Revenue model · the proposal · August 2026
One product,
five meters
The business plan laid out nine ways to charge. This is the one I'd actually pick, with the prices, the entitlement boundaries, the numbers over three years and the order to switch each meter on. It is a decision, not a menu.
The catalogue is bought once. The API bills are metered. The market data is rented. The insurance document is referred. The institutions are invoiced.
Five separate meters, each attached to a different kind of cost. No meter ever charges rent on something that costs you nothing to keep running.
All figures net of Apple's 15% (Small Business Program) and deliberately conservative. They're here to be argued with.
01 · the five meters
Every SKU, and the
cost it's attached to
M1
CLCTN Complete — the one-time unlock
€29.99 ONCE
WHAT IT UNLOCKS
Unlimited objects past the free 100. Lending, wishlists, shared shelves, 3D shelf, export, multiple collections. Non-consumable, no account, restores on any device.
ATTACHED COST
None recurring — the data lives in the user's iCloud. That's exactly why it's a purchase and not a subscription.
WHY €29.99, NOT €24.99
Conversion is flat between the two, and €29.99 leaves you a launch price (€19.99), a Black Friday, and a bundle without ever devaluing the app. Price down from strength, never up from weakness.
≈ 80% of year-one revenue. This is the spine — everything else is added on top of a product that already stands alone.
M2
Lookup credits — the metered bill
€4.99 / 500
HOW IT METERS
Barcode scans, photo matches and catalogue enrichment cost one credit. 100 free credits a month for everyone, 300 for unlock owners. Manual entry is always unlimited and free.
ATTACHED COST
€0.02–0.06 per call across Discogs, IGDB, comic and coin sources plus your own image matching. Real money, forever, per use.
WHAT IT'S REALLY FOR
Insurance, not income. It caps the damage a user cataloguing 4,000 comics can do to your margin, and it's the socket rewarded video plugs into later in price-sensitive storefronts.
Build the meter in v1 even if you never sell a pack. Retrofitting a counter onto a shipped free app is the expensive version of this.
M3
Market Data — the only subscription
€6.99/MO · €54.99/YR
WHAT'S RENTED
Live comps, sold-price history, condition-adjusted values, portfolio value over time, bulk listing tools, insurance-grade export. Listing itself stays free and unlimited.
ATTACHED COST
Data licences and an index that must be refreshed whether anyone opens it or not. A monthly obligation, honestly priced monthly.
SHIP IT IN SLICES
Slice one is read-only: comps and price history, no listings, no moderation, no index. That alone is what sellers pay for — and it earns while you build the rest.
The only recurring charge in the whole model, and the only one with a recurring cost behind it. Annual option because cataloguing happens in bursts.
M4
Insurance & valuation referral
€25–80 / POLICY
WHERE IT LIVES
One line at the bottom of the export screen, shown only to someone whose collection has crossed a real value threshold. Never a banner, never a notification, never in the catalogue.
ATTACHED COST
Nothing to build — the export already exists. The cost is two partner conversations and per-country rules on insurance introductions.
HIGHEST €/USER ON THE PAGE
One bound policy is worth two unlocks. It also earns without a marketplace existing, which makes it the cheapest hedge against v2 not happening at all.
Have the conversation in month 3 — before v2 design. What the insurer needs in that file determines the export screen.
M5
Organisation seats — invoiced, not sold
€0 OR €29/MO
TWO PRICES, ONE PRODUCT
Free forever for associations under ~500 objects, with a badge. €29/mo invoiced for anyone with a budget line: schools, clubs, games cafés, subsidised tool and toy libraries, shops.
ATTACHED COST
Three transactional emails, roles, an invoice, and support someone expects an answer to. Off the App Store, so no 15% — and no IAP rules either.
WHAT YOU'RE ACTUALLY BUYING
Distribution. Each org puts a rendered object card in front of 300–400 local, object-obsessed members every month, and it converts on meter 1.
Count members reached, not seats sold. If you ever optimise this for seat revenue you've traded a growth channel for €7k.
Deliberately absent: no transaction take rate (you'd have to force the deal in-app, and you've said you won't), no ads against inventory, no subscription on the catalogue, no promoted listings until the index has real supply. Each of those is a live option that costs more trust than it earns at your scale.
02 · the boundary
Exactly where free ends
This is the single most consequential page of the model, because it's the only one users read. The rule: the free tier is a complete product at a smaller size, never a crippled product at full size. Nothing is teased, blurred, watermarked or locked behind a "Pro" chip.
FREE, FOREVER, NO ACCOUNT
Up to 100 objects — everything works
All 14 item types · full type-specific detail fields · 3D shelf and sleeve rendering · photo-first add flow · 100 lookup credits a month · lending and handshakes · one wishlist · receive and open any shared object link · iCloud sync · CSV import
BEHIND THE €29.99 UNLOCK
Scale, sharing and getting data out
Unlimited objects · multiple named collections · 300 credits a month · shared and collaborative wishlists · public shelf links · full export (CSV, PDF, insurance format) · valuation totals · custom fields · priority in the type-proposal queue
The cap is a ceiling, not a wall: at object 101 the app doesn't stop — it saves the object, shows it, and tells you the next one needs the unlock. Nobody ever loses work they've already typed in. That single behaviour is worth more than any pricing tweak on this page.
Why 100 and not 25: audience C — the inherited crate, one shoebox of coins — must be able to finish their whole collection for free and love you for it. They're the ones who tell audience A about the app. A cap that catches casual users converts worse and costs you the word of mouth.
03 · three years
The model, run forward
BASE CASE · NET OF APPLE'S CUT AND DIRECT COSTS
YEAR 1
YEAR 2
YEAR 3
Cumulative downloads
25,000
70,000
140,000
M1 · unlocks at €29.99 (cumulative payers)
990
2,800
5,600
M1 revenue
€25,200
€28,000
€36,000
M2 · lookup credits
€800
€2,200
€4,500
M3 · market data subscribers (avg active)
—
560
1,450
M3 revenue
—
€31,400
€79,000
M4 · insurance referrals (policies bound)
—
€18,000
€31,000
M5 · invoiced organisations
—
€5,200
€15,600
Direct costs — lookups, data licences, index, mail, Apple dev
−€2,100
−€16,000
−€39,000
Net
≈ €24k
≈ €69k
≈ €128k
Year 3 is the first year this is a salary plus a budget to pay someone else for something. Note the shape: revenue starts as one-time and becomes recurring, which is the right direction — but only because the recurring part has a real cost behind it.
The two cases either side
DOWNSIDE · HALF THE DOWNLOADS, NO INSURANCE DEAL
€11k → €22k → €35k
Still cash-positive from month two, because the only fixed costs are €99/year and your evenings. That's the point of a model with no staff and no burn: the downside is disappointing, never fatal.
UPSIDE · ONE CATEGORY MOMENT OR ONE PRESS HIT
€60k → €210k → €450k
10× downloads with identical pricing. The multiple never comes from the price — it comes from one community adopting you wholesale, or the 3D shelf being the thing people post. Design for that; don't budget for it.
04 · unit economics
What one user is worth,
and what one costs
REVENUE PER DOWNLOAD
€1.04
Year one, all meters. Rises to ≈€1.90 by year three as subscriptions and referrals layer on the same install base.
COST OF A USER WHO NEVER PAYS
€0.35
≈12 lookups in their first week, then silence. Storage and sync are Apple's problem, which is the whole reason this works.
BREAK-EVEN RATIO
1 : 72
One €29.99 unlock pays for 72 users who never spend a cent. You are structurally allowed to be extremely generous with the free tier.
LIFETIME VALUE OF A CATALOGUER
€78
Unlock + a credit pack + 7 months of market data + a 1-in-8 chance of an insurance referral. Deep collections are the only user segment worth optimising for.
Every meter in this model is priced off objects catalogued, not users acquired. So there is exactly one growth metric: median objects per user. Everything else is downstream of it.
05 · sequence
When each meter
switches on
v1 launch — M1 and M2 only
Two SKUs total: the unlock and the credit pack. No accounts, no server-side entitlement, no subscription plumbing. Ship the credit meter even at zero price so the counter exists in the wild from day one.
Month 3 — the two conversations, before any v2 code
Two specialist collectibles insurers: what must an itemised export contain, and what do you pay per introduction. One market-data provider: what does a licence cost per thousand lookups. Those two answers set M3's price and M4's existence — and neither needs a line of code.
Month 8 gate — M3 slice one, read-only
Only if latent supply ≥8% and latent demand ≥30%. Ship comps, price history and the insurance export as the subscription — no listings, no index, no moderation. M4 turns on with it, as one line on the export screen.
Month 12 — M5, invoiced by hand
Roles, a queue, three emails. Free for associations from the first day; invoice the first five budget-holding orgs manually in a spreadsheet. Do not build billing for five customers.
Month 15 — decide what you're building, once, with data
Index with real supply, or seller tools with no market, or an org product. The meters above earn in all three futures, which is the actual reason to choose this model over any single-bet alternative.
06 · the rules
Four lines that don't move
RULE 1
Never rent what costs nothing
No subscription on the catalogue. If the servers vanish tomorrow, a paying user's collection still opens, still syncs, still exports.
RULE 2
Never earn from the deal closing
No take rate, no forced checkout. Two people finishing in a car park must cost you exactly nothing — otherwise you spend years fighting your own users.
RULE 3
Never monetise the inventory itself
No ads against someone's objects, no selling or licensing what they own. The privacy promise is load-bearing — it's half of why anyone types 200 objects into a phone.
RULE 4
Never take work away from someone
Objects already entered stay visible, editable and exportable forever, at any tier. A cap is on adding, never on keeping.
Where this model breaks
If nobody reaches 100 objects
The whole model rests on the cap being reachable. If median objects per user at day 30 comes in under 30, the problem is the add flow, not the price — and no repricing will save it. Fix onboarding and importers first.
If market data licences are expensive
M3 is the only meter with a cost you don't control. If a licence lands above ~€3/subscriber/month, drop live comps and sell the insurance export and portfolio history alone — same price, cost you own.
If insurers won't pay for introductions
M4 is a quarter of year-two revenue and rests on two conversations you haven't had. Have them in month 3 so a no costs you nothing but a rewrite of this page.
07 · founding prices
Price goes up as the
app gets bigger
Yes — do this. A rising price is the single most effective conversion mechanic a paid app has, it's honest here because the app genuinely gains capability over the same period, and it fixes the one real weakness of a one-time unlock: you get paid once, at whatever price you happened to set while you knew least.
One change to your plan: trigger on unlocks sold, not users. A user count can be 10,000 people who never paid — raising the price on the back of that is raising it before you've proved anything.
Unlocks sold is the same number you wanted (a headcount that means you're comfortable), except every unit of it is evidence that the price works. It's also the only number you can see the same day, in App Store Connect, without instrumentation.
The ladder — three steps, then stop
COHORT
TRIGGER TO CLOSE IT
UNLOCK
MARKET DATA
WHAT THEY'RE BUYING INTO
Founding
MONTHS 1–6
First 1,000 unlocks sold, or 6 months, whichever comes first
€19.99
€4.99/mo
for life
A catalogue with no market data, no comps, no org features. They're paying for a promise and taking the risk with you — the discount is the payment for that risk, not a marketing trick.
Early
TO ~MONTH 14
First 4,000 unlocks, or the day market data ships
€24.99
€5.99/mo
for life
A proven catalogue with importers, shared shelves and a year of fixes behind it. Still no market layer. Price rises because the tool got better, and that sentence is the whole justification.
Standard
FOREVER AFTER
No further rises. This is the price.
€29.99
€6.99/mo
The full model from section 01, with comps, valuations, insurance export and the index. The €29.99 was always the real price — the two steps below it were paid rehearsals.
Cost of the discount, on the base case: ~1,000 founding unlocks at €10 off is €8,500 of forgone revenue in year one — about a third of year-one net. That is the price of the mechanic, and it buys you a fast, motivated first cohort plus a genuine reason to talk to press and communities. Worth it; not free.
What each cohort keeps forever — the part that has to be airtight
M1 · THE UNLOCK
Nothing to do — it's already permanent
A non-consumable is bought, not licensed. Raising its price never touches anyone who already owns it, and it restores on every future device at €0. This is the easy meter.
M3 · MARKET DATA
Founding rate, locked, for as long as they stay
Apple lets you raise a subscription price for new subscribers while preserving the old one for existing ones. Use it. A locked €4.99 that never rises is the perk with actual teeth — and the reason a founding buyer stays subscribed for years.
M2 · CREDITS
Never in the ladder
Credits are a cost pass-through. Raising them looks like gouging and earns you nothing worth having. One price, everywhere, forever — and give the founding cohort a permanent double allowance instead.
THE BADGE
Founder № 0341
Numbered, in the app, on their shared shelf links. Costs nothing, can never be bought later, and this audience — people who catalogue and number things — is precisely the audience that cares.
How to actually run it
Say the ceiling out loud from day one
"€19.99 for the first 1,000. It becomes €29.99." On the App Store page, in the unlock sheet, in every community post. A rising price only converts if people knew it was rising — otherwise it's just a price change that annoys whoever missed it.
Announce the rise two weeks out, and then do it
"Founding price ends on the 14th" is the best-converting message a paid app ever sends — expect a spike in the final week that partly pays for the discount. Then actually raise it. One deadline you don't honour and the next one earns nothing.
Flip the price by hand, not by counter
You watch unlocks in App Store Connect and change the price yourself on a chosen date. No server, no live counter, no "347 spots left" widget — a stale or manipulable countdown is worse than none, and an automated flip will fire on the worst possible morning.
Let the storefronts ladder separately
1,000 unlocks is a global count, but the price tiers are per-market (section 02 of the business plan). Raise the family-A storefronts on the ladder and leave price-sensitive markets at their local floor — those users were never the ones making you comfortable.
THE FOUR WAYS THIS GOES WRONG
More than three steps and it stops reading as a founding price and starts reading as a pricing gimmick — pick three, publish them, stop. Rolling it back when sales slow teaches everyone to wait for the next sale; if €29.99 turns out to be wrong, fix it with a proper permanent price change, not by reopening the promo. Raising while the app is still rough — the trigger must be unlocks sold and the app being visibly better than at launch, because "it got better" is the only sentence that makes step two defensible. Raising before the cap converts — if unlock conversion among users who hit 100 objects is under 15%, a higher price is not what's wrong, and the ladder will just make the hole deeper.