresh in their head and they're expecting to hear from you. The second message drops to about 35%. Number three lands at 25%. By the fifth, you're sending to a room that has already decided whether they're sticking with the product or quietly walking away. Most programs respond to that moment with a feature recitation that assumes neither outcome. The result is predictable: low action, high unsubscribe, wasted attention at the one point in the relationship where attention is the easiest thing in the world to get.

The shape that works is shorter. Fewer messages. Each one tied to a specific decision the new user has to make, each conditional on whether they've taken the previous step, each written like a response to actual intent rather than a generic brand introduction. Five messages over seven days, with real stop rules. That's the outline. The rest of the guide is what goes in each one and why.

How do you know whether the sequence is doing its job? The metric that matters is activation rate within 7 days of signup — the percentage of new users who reach whatever "first real action" means in your product (first purchase, first project created, first integration connected, whatever the equivalent is for you). Watch it as your primary number. Three secondary signals are worth tracking underneath it: open and click rate per message (the drop between message one and message three tells you whether each new email is earning its place), unsubscribe rate (a spike on messages three or four means you're over-sending), and how well users who activated through this sequence retain compared to a baseline. Open rate on its own is noise. Activation is the outcome.

The first email fires the second a new account is confirmed. Most programs try to make this email do five jobs at once: welcome the user, set expectations for future sends, introduce the team, list the headline features, drive the first action. The over-packing costs the one outcome that matters — getting the user to take the first meaningful step inside the product itself.

The structure that performs: a short welcome sentence, one clear statement of what the user should do next (tied to whatever "activated" means for your product), one button, nothing else. Brand-voice introductions, team photos, and feature tours belong in later messages, once the user has some skin in the game. A welcome email is not the place to explain the company. It's the place to get someone to do the thing they signed up to do.

Subject line: write to the action, not to the brand. "Start your first [thing]" outperforms "Welcome to [brand]" by a meaningful margin because it previews what the email is for, not who sent it. Useful test — if the subject line works as a standalone instruction, it's probably the right subject line.

What about a discount? Usually no. Discounts on the welcome email train users to wait for offers and quietly undermine the framing that activation, not transaction, is the goal of the relationship. The exception is ecommerce, where first-purchase incentives are expected and the product is transactional immediately — but for software, content, or community products, a welcome-email discount is almost always a tax on future engagement that you don't see until later.

The day-after email is a proof-of-value moment, not a second prompt. Show the new user what they'll get from the product if they invest the setup time. A screenshot of a real dashboard. Maybe a sample report. Or a named customer outcome. Anything concrete that answers the question the user is actually asking — "why did I sign up for this in the first place?" — without asking for more action.

Users who get a "no ask" email in this slot activate at meaningfully higher rates than users who get a second prompt. The mechanism is almost too human to be interesting: the previous email asked for action, this one builds a small amount of trust before asking again, and the next ask lands harder for it. Most programs skip this email because it reads as a wasted send. It isn't. It's the send that makes the next ask work.

If you genuinely can't picture what this email contains without asking the user for something, your product probably hasn't figured out its aha moment yet — the specific first experience that flips a new user from curious to convinced. The 72-hour aha moment guide covers how to find it.

This is where most welcome sequences fall apart. The "third email" gets sent on schedule regardless of what the new user has done in the meantime — a generic feature tour that reads as oblivious to users who've already taken the first action, and reads as too heavy for users who haven't. Same email, two different audiences, no fit for either.

The fix is a branch — sending route A or route B depending on what the user has done so far. Has the user taken the first meaningful action, even partially? Then this email is the second step toward full activation: concrete, specific to where they got to, written as if you're carrying on a conversation they started. If they haven't taken the first action yet, the message a day later is a re-pitch of the original goal from a different angle. A peer example works better here than a repeated CTA — "here's how Sarah at Acme did it in 10 minutes" lands in a way that "click here to start" never will.

Branching used to require a lifecycle engineer and a small Looker query. It still does, in some setups. Every modern ESP (email service provider — the platform that actually sends your email, like Braze, Iterable, or Klaviyo) supports event-triggered splits, but wiring them up so the right person gets route A and not route B is the work most teams quietly skip.

The Orbit Multi-Channel Orchestration skill handles the state-awareness that makes branching work — without six independent systems firing at the same user on the same day.

This is the right moment for social proof because the user has had enough exposure to be actively deciding whether to stick with the product. Generic testimonials read as weak at this point — the user is mid-evaluation and wants specificity, not slogans.

What works is named, specific, recent, and concrete. A real user, a real outcome, a real timeline, a real use case. "Sarah at Acme used Orbit to audit 340 segments in two hours" lands in a way that "our customers love Orbit" never will, because the first sentence is verifiable and the second one is a sticker. If you don't have named proof yet, a short summary of an anonymised real program still beats a generic quote every time. The pattern to avoid is the carousel of star-ratings and customer logos — it tells the user nothing they couldn't already infer from your homepage.

For users who haven't activated by the seventh day, the final email is short, friendly, and offers a concretely easier path in. "Not sure where to start? Here's the fastest 5-minute entry." The job here is to lower the bar, not to pitch features one more time. Users who activate after the seventh day are a minority — but this is the email that captures the ones who were close and just needed a nudge.

After that, the welcome sequence ends. Full stop. Users who still haven't activated should move into a separate re-engagement track, covered in the winback flows guide.

Do you need to tell users what emails they'll get from you? Not inside the sequence itself. Tell them at signup (expectations up front) and in the preference centre (control after the fact). An email listing future emails reads as filler. The user just signed up. They don't want a schedule. They want to use the product.

Two conditions should stop a welcome sequence and almost always don't. The first is the activation stop, covered above — once the user activates, the sequence ends, immediately, no "oh they'll enjoy the social proof anyway" reasoning. The second is less obvious and more important: when a user has zero opens across messages one through three, pause the sequence for 14 days and resume with a softer check-in instead of blasting them through the back half at full cadence. People who open nothing early are telling you they're not ready. Ignoring that signal produces the vast majority of early unsubscribes you'll see in any onboarding program. A two-week pause costs almost nothing and recovers a meaningful slice of the audience that would otherwise have churned out of the inbox altogether.