Welcome emails were opened by 83.63% of recipients and clicked by 16.6% in GetResponse's analysis of 4.4 billion messages sent in 2023, against 40.08% and 3.84% for newsletters from the same senders. The attention fades fast: 21.2% of opens and 44.14% of clicks happen within an hour of sending. A SaaS onboarding email sequence is the programme that spends that attention, and most sequences spend it on a calendar rather than on what the user has done.
This guide gives six touches with the trigger, fallback timing, job and copy for each, then examples, benchmarks and deliverability rules. Engagement data comes from GetResponse, Omnisend and Mailchimp, product data from Userpilot and ChartMogul, the one sequence published with its own numbers from PostHog, and examples from Really Good Emails and Knock. Vendor figures are labelled.
Build the sequence on activation events, not a calendar
The usual sequence is a drip on days 0, 1, 3, 7 and 14. It is wrong for about two users in three: it tells people who have already set up what to do next, and tells people who never came back that time is running out.
Userpilot's 2025 benchmark of 62 B2B SaaS companies puts average activation at 37.5% and average time to value at 1 day, 12 hours and 23 minutes; only 19.2% of users finish an onboarding checklist (188 companies). These are vendor figures from Userpilot's own, mostly established, customers, so a young product should expect lower. Our SaaS onboarding statistics post has the full set.
ChartMogul and ProductLed's Conversion Report, a January 2026 survey of 200 software products, sets the frame: 62% run a 14-day trial, the median free-to-paid conversion is 8%, and card-required trials convert at about 30%, more than five times the no-card rate, though only 20% of products ask for a card.
So the sequence has three real anchors: sign-up, the activation event (the action that shows the product's value) and the trial end. Everything between them fires on whether activation has happened, with a day-based fallback for users who do nothing. A sequence that cannot see product events is a newsletter with a countdown.
The six-touch SaaS onboarding email sequence
Six is the ceiling, not the target. PostHog's flow sends at most six over several weeks, and GetResponse's data shows engagement falling with every message added. Users who activate early should see fewer.
| Touch | Trigger | Fallback timing | Job | One CTA | Skip when |
|---|---|---|---|---|---|
| 1. Welcome | Sign-up, after the first session | One hour after sign-up | Confirm the account, name the first action | Do the first action | Never |
| 2. Unblock | Setup not completed | 24 hours with no setup event | Remove the blocker, offer a human or demo data | Finish setup | Setup is done |
| 3. First value | Activation event fires | Day 3: repeat touch 2 instead | Show what the result unlocks next | Use the next feature | Not activated |
| 4. Expand | Activated, still one user | Day 5 | Bring in the team or a second use case | Invite a colleague | Team already added |
| 5. Midpoint | Half the trial elapsed | Day 7 of 14 | Offer a call or extension, ask what is missing | Book or reply | Converted |
| 6. Decision | Three days before trial end | Day 11 of 14, then day 14 | State the plan, the price and what is lost | Choose a plan | Converted |
Source: Octo Agency sequence design; timing anchored to PostHog's published flow (2025), Userpilot's time-to-value data (2025) and ChartMogul's 14-day trial share (2026)
Touch 1, the welcome. Send it after the first session, not the second the form submits. PostHog delays its welcome by one hour so users can send data first, and Knock's review of 16 SaaS welcome emails says the same. Name one action: Vercel's "Ship your first Vercel deployment in minutes" and Close's "Don't overthink your Close trial. Do one thing." do it in the subject line. Loom's is a video from the CEO, the product demonstrating itself.
Touch 2, the unblock. Only for users who have not finished setup, written as help rather than a reminder. Cursor's "Connect your repos to Cursor" and Atlassian's "Reminder: create your first Jira project" name the missing step. PostHog's second version checked for data at 24, 96 and 168 hours, then scrapped the checks because they misfired. Test the event before trusting it.
Touch 3, first value. Fires on the activation event and shows what the first result unlocks: Zapier's "How top teams automate" to someone who has just built an automation, Dropbox's "You can edit files directly in Dropbox" once the first feature is in use.
Touch 4, expand. Miro's welcome email moves in three steps from creating alone to working with a team, with Fortune 100 logos as proof. Send it only to activated single-user accounts; an invite prompt before value is a request for a favour.
Touch 5, the midpoint. Day 7 of 14 is when the human offer lands: a call, an extension, or a plain question about what is missing. monday CRM's "monday CRM is easier than you think" offers both. Send it from a person: PostHog's personal email to engineers reached a 68% open rate and 16% click-through against 56% and 7.4% for the flow as a whole.
Touch 6, the decision. Three days before the trial ends, then on the day. State the recommended plan, the price and what the free tier loses. Apollo's "Your Apollo trial has ended" and Gamma's "Did someone say more?" are the factual close and the upgrade pitch. Link the pricing page; our pricing page statistics post covers what it has to carry.
Timing: what 4.4 billion messages say about when people read
Engagement is immediate or absent. In GetResponse's data 21.2% of opens and 44.14% of clicks arrive within an hour, and 71.03% of opens and 86.05% of clicks within 24 hours. An email that has not been clicked in a day will not be, so the next trigger should not wait for it.
Fewer messages, higher rates. Cycles of one to three messages are 57.9% of all autoresponder cycles and carry the best rates: click-through of 29.39% for one message, 12.98% at three, under 10% from four onward and 3.95% at 19 or more. Part of that is selection, since a short cycle is often a welcome email alone, but the direction holds in every band.
Send windows belong to the audience. GetResponse finds Tuesday strongest, two windows at 4 to 6 AM and 5 to 7 PM, and weekends lower. Send day-based fallbacks in the recipient's morning.
| Email type | Open rate | Click rate | Click-to-open | Unsubscribe | Source and sample |
|---|---|---|---|---|---|
| Welcome email | 83.63% | 16.6% | 19.85% | 0.94% | GetResponse, 4.4 billion messages, 2023 |
| Autoresponder cycle | 51.05% | 5.59% | 10.94% | 0.37% | GetResponse, 2023 |
| Triggered email | 45.38% | 5.02% | 11.07% | 0.38% | GetResponse, 2023 |
| Newsletter | 40.08% | 3.84% | 9.57% | 0.12% | GetResponse, 2023 |
| Technology and high tech, all email | 44.72% | 7.40% | 16.54% | 0.15% | GetResponse, 2023, nearest proxy for SaaS |
| Automated vs campaign, all industries | 38% vs 30.7% | Not given | Not given | Not given | Omnisend, 2025 data, ecommerce-weighted |
| All Mailchimp users | 35.63% | 2.62% | Not given | 0.22% | Mailchimp, data to December 2023 |
| PostHog onboarding 3.0 | 56% | 7.4% | Not given | 0.6% | PostHog, own flow, 2023 |
| PostHog onboarding 6.0 | 52% | 3.1% | Not given | 0.3% | PostHog, own flow, 2024 |
Sources: GetResponse Email Marketing Benchmarks, 2024 edition; Omnisend email marketing statistics, January 2026; Mailchimp email marketing benchmarks; PostHog, February 2025 update
None of the three platforms publishes a SaaS category; GetResponse's Technology and High Tech row is the nearest proxy, and a B2B software audience should expect the upper end of each range. The PostHog rows are the only ones measured on a real onboarding sequence.
Copy rules that held up in the data
One email, one action. Knock's review found Supabase's welcome buried its primary action as the fourth bullet in a list and named that as the one change that would bring more users back. Put the action in a button and in the subject, so a reader scanning a phone inbox knows what the email wants before opening it.
Send from a person once the account has value. PostHog's personal emails beat its designed flow on every published metric, and its latest version sends every email from a named employee. Keep the welcome branded if the product is visual; make the midpoint and the decision personal.
Personalise on behaviour, not first name. PostHog's summary: relevant content at the right time, by branching on the product chosen and skipping emails about features already in use.
Say how many emails are coming. PostHog's unsubscribe notice states one a week for six weeks; unsubscribes stayed between 0.3% and 0.6%, under the 0.94% GetResponse records for welcome emails alone.
Build for a phone. Knock's guidance: single column, buttons at least 44 by 44 pixels, body text at least 14 pixels. Litmus reports mobile as the primary reading environment and dark mode above a quarter of readers. Illustrations beat screenshots in PostHog's tests by only 1 to 2 points of click-through.
Twelve onboarding emails and what each one asks for
These are observed emails from public collections, not performance data; each shows a sequence position done well, or in one case badly.
| Position | Brand | Subject line or headline | What it asks the user to do |
|---|---|---|---|
| Welcome | Vercel | Ship your first Vercel deployment in minutes | Deploy once |
| Welcome | Close | Don't overthink your Close trial. Do one thing. | One first action |
| Welcome | Loom | A special welcome to (first name) | Watch the CEO's video, install the app |
| Welcome, weak | Supabase | Welcome to Supabase | Primary action buried as bullet four |
| Unblock | Cursor | Connect your repos to Cursor | Link a repository |
| Unblock | Atlassian Jira | Reminder: create your first Jira project | Create the first project |
| First value | Zapier | How top teams automate | Copy a real automation |
| First value | Dropbox | You can edit files directly in Dropbox | Try the second feature |
| Expand | Miro | Welcome to Miro! See Our Quickstart Resources. | Move from solo to team use in three steps |
| Midpoint | monday CRM | monday CRM is easier than you think | Take the extension and onboarding help |
| Decision | Apollo | Your Apollo trial has ended | Upgrade or lose access |
| Decision | Gamma | Did someone say more? | Move to a paid plan |
Sources: Really Good Emails, onboarding category, read 2026-11-08; Knock, 16 examples of effective SaaS welcome emails, November 2025
The strong ones share two traits: the subject is the task and the body has one button. The weak pattern is the welcome that lists everything the product does, a brochure the user already read on the landing page.
Measure clicks and activation, because opens are broken
Apple's Mail Privacy Protection pre-loads the tracking pixel on arrival, read or not. Litmus's September 2026 market share data puts Apple Mail, iPhone and iPad at 76.56% of tracked opens, Gmail at 16.10% and desktop Outlook at 3.48%, and says MPP affects roughly 55% to 60% of all opens. GetResponse's data shows the result: one- and two-message cycles record open rates of 103.04% and 103.44%. Stripo's 2026 welcome email report adds that welcome emails are the most affected because they fire the moment someone subscribes.
So the 83.63% is partly machines, and so is yours. Judge the sequence on four numbers.
- Click rate per email, against the rows above. PostHog's working targets, from experience rather than a benchmark: above 4% is acceptable, 6% is the goal.
- Activation within seven days, split by who received each touch. This is the number the sequence exists to move.
- Trial-to-paid conversion against ChartMogul's benchmarks: 8% median, 4% to 6% good and 10% to 15% great for a no-card trial, 25% to 35% good with a card. Our SaaS website statistics post has the trial data by model.
- Unsubscribes and spam complaints: PostHog treats under 1% unsubscribe as acceptable.
Litmus's advice when MPP launched in 2021 still stands: keep a segment of non-Apple subscribers whose opens are real and run subject line tests on it.
Deliverability and the legal lines
Google's sender rules. Since 1 February 2024, anyone sending 5,000 or more messages a day to Gmail must authenticate with SPF, DKIM and DMARC (a policy of none is allowed), send over TLS, and support one-click unsubscribe with the List-Unsubscribe and List-Unsubscribe-Post headers under RFC 8058. Spam complaints in Postmaster Tools must stay below 0.30%; Google recommends under 0.10%. A product with a few hundred sign-ups a day crosses 5,000 once notifications are counted.
Which emails may skip the unsubscribe link. Under the US CAN-SPAM Act a message is transactional only if its whole content confirms a transaction, gives account or security information, or delivers the service, and the FTC reads those categories narrowly. An onboarding email that promotes a feature or a plan is commercial: it needs an unsubscribe link, a postal address, an honest subject line and an opt-out honoured within 10 business days, with penalties up to $53,088 per email. Keep receipts and security alerts in a separate transactional stream so a user who leaves the sequence still gets them.
One system for product and marketing email. Knock and PostHog agree the sequence belongs with product notifications, so it can branch on events and share one unsubscribe state. PostHog left Mailchimp when costs rose after importing more than 30,000 users and moved to Customer.io.
What this means for a team building its first sequence
Instrument the activation event before writing copy. Write the welcome, the unblock and the decision first; they carry most of the clicks and all of the revenue. Add the other three touches once you can see who has activated, cap the sequence at six, and measure clicks, seven-day activation and trial conversion rather than opens. When we build an application in Next.js, we would treat the onboarding event feed as part of the product scope rather than a marketing task added later; our web apps service page explains how we scope onboarding flows.


