Self-serve Onboarding

Back when software was installed, having a person involved in rollout was simply how it worked. Someone visited, put the program on a server, created the accounts, and taught people to use it. A single contract ran into tens of thousands of dollars, so that labor fit inside the numbers.

The same approach doesn't hold in a product where one person pays $15 a month. One hour of a specialist's time consumes a sizable share of everything that account will pay over a year. Calling 3,000 free signups one at a time, twenty calls a day, takes 150 days.

So the direction shifted toward having the product handle first-time setup by itself. Build the screens well once and you can take 3,000 people or 30,000 for the same cost, which is the entire case for this approach.

It gets built in four steps. Reverse the order and you end up choosing screen mechanics before deciding on the destination.

1. Pick the single action where a user realizes what the product is worth. At Pagedeck that's written as one on-screen action, like inviting a colleague and editing the same document together.
2. Keep only the steps that have to happen between signup and that action, and strip out the rest. Profile photos and notification settings usually go here.
3. Attach one screen mechanic per remaining step. The three in common use are a product tour, which points at what to click in sequence with tooltips, an onboarding checklist, which lays out the setup items and shows how many are done, and empty state onboarding, which turns a screen with no data yet into guidance.
4. Measure how many people drop at each step and fix the worst cell first.

None of this happens only on screen. The email that goes out the day after signup, and the nudge that goes only to people who have done nothing after three days, are part of the same procedure.

"This tier goes self-serve. The numbers don't work with a person attached."

Each customer pays too little to carry any labor cost.

"We cut the tour from six steps to three and completion doubled."

Removing steps is the main work in this approach.

"Set the email to go only to accounts that have done nothing after three days."

Reach the people the screens missed from outside the screens.

At Pagedeck, Mia and Noa keep this procedure drawn as five cells with a pass rate on each. After three sample documents were placed on the empty screen, the share of accounts creating a first document went from 28 percent to 41 percent.