Gamification

For a long time the working assumption was that a useful product gets used. What actually decided the outcome was when the usefulness arrives. Search returns value in the first second, so it lives up to the assumption. Products whose value shows up weeks later do not. A document tool is only useful once documents pile up and a team moves in; a budgeting app needs three months of entries before its chart says anything.

Those weeks in between were the problem. At Pagedeck, half the people who signed up created one or two documents and never came back after the third day. Nothing was wrong with the product. Nothing showed them where the work they had just done was heading. Someone who created two documents today got back exactly two documents.

Games are built the other way around. Within five minutes a score moves, a bar fills, a new level opens. Gamification is the idea of lifting that structure out and attaching it to a product: fill the empty stretch before the value arrives with signals that come back immediately, so the user survives that stretch.

It works by attaching an immediate signal to behavior that has no payoff yet. Four parts usually get attached.

1. Pick one countable behavior. Documents created, colleagues invited, one entry a day — it has to be something a machine can count, and repeating it has to genuinely benefit the user. Whatever gets picked here is what the user will do more of.
2. Make progress visible. A five-item checklist with two items done is the standard form. People are uncomfortable leaving a bar more than half full, and the closer the end gets, the faster they finish the remainder.
3. Return a signal right after the action — a confirmation, a point, a completed check. The value may be three weeks out; the signal is now. An identical signal every time goes stale quickly, so teams mix in a variable reward, where the content differs from visit to visit.
4. Create something to lose. A streak, counting how many days in a row someone has done the thing, is the usual choice. Once a number has accumulated, stopping starts to feel like a loss rather than a neutral act.

None of these parts starts itself. A trigger — the cue that begins the behavior — has to come from somewhere, since only people already looking at the screen know that two checklist items are still empty. That's why gamification and notification strategy usually travel together.

There's one condition for any of it to work: it has to sit on top of something the user already wanted to do. Show a progress bar to someone who came to organize documents and the organizing speeds up. Attach points to something with no reason behind it and the behavior stops the moment the points arrive. The worse case is subtler. Add points to something a person was doing for its own sake, then remove the points later, and they do it less than before, because the original reason got quietly replaced.

The most common home is onboarding, the process of helping a new user settle into a product. Five getting-started items shown as a checklist, filled in one at a time. This stretch is exactly the one where there is plenty to do and no value yet, so simply showing progress moves completion rates noticeably.

Next are products that only pay off through daily use — language learning, workout logs, expense tracking, anything where a missed day is a real loss. Streaks and unlocking work hardest here. Attach the same device to a tool people need twice a month and what grows is pointless visits made to keep a counter alive.

Third are contribution behaviors people have weak reasons to perform on their own: community answers, profile completion, data entry in internal tools. Points and tiers on answers are the familiar version, and the question to ask first is what exactly will increase. Attach points to the number of answers and short useless answers increase along with the good ones.

Leaderboards rarely fit collaboration tools. Rank people inside one team by documents created and what grows is the document count, not the collaboration. The question that decides whether to attach a device at all is short: if this number goes up, does something genuinely good happen for the user?

"We added the checklist to onboarding and first-document rate went from 38 to 61 percent."

A report that the device actually moved completion, argued with the before and after numbers.

"I get the badge idea, but is that something anyone wants a badge for?"

Asking whether the behavior is one the user already wanted to perform.

"No leaderboard. There's no reason for these people to compete."

A judgment that the device doesn't match how the product is used.

Noa once built a getting-started checklist at Pagedeck. Three items: create a document, invite a colleague, try a template. People who finished two generally finished the third, and return visits on day three rose noticeably.

The next quarter, when Ravi proposed a leaderboard ranking teams by documents written, the answer went the other way. Sofia asked whether a team that writes more documents is a team doing better work, and what happens when a manager sees that table. Neither question had a good answer, so the leaderboard was dropped. Same category of device: the checklist helped people do what they came to do, and the leaderboard was about to invent work nobody needed.

The most common belief is that points and badges turn non-users into users. Watching how absorbed people get in games makes the devices look powerful on their own. What actually moves is the group already inclined to use the product. A badge means nothing to someone with no reason to be there, and inflating rewards to pull those people in mostly recruits users who take the reward and leave. Left uncorrected, this sends teams adding one more device instead of asking why the product goes unused.

Reading a rising number as proof the device worked is the same failure. What rises is whatever is easy to count. Visits and document counts go up; whether those documents were worth writing never enters the table. So after attaching a device, teams watch retention alongside the behavior they instrumented. If people showing up to protect a streak leave for good the first day it breaks, three months of climbing numbers unwind in a week.