Why are we changing our revenue model?
I think the biggest dogma in the SaaS world might be "Don't mess with your pricing model." The caveat "Look, don't touch the prices or the packages," wherever it comes from, always carries a certain anxiety inside it. Sometimes that anxiety hides a "Who's going to define those new packages now?", and sometimes a "Ugh, how are we supposed to migrate the people on Plan X over to Plan Y now?!?" It shouldn't be this way. And for us, it isn't.
Our situation
The top 10 users account for 41% of bills, the top 20 for 58%, and the top 50 for 80%. What's even more striking on a plan basis is this: starter/canceledaverage of users 573 invoices—that is, those who work the hardest and churn the most. 133 fellows/inactive users are currently at an average of 99, while only 6 are actively paying contributions (4 starter active + 2 fellow active + 1 trialing). These numbers directly impact tokenomics design in three ways.
1. The "Flat $9 Starter" model was mispriced in the most valuable segment.
The 8 users churning from Starter processed an average of 573 lifetime invoices. If these users are active for 6 months, they process approximately 95 invoices per month, and if active for 3 months, approximately 190 invoices per month — meaning a "heavy user" profile in terms of volume. This profile gained value from Clint, but the feature band below the $9 single price wasn't enough to sustain it. Tokenomics is targeting this segment. Pro ($12) or Unlimited ($24) We could have kept it in exchange for a larger package with the extra layers — or even if we couldn't, we would have at least gotten 2-3x more ARPU. Conclusion: The Pro tier is mandatory, not optional.
2. The monthly credits for the free tier should comfortably cover the median user.
Lifetime median is 67; assuming an average user's active lifespan is 6 months, that's approximately 11 bills per month. But this is lifetime data — actual usage within the active month is much higher. Safe assumption: median active user processes 15-25 bills per month. Free = 30 credits/week This perfectly matches the median, but it leaves it incomplete. Increasing the free offer to 40 credits per month.— The median user leaves a 30% margin aside while trying the product without feeling "not enough," so the upgrade signal remains organic. This also activates the 22% "under 10 bills" cohort without any effort.
3. There's a "ceiling problem" at the top — Unlimited is really necessary.
The top 6 users alone generate 32% of the total volume with 601+ lifetime invoices (top-1 user 2,252 invoices). These average 100-400 invoices per month. Pro 500 credits/week Even for the 3-4 heaviest users, it wouldn't be enough. Without Unlimited's "fair use 3,000/month soft cap," they would either churn or constantly buy top-ups (which creates friction). Unlimited is for the upper segment: a soothing top cover. It's necessary, otherwise they'll abandon the product.
Clint — Monetization v2
Credit-Based Tokenomics
Instead of a free trial, it's "all-in-one with token limits." All features are publicly available; scalability is managed with credits consumed per processed invoice. Credits are refunded when an incorrectly processed invoice is deleted.
Internal Working Document · April 2026 · Calibrated with 148 users / 18,890 invoice data
Philosophy
Pricing is an invitation, not a gateway. The free plan is not a demo, but a real starting point. The paid plan is not a leap of faith, but a natural continuation. All features are available to all users; upgrading to a higher plan is a volume decision, not a feature decision.
Auto-refund promise: Credit is refunded for every incorrectly processed invoice that is deleted. The AI's error does not affect the user's bill. This is a tangible example of the "Pricing That Asks Less Than It Gives" philosophy and a layer of trust that no one else in the industry offers.
Data summary — what it was calibrated to
Intensity: The top 10 users generate 41% of the total volume, the top 20 58%, and the top 50 80%. Classic Pareto analysis, but a bit harsh at the extremes.
Critical observation: Although the starter/canceled cohort was the most heavily used segment with an average of 573 invoices, it churned out due to a value/cost mismatch with the single-price $9 Starter model. The new model captures this segment in the Pro/Unlimited tiers.
New packaging structure
Top-up: All plans offer top-ups of $5 = 100 credits, $9 = 200 credits, and $19 = 500 credits if used up within the month. Top-ups do not expire.
Resource weighting
The fact that "Manual" is set to 0 embodies the promise of "truly free startup": there can always be a user who can manually manage all bills without paying anything or having any limits.
Refund & Restore logic
Basic principle: Credit is spent on "correctly processed invoices," not on "AI experiments." The cost of an incorrectly processed invoice is not passed on to the user.
Flow
1 · Delete: Instant refund. Credit returns to monthly pool (not rollover). Toast notification: "1 credit refunded."
2 · Trash (30 days):Wait. The invoice can be returned; the credit has already been refunded.
3 · Restore: Recharge. Net zero in the same month; deducted from the new month's balance in a different month.
Action-credit map
Financial impact: Assuming 95% AI accuracy, processing approximately 3,150 invoices per month would result in approximately 158 deletions and a loss of approximately 50–80 effective credits (less than 0.5 credits per user). The refund promise is financially worthless, and the brand impact is significant.
Fellow Grandfather Plan
Risk: 133 out of 148 users (90%) are in the "fellow" legacy. The group that would suffer the most brand damage if they suddenly hit the limit during the tokenomics transition — Clint's early supporters.
Solution: The Fellow plan is permanently maintained. 200 credits are gifted monthly, with no fees required. Only 2 fellows are active in the data → the practical financial burden is close to zero, and the story value is significant.
Contact message:"We started with you. We're not leaving you in the same place. The Fellow program is always free."
Old model vs New model
The truth will set us free and hopefully the destiny will favor the bold.
