How to run SaaS onboarding calls that lift activation

A SaaS onboarding call lifts activation when it ends at your activation milestone, reached on the customer’s own data, rather than at the end of a product tour. Offer it inside the first 48 hours, then judge it against a holdout group. From a 30% baseline, proving a 6-point lift takes about 963 accounts in each arm.

What does an onboarding call need to do to lift activation?

An onboarding call lifts activation only if it moves the account across the one milestone that predicts retention. Rachitsky and Timen offer a useful test for that milestone: users who reach it should retain at least twice as well as users who don’t. “Watched the demo” usually fails that test; “imported their own contacts and sent the first campaign” usually passes it.

Most onboarding calls are tours: the CSM shows features and the account ends the call in the state it began. An onboarding call that ends before the milestone is a demo with better manners. This page treats the milestone as the finish line and builds the call backwards from it.

Unblocking a trial that has stalled mid-evaluation is a separate job, covered in our guide to mid-trial outbound conversations and trial-to-paid conversion. This page is about the first call, before anything has stalled.

How to run the First-Outcome Call, step by step

The First-Outcome Call has one rule: the call ends when the customer has produced their first real output, on their own data, with their own hands on the keyboard. Here is the sequence, starting today.

1. Write the milestone as one sentence before you book a single call (for example, “the account has sent one live invoice to a real client”). If two people on your team write different sentences, settle that first.
2. List what has to be true before the milestone is possible: data imported, an integration authorised, a teammate invited, a template configured. Those prerequisites are the call agenda. Product features are not.
3. Ask the customer to bring the real thing in the booking confirmation: the CSV, the admin login, the colleague who owns the integration. A call where the customer turns up without their data cannot reach the milestone.
4. Pull the account’s event log ten minutes before the call. Open with what you can see (“you started the import yesterday and it stopped at field mapping”), not with “tell me about your business”.
5. Hand over control. The customer shares their screen and does each step while you direct. Whatever they do themselves on the call, they can do again on Thursday.
6. End at the milestone, or book the blocker. If something outside the call is in the way (an admin who isn’t there, a file that doesn’t exist yet), book a 15-minute follow-up within 48 hours with that person invited.
7. Log one outcome code per call: milestone reached, blocked (and by what), no-show, or not a fit.

Finish state for week one: a written milestone, a prerequisite checklist, a booking link with the “bring this” note, and an outcome-code field in your CRM.

Which new SaaS accounts should get an onboarding call?

An onboarding call is a scarce resource, so give it to accounts where a human can change the outcome. That rules out accounts already past the milestone and accounts outside your target customer.

If you get more signups than your team can triage by hand, the routing logic is set out in our three-lane triage for chasing SaaS leads without hiring SDRs.

How do you get new accounts to book, and turn up to, an onboarding call?

An onboarding call only lifts activation for accounts that actually attend it, so the booked rate and the show rate decide how much effect you can ever measure. We know of no credible public benchmark for the booking or show rate of SaaS onboarding calls specifically, so treat any figure you see quoted as somebody’s own funnel, not a norm.

What reliably moves both rates is the same mechanics that move any appointment:

• Offer the call at the moment of intent: on the post-signup screen, again at the first failed setup step, and once more in a short message the next morning. One offer buried in a welcome email is not an offer.
• Chase unbooked target accounts by phone or text the same day.
• Two reminders: one the day before, carrying the “bring this” checklist, and one an hour before.
• Rebook no-shows within the hour rather than letting the slot lapse.

Booking and reminding is appointment-setting work, and it is the part of the process that is easiest to hand to automation or to an AI appointment-setting service. LeadsNow has booked 50,769+ AI-booked sales appointments since 2017. The call itself still belongs to someone who knows the product.

Why “attendees activated more” proves nothing

Comparing attendees with non-attendees always flatters the onboarding call, because the accounts that book and turn up were already the motivated ones. They would have activated at a higher rate with no call at all.

The fix is a rule we call measure the offer, not the attendee. Each week, randomly hold back a share of eligible accounts (half is simplest) and don’t offer them a call. Then compare activation across everyone offered with everyone held back, attendees and no-shows alike. Statisticians call this an intent-to-treat comparison.

The cost of honesty is dilution. If only about 30% of offered accounts actually hold a call, a 20-point lift among attendees shows up as roughly a 6-point lift across the offered group, and small lifts need large samples to detect.

How many accounts do you need to prove an onboarding call works?

This is the worked calculation, end to end. The inputs are labelled assumptions, because no credible public source exists for onboarding-call booking rates, show rates or the effect of a call on attendees. Replace them with your own. The baseline is 30%, the SaaS-only median in the Rachitsky and Timen activation survey. Sample sizes use the standard two-proportion formula at 95% confidence (two-sided) and 80% power.

To check the arithmetic: in the mid band, 40% × 80% = 32% of offered accounts hold a call, and 32% × 20 points = a 6.4-point lift across the offered group. At a 30% baseline that lift needs 849 accounts per arm, which is 849 ÷ 300 = 2.8 months of signups. The capsule figure of 963 per arm is the same formula applied to a flat 6-point lift.

Booking and show rates decide whether an onboarding-call programme can be proven inside a quarter. In the low band, the honest result after one quarter is “we can’t tell yet”. Under about 300 target signups a month, even the mid band takes more than five months to read. Run the call if customers want it, but don’t claim a number.

If activated accounts still aren’t paying, the leak sits further down the funnel. Our guide to why trial-to-paid conversion is low splits that cause out.

What running onboarding calls yourself really costs

A First-Outcome Call costs about 45 minutes of skilled time: ten minutes preparing from the event log, thirty on the call, and five logging the outcome. A no-show still costs the preparation. In the mid band above, that is 96 held calls and roughly 72 hours a month. The low band is about 39 hours and the high band about 112 hours, before any time spent chasing unbooked accounts.

• Skill: the caller has to be able to fix a failed import or an integration permission live. A rep who can only demo will run tours.
• Tooling: a scheduler with reminders, an event feed from product analytics into the CRM, and an outcome-code field. Most stacks already have all three.
• What breaks at volume: chasing. Booking 120 calls from 300 offers (96 of them held) means reaching the other 180 accounts more than once, and that work is usually the first thing a busy CS team drops.
• The hidden cost: holding back half your eligible accounts for a quarter. That cost is real, and it is the price of knowing whether the calls work.

Where onboarding calls sit among the other lifecycle moments a live conversation can rescue (signup stalls, failed payments, cancellations) is mapped in our SaaS lifecycle outbound playbook.

Frequently asked questions

What is a good activation rate for a SaaS product?

In the activation benchmark survey by Lenny Rachitsky and Yuriy Timen (500+ responses, published October 2022), SaaS-only products averaged 36% activation, with a median of 30%. The authors treat the 60th percentile as good and the 80th as great. Your own rate depends on how strictly you define the milestone, so compare definitions before you compare numbers.

How long should a SaaS onboarding call be?

Thirty minutes is our working length for a First-Outcome Call. That is a recommendation, not a benchmark. The right length is however long it takes the customer to reach the milestone on their own data. If that regularly takes more than an hour, the product’s setup is the problem, and no call length will fix it.

Should sales or customer success run onboarding calls?

Whoever can fix a setup problem live should run them. Human help in self-serve funnels is common: ChartMogul’s 2026 SaaS Conversion Report found that 80% of free-trial products add human touchpoints when an enterprise user enters the trial. Booking and reminders can sit with anyone, or with automation. The call itself needs product depth.

Can I compare activation for people who attended the onboarding call with people who didn’t?

No. Accounts that book and attend were already more motivated, so they would have activated more without the call. Compare everyone offered a call with a randomly held-back group instead, counting no-shows in the offered group. That intent-to-treat comparison is the only one that isolates the effect of the call.

What if most new accounts never book the onboarding call?

A low booking rate dilutes the effect you can measure. In our worked example, a 25% booking rate and a 70% show rate turn a 20-point lift among attendees into a 3.5-point lift across all offered accounts, and proving that takes 2,776 accounts per arm. Fix the offer and the chasing before you judge the call.