The success fee is a tax on your own money

There's a pricing model spreading through small-business software that I think deserves a plain-language teardown: the success fee. You've seen the pitch — “we only make money when you do.” No subscription, no risk, just a cut of whatever the tool recovers, wins, or collects for you. It sounds like alignment. I want to walk through why, for one whole category I've been building in, it's the opposite — and why the cut lands on the worst possible dollars: money that was already yours.

The category: chargebacks

When a customer disputes a card charge, the merchant gets to respond with evidence — the order, the delivery confirmation, the correspondence. Assembling that evidence pack is real work, and a wave of tools now automate it. Here is what the biggest ones charge, from their own published pricing as of July 2026:

• Chargeflow takes 25% of every recovered chargeback, plus $29 per deflected dispute from its alerts product. Recover $10,000 in a quarter and you hand over $2,500 — for evidence packs assembled by software.
• Disputifierruns a success fee too, capped at $250 per dispute — the cap is the confession that an uncapped percentage doesn't track the cost of anything.
• Stripe — the platform itself — now offers Smart Disputes in its own dashboard for 30% of what it wins back. When the payment rail adopts the skim, the model has fully arrived.

Sit with the Stripe number for a second. The dispute data — the transaction, the customer history, the delivery signal — is already inside Stripe. The marginal cost of assembling it into a response rounds to an API call. The 30% isn't pricing the work. It's pricing your reluctance to do paperwork.

The steelman, because it deserves one

Success fees have a real virtue: zero risk at zero volume. If you get two disputes a year, a percentage of something beats a subscription for nothing, and I'll say so plainly — at very low volume, the skim is the right buy. The model also self-installs: there's no budget line to approve, so it spreads through app stores frictionlessly. That's why it won.

But “we only make money when you do” quietly claims something false: that the vendor's incentives now point at your net outcome. They don't. A success-fee vendor is paid on gross recoveries across all customers. Their optimal move is maximum volume through the pipeline, not maximum care per case — your individual win rate is a rounding error in their revenue. The alignment story inverts exactly where you'd want it to hold.

What the reviews say

You don't have to take the structural argument on faith; the receipts are public. On Chargeflow's own Shopify review page this spring: one store wrote it “lost basically 10 out of 10” disputes (OutStorm, May 2026); another that “they charge you 29$ on each dispute! without any action” (Vivaly, June 2026). A Disputifier reviewer on Capterra: most disputes “still go through anyway, making the service almost pointless.” These are one-star outliers on products with plenty of happy customers — but notice what they're angry about. Not bugs. The bill arriving regardless of care. That's the model, experienced from the paying end.

Flat price is a discipline, not a discount

The alternative isn't charity — it's a different discipline. A flat monthly price means the vendor eats the variance. If the software is inefficient, the vendor pays, not you. It forces the thing success fees let you skip: making the cost of the work visible and then charging near it. Software assembling an evidence pack costs cents. A subscription priced in tens of dollars admits that. A 25–30% skim on a four-figure recovery hopes you never do the division.

Full disclosure, since this essay obviously has a horse in the race: Kynth Chargebacksis mine. It connects to your Stripe account, AI assembles the evidence pack, you press submit, and it costs $29 a month whether you recover $500 or $50,000 — you keep 100% of every recovered dispute. I'm not claiming it beats the incumbents at everything; they cover more platforms and more automation. I'm claiming the pricing model is honest, and that in this category honesty is the feature.

The test to run on any success-fee pitch

Three questions, any category:

• Whose money is the fee taken from?If the answer is “money that was already yours, recovered by a form,” the fee is a tax, not a share of created value.
• What does the marginal case cost the vendor?If it's an API call, a percentage of your revenue is not a cost-based price.
• What happens to their revenue if your win rate falls? If volume keeps them whole while you lose, the alignment pitch is backwards.

The success fee won app stores because it feels free on day one. It keeps winning because nobody does the math on day ninety. Do the math.

Isaiah Kim — one-person AI product studio. Prices and quotes verified July 2026 on the vendors' own pages; the engine behind the products is at api.kynth.studio.