Why We Refused to Add Points, Badges, and Streaks

The Prompt

We were looking at gamification because our retention numbers were bad — 17% of activated users edited a note after day one, 8% after day seven.

Gamification is the obvious lever. Points, badges, leaderboards, streaks. Every productivity app has them and they are well understood.

They are also, by a wide margin, the most documented failures in the category. The three mechanics teams reach for first are the three with the longest record of not working. So before adding any of them, we tried to work out what the successful implementations were actually doing, because it is clearly not "points."

The Reframe

The useful version is not "add better gamification." It is:

> Make progress toward the user's real outcome visible, easy to finish, and satisfying to close.

Every part of that sentence is doing work. The user's real outcome, not a proxy for it. Visible, so they can see where they are. Easy to finish, so the loop closes. Satisfying to close, so closing it feels like something.

Notice what is absent: any external reward. The mechanism is a small open loop tied to work the person already wanted to do.

Naming Our Actual Outcome

For OmniCanvas, the outcome is not "create notes" and it is definitely not "open the app." Those are our metrics, not the user's goals. Nobody has ever wanted to have created a note.

The real outcome is turning captured material into decisions and next actions. Somebody records a meeting because they want the meeting to produce something. The recording is instrumental. The decision is the point.

Once that is written down, the design follows.

What We Built

A lightweight per-meeting "close the loop" flow with four steps:

1. Recording captured
2. Transcript ready
3. Recap and action items reviewed
4. Next step checked off or added to the calendar

The capture package shows something like "Meeting closed: 3/4." Steps can be skipped when irrelevant. The whole mechanic can be dismissed. Completion gets a subtle acknowledgment and then gets out of the way.

The progress indicator is not the interesting part — it is the most replaceable element of the design. What matters is that it is a small, achievable, dismissible open loop attached to real work. Three of those four adjectives are usually missing from gamification implementations.

The other thing worth noting: we already had most of it. The product generates action-oriented capture packages with checkboxes, places them beside recordings automatically, and supports interactive checklists in the editor. This was assembling an existing workflow into a visible loop, not building a gamification system. That is generally a sign you are on the right track — the mechanic should be revealing structure that is already there, not imposing structure that is not.

What We Explicitly Refused

Points for notes, recording minutes, or AI operations. These reward volume. Volume is not the outcome, and rewarding it produces users who create more notes without getting more value — which shows up in our metrics as improvement while the user's actual experience gets slightly worse.

Badges for opening the app or creating a lot of content. Same problem, with a collectible attached.

Daily streaks. This is the one we feel most strongly about. Serious note-taking is episodic. You have a week of meetings and then a week of heads-down work. A streak counter punishes the second week, which is often the more productive one. Worse, streaks generate anxiety about the tool rather than engagement with the work, and a broken streak is a well-documented reason people abandon an app entirely. Building a mechanic whose failure state is churn is a strange thing to do on purpose.

Leaderboards between knowledge workers. Ranking colleagues by note count is a bad idea in ways that should not need explaining.

Loot-box-style variable rewards. Variable reward schedules are effective at driving compulsive use, and they are profoundly mismatched with a tool people are trusting with their work. If someone is deciding whether to put confidential meeting notes in your product, the last impression you want to create is that it is designed to be compulsive.

Surprise, Done Honestly

Variable rewards do address something real: predictable systems get boring. The question is what the surprise consists of.

The wrong answer is confetti, random bonuses, or artificial scarcity. The right answer is intrinsically useful discovery:

• "We found a related decision from last month"
• "This unresolved question has appeared in three meetings"

That is genuinely surprising, genuinely variable, and the reward is information the user actually wanted. It cannot be gamed, because generating more notes does not generate more insights — only relevant ones do.

Competence Feedback Beats Achievement

This is the biggest opportunity and the one furthest from conventional gamification.

After enough usage, show evidence that the user's system is working:

• "4 of 5 meetings ended with a clear next step"
• "7 action items completed"
• "3 open questions resolved"
• "This note contributed to two later decisions"

Compare that to "you created 100 notes."

The first set demonstrates effectiveness. It tells someone their process is producing results, which is information they want and cannot easily get anywhere else. The second is badge theater — it tells them they used a tool a lot, which they already knew and which proves nothing.

For collaboration, acknowledgment can work the same way: a teammate confirms a decision, accepts an action item, resolves an open question, or says thanks. Grounded in the work. No ranking required.

The Metric We Flagged as Wrong

One honest note. Our current definition of "meaningful work" leans on object count and text length — did you place enough things or type enough characters.

That is a volume heuristic, and this entire post is an argument against volume heuristics. It is on the list to replace with something intent-shaped: did the user do something that implies they want to keep this? Naming it. Structuring it. Returning to it.

We shipped the weak version because it was ready. Writing down why it is wrong is how it gets fixed.

The Short Version

Do not ask "what mechanic should we add." Ask "what is the user actually trying to finish, and can they see how close they are."

If the answer to the second question is yes, you may not need a mechanic at all. If it is no, no amount of points will help — you will just be measuring the wrong thing more enthusiastically.

Open a meeting-notes canvas and see what closing the loop looks like.

30 days free. No credit card required.