Habits and flywheels
I wrote this to work out what comes after onboarding, when use has to turn into a habit and a habit has to turn into a flywheel.
What happened next: 1.0 shipped the fast capture paths this doc argued for: photos, book scanning and adding several items at once.
Companion to the progressive onboarding plan. Onboarding gets a user to the right feature at the right time; this is what happens next: turning use into a habit, and habit into a flywheel that compounds.
0 · Three concepts, cleanly separated (then wired together)
These get blurred constantly. I want them distinct, because they need different design moves.
| Progressive onboarding | Habit | Flywheel | |
|---|---|---|---|
| Question it answers | Does the user discover the right capability at the right moment? | Does the user come back and repeat the core action without being told? | Does each turn make the next turn easier and more valuable? |
| Unit | A reveal (one feature, one moment) | A loop (cue → action → reward → investment) | A compounding system (value feeds motivation feeds value) |
| Dimension | Breadth: activation across features | Depth: repetition of one behavior | Momentum: self-reinforcement over time |
| Time horizon | First hours/days | Weeks | Months+ |
| Failure mode | Feature blindness (it's on screen but never "registers") | One-and-done; never returns | Effort in, nothing compounds; it feels like a chore forever |
| My lever | Timing & context (the reveal table) | Friction & reward on the core action | Designing loops where output re-enters as input |
How they interrelate: the sequence and the cross-feeds. Onboarding is the ignition, habit is the engine, the flywheel is the momentum the engine builds. In order: onboarding gets the user to the first few turns of a loop; habit keeps them turning it; the flywheel is the compounding payoff that makes the habit worth having. And they cross-feed each other:
- A flywheel win is a celebration moment, and a celebration moment is exactly where onboarding reveals the next capability (see the onboarding plan). So the flywheel generates onboarding opportunities.
- Onboarding a shortcut reduces habit friction: teaching "Add Several" at the right moment lowers the cost of the core action, which speeds the flywheel.
- A new habit unlocks a new flywheel: once scanning books is automatic, the bookshelf-completeness loop starts spinning.
So they're three views of the same growth system rather than competing frameworks. Onboarding = what to show. Habit = how to make one thing automatic. Flywheel = how the whole thing compounds.
1 · How to build a good habit in HomeBase
The durable model is the loop: Cue → Action → Reward → Investment (the last step is what separates a habit that compounds from a treadmill). Applied to a home-inventory app, which is a deliberate, occasional behavior, not a slot machine, and I want to keep it that way:
Cue: attach to moments that already exist. People don't naturally think "open my inventory app." They do think about their stuff at real moments: a package arrives, they buy something, they pack for a trip, they declutter a drawer, they can't find a thing. The habit design is to ride those existing cues rather than manufacture new ones: - A share-extension / "Add to TuckBuddy" from Photos, Safari, or a retailer's order page: capture at the moment of acquisition. - A home-screen widget / quick-add so the app is one tap from the moment, not five. - Gentle, soft-proxied reminders tied to real events (a move, a season), never nagging, consistent with the onboarding governance.
Action: make the core behavior nearly frictionless. The habit is "add an item," and every bit of friction there is tax on the whole system. This is why capture has three fast paths (photo, scan, describe) and why the repetition→shortcut reveal exists: the moment someone grinds manual entry, I want them holding a faster tool. The cheaper the action, the more the loop turns.
Reward: immediate, felt, and sometimes delightful. Two kinds: the intrinsic reward (the item is saved, the collection grew, I can find it now) and the designed reward (the satisfying save, the milestone celebration). The rule from onboarding holds: celebrate, but never block the next add, and ration the teaching so the reward stays a reward.
Investment: the step most apps skip. Every item entered makes the next session more valuable: search returns more, the catalog is more complete, the thing is more mine. That stored value is both the reward for returning and the switching cost that keeps me here. Design so that using the app deposits value that pays out on the next visit: that's what turns a habit into a flywheel.
One caveat: this is a mass-entry app that should reward bulk, repetitive work, so the habit I want is "capture as I go / clear a backlog in a sitting," and the anti-pattern is interrupting that flow to teach or celebrate. The habit and the onboarding governance are the same discipline seen twice.
2 · Flywheels I see in HomeBase
A flywheel is a loop where the output re-enters as input and each turn lowers the effort or raises the payoff of the next. Here are the ones I can already see, strongest first.
① The Completeness flywheel (the core loop). More items entered → search & "where is it / do I own it" work → more real wins → more trust that it's worth entering → more items. This is the engine; everything else supports it. The threshold to protect: it has to cross the point where finding-in-app beats finding-by-memory. Below that it's a chore; above it, it's a reflex.
② The Capture-friction flywheel (onboarding feeds this). Learn a faster capture path (scan / describe / batch) → entry gets cheaper → enter more → more value → worth learning the next shortcut. This is the direct bridge from progressive onboarding into the flywheel: every well-timed shortcut reveal speeds the core loop.
③ The Organization flywheel. As items accumulate, organizing (zones, tags, lists) pays off more → the payoff motivates more organizing and more careful entry → a cleaner catalog makes the next thing easier to place. The "Unsorted" nudge and data-management reveals are the accelerants here.
④ The Books micro-flywheel (unusually sticky). Scanning a spine is fast and a little fun → the bookshelf fills → the shelf is delightful to browse → that delight motivates scanning more. It's a self-contained loop with a strong intrinsic reward (a beautiful shelf), which makes it a great on-ramp to the whole habit. Worth leaning into.
⑤ The Data-gravity / lock-in flywheel. The bigger the catalog, the more valuable it is and the more it hurts to abandon → backup and (later) sync deepen the investment → the moat is your own data. This is more a retention loop than a growth one, but it's real: switching cost = your entire inventory.
⑥ The Household flywheel (future, multi-user). A shared inventory → both partners contribute → more complete → more useful to both → both rely on it daily → they pull in the rest of the household (roommates, family). A within-household network effect: each additional contributor raises the value for everyone already in. This is the one with the most upside and the most design risk (ownership, trust): it's why the multi-user work is gated and careful.
⑦ The AI-intelligence flywheel (future, premium). More entries → the AI has more context → better suggestions, duplicate detection, implied-zone offers, richer auto-fill → capture gets smarter and cheaper → more entries. This is where the premium tier and the flywheel align: the paid feature gets better the more the free habit runs.
3 · Where these show up in the build
Good news: the same infrastructure serves all three. The signals layer + coach engine from the onboarding plan is exactly what a habit/flywheel program needs too: the difference is only what the coach optimizes for at each moment:
- Onboarding asks the coach: is there a capability to reveal here?
- Habits ask: is the core action cheap enough, cued at the right moment, and rewarded? → share-extension, widget/quick-add, frictionless capture, the milestone celebrations, and soft-proxied event reminders.
- Flywheels ask: does this session deposit value that pays out next time, and is a compounding win worth celebrating? → surfacing the completeness payoff ("you can now find X in seconds"), the bookshelf delight, backup/sync as investment, and (later) household and AI loops.
Backlog shape: this rides the same growth epic I proposed for onboarding, plus a few habit-specific build items worth logging when we get there: a share-extension / quick-add widget (cue + friction, probably the single highest-leverage habit lever), the milestone celebration system (reward), and the completeness-payoff surfacing (make the flywheel's compounding legible so the user feels it). The household (⑥) and AI (⑦) flywheels map to the already-gated multi-user and AI-assistant epics.
4 · The one-paragraph mental model to keep
Onboarding lights the fire, habit keeps it burning, the flywheel is the heat it throws off, and the heat lights new fires. Reveal the right tool at the right moment (breadth), make the core action automatic and rewarding (depth), and design every loop so its output re-enters as input (momentum). Keep them distinct when I'm designing a specific moment, but remember they're one system: a flywheel win is an onboarding moment, an onboarding shortcut is a habit accelerant, and a new habit opens a new flywheel.