SaaS onboarding is everything between a new sign-up and the moment that user gets real value from the product: setup steps, first data, the first finished task.
The best practice with the most test support in our database is short and narrow.
Cut the setup down to the one or two actions that lead to value, show users how far along they are, and put real data in front of them on day one.
That's what the tests below keep showing. They're also the reason this page is short on rules.
Our database has only a handful of quantified web SaaS onboarding tests, and most of them changed several things at once.
So we say how strong each piece of evidence is, along with the number.
What product onboarding is, and what to measure
Product onboarding is the in-app part: what users see after they create an account.
It's separate from the sign-up flow that comes before it and from the trial-to-paid push that comes after.
This page covers web SaaS. Mobile apps have their own patterns, like permission prompts and paywalls in the first session, which we cover on app onboarding.
The metric matters more than the design. Most teams pick one of three:
- Activation rate: the share of new users who complete a defined set of setup actions. It's quick to read, but you choose the definition, so you can make it look good.
- Funnel completion: the share who finish the onboarding steps. It's useful for spotting where people get stuck, but it's the easiest of the three to inflate.
- Trial-to-paid or retained users: slower and noisier, but it's the number the business actually depends on.
Patrick McKenzie's 2012 full redesign of Bingo Card Creator shows why this matters.
The new site and app, which added a step progress bar, raised trial sign-ups on AdWords landing pages from 20.9% to 23.1% and lifted user success by 15%.
Purchases stayed flat (50 vs 46). Patrick McKenzie's explanation was that fewer users hit the trial limit.
A better onboarding metric didn't turn into more revenue.
So track an activation metric to steer the work, but judge the test on paid conversion or retention whenever you have enough traffic.
If you can't reach a sound sample size on revenue, say so in the write-up and don't pass off an activation lift as a revenue lift.
SaaS onboarding best practices, backed by tests
Each practice below comes with the tests behind it and how much weight they can carry.
Browse the full set on the SaaS onboarding tests hub.
1. Give trial users one checklist of setup steps
This is the pattern with the most support in our data.
AURUM, maker of the legal software Astrea, added a floating checklist that guided trial users to add a task, process a clipping and add a fee, with a "First Steps" progress indicator.
Activation rose 187%, and the key setup events went up 21% and 25%.
One caveat: AURUM introduced it as a controlled rollout, not a strict A/B split.
VWO ran the same idea on its own trial.
A central "Get Started" tab pulled scattered setup tasks into one step-by-step checklist covering account setup, code installation and feature discovery.
Free trial to paid conversion rose 127% in a 25-day split test.
That's a vendor testing its own product, and the new flow included other UX changes, so treat the 127% as the effect of the whole package.
Brainshark is a third, looser example. A three-step setup wizard with checkmarks, rolled out together with targeted in-app tips and upgrade offers, came with a 150% rise in enterprise free trial sign-ups.
With that many changes at once, you can't credit the wizard alone.
Why might checklists work? One candidate is the endowed progress effect.
In a 2006 field study by Joseph Nunes and Xavier Drèze, car wash customers whose 10-stamp loyalty cards came with 2 stamps already filled in redeemed at 34%, against 19% for customers with plain 8-stamp cards.
That was one study, of loyalty cards rather than software.
It suggests why marking "account created" as step one could help, but it doesn't promise a lift.
See goal gradient and the Zeigarnik effect for the wider research and its limits.
2. Get to real value faster, and cut steps that don't lead there
The most convincing test here is a product change, not a design change.
Legal clippings used to show up in Astrea up to 48 hours after sign-up.
AURUM started pulling in recent clippings from before sign-up so they appeared on first login.
The share of trial users engaging with a clipping rose by 16 percentage points (125% relative).
Nothing about the screens changed. Users just saw their own data right away.
AURUM's first-run redesign points the same way. It removed large illustrations, merged setup into one continuous flow with a progress bar, and explained why the mandatory bar-association step was there.
Activation during the 10-day trial rose 105% (7 percentage points).
It was a bundled redesign, so it shows the direction rather than which part did the work.
The practical test is to list every onboarding screen and ask whether it moves the user closer to their first useful result.
If a step only collects data for marketing, move it after the user has seen value.
Our friction entry covers how to tell useful friction from waste.
3. Fix the step where people get stuck before you redesign everything
DashThis used session recordings to find a weak point in its trial onboarding.
The buttons for adding an integration were pale and sometimes below the fold on small screens, in a hard-to-scan list of 30+ integrations.
Making the buttons bold and always visible, and reorganizing the list, came with a 50% rise in onboarding funnel completion and 25% more first-step completion.
That was measured before and after over about 10 months, with several small changes shipped together, so it's weaker than a split test.
The lesson is about method. Find your biggest drop-off between steps, watch recordings of that one step, and fix what you see.
It's cheaper than a redesign, and you'll learn more.
4. Guide users inside the product, not only by email
Citrix tested a guided walkthrough inside the GoToWebinar free trial against the same 30-day trial with no in-product help.
The walkthrough had step-by-step pop-ups, key selling points and a progress toolbar, and the welcome email linked to it.
Trial-to-paid conversion doubled (+100%), product usage rose 40%, and 61% more users reached the audio setup page.
The walkthrough worked because it pointed at specific setup tasks.
A tour that only lists features is a different thing, and none of our tests support it.
User onboarding best practices that depend on your model
Some onboarding questions don't have one answer. They depend on your pricing, your trial and who signs up.
The trial limit is part of onboarding
Patrick McKenzie cut the Bingo Card Creator free trial from 15 cards to 8.
Free trial to paid conversion went from 1.04% to 1.89%, at 99% confidence with a sample in the 5,000s.
A tighter limit got users to the upgrade point while they were still engaged.
It also fits the 2012 redesign result above: when onboarding gets better, users may never reach your limit, so check both together.
The offer itself can matter more than any screen.
Acuity Scheduling replaced its free-forever plan with a 14-day full-feature Pro trial with card details up front, and paid Professional sign-ups rose 268%.
That changes who enters onboarding in the first place.
Our free trial conversion guide and freemium vs free trial page go deeper.
Personalize by role or goal only if the paths really differ
Asking users at sign-up what they want to do, then showing a matching checklist, is common advice.
Our database has no clean A/B test that isolates it for web SaaS.
The Brainshark result above bundles targeting with other changes.
If your users have clearly different first jobs to do, a one-question fork is worth testing.
If they don't, it's just one more step.
See segmentation for how to pick segments worth targeting.
Don't hold back value to force a habit
OnSwitch, a freemium lighting app, locked some popular free content for the first three days and then unlocked it with a push notification.
Users came back slightly more often but bought less, and a repeat with different content gave similar results.
It's a mobile app, but the lesson fits SaaS too: a return-visit metric can go up while revenue goes down.
How we'd run your next onboarding test
| Step | What to do | Why |
|---|---|---|
| Define activation | Pick the 1-3 actions that come before retention in your own data | A checklist only helps if its steps are the ones that matter |
| Find the drop-off | Use step-by-step funnel data plus session recordings | DashThis found its fix by watching one step |
| Change one thing first | A checklist, a faster path to data, or one simpler step | Bundled tests give a number but no cause |
| Judge on money | Report activation, then trial-to-paid or retention | Bingo Card Creator's redesign raised sign-ups while sales stayed flat |
For more tests from software companies, see the SaaS CRO hub, our roundup of SaaS A/B test results, or search all tests.