Per alert or per deflection, which alert pricing model actually costs less
A merchant sent me two quotes last month and asked which was cheaper. One was $15 an alert. The other was $29. He had already decided.
He had decided wrong, or at least he had decided without the number that settles it. The two vendors were not selling the same event.
Two different meters wearing the same word
Most of this category bills you when an alert arrives. Chargeback.io at $15 for Visa RDR and CDRN and $29 for Ethoca, Disputely at $16, $25 and $25, and our own card at $15.00 and $24.00, all charge on receipt. The signal reaches you, the meter ticks, and what you do next is your business.
Chargeflow bills differently: $29 per deflected chargeback, and in its own words you "only pay for alerts that effectively deflect chargebacks." Chargeblast's Shopify listing runs a third variant, a monthly plan starting at $19 plus $14 per chargeback stopped. All checked August 17, 2026.
Those are not competing prices. They are competing definitions of what you are buying.
The reason the difference is real: not every alert becomes a deflection. Some arrive on orders you already refunded. Some lose the race to the chargeback filing. Some are duplicates of an alert you already actioned, since Ethoca and CDRN both see Visa disputes. And on some you will look at a $28 order and decide the refund is not worth it.
Every one of those is a billable event under pay-per-alert and a free one under pay-per-deflection.
The arithmetic that settles it
Call your deflection rate d, the share of alerts you receive that actually stop a chargeback.
Under pay-per-alert, your cost per dispute stopped is the fee divided by d. At $24.00 an alert and a 70% deflection rate, each stopped dispute cost you $34.29, not $24.00.
Under pay-per-deflection, your cost per dispute stopped is the fee. Flat. $29 whatever d does.
Set them equal and you get the crossover.
| Your alert fee | Pay-per-alert wins above this deflection rate |
|---|---|
| $24.00 (Ethoca) | 83% |
| $15.00 (Visa RDR, CDRN) | 52% |
| $29.00 (Ethoca elsewhere) | 100%, so effectively never |
Against a $29 per-deflection fee, a $24.00 Ethoca alert only beats it if more than 83 alerts in every 100 actually stop a chargeback. A $15 Visa alert clears the bar at 52%, which is a far easier standard.
And a $29 alert billed on receipt cannot beat a $29 fee billed on success. It would need a deflection rate above 100%. That is worth sitting with if you are looking at a rate card where Ethoca is priced at $29 per alert received.
The number nobody measures
Here is the problem with everything above. Almost no merchant knows their d.
Vendors quote effectiveness ranges, and those ranges describe the network's reach rather than your operation. Your actual deflection rate is set by things inside your control: how fast someone acts on the alert, whether refunds are automated or wait for a human, whether you refund on low-value orders at all, and how many duplicates your vendor waives.
A store with rule-based auto-refunds running against a clean order feed will land far above a store where alerts arrive as email to a shared inbox and get picked up the next morning. Ethoca's guidance targets a refund inside roughly 24 hours, and missing that window converts a billed alert into a chargeback you also pay for.
That is the real trade, and it is not primarily about price. Per-deflection billing is insurance against your own execution. Per-alert billing is a bet that your execution is good, and it pays out when it is.
This is where I should say what we do. Deflecting disputes before they are filed bills on alerts received, at the rates on our published card, and it supports rule-based automatic refunds precisely because the model only works in your favour when the response is fast. We waive duplicate fees, which lifts d by removing the most common category of alert that can never deflect anything.
I am not going to claim that makes per-alert billing right for everyone. If your alerts land in a shared inbox and get handled when someone gets to them, the per-deflection vendor is straightforwardly the better buy, and you should take it.
The honest part
The crossover math above assumes the two vendors deflect at the same rate on the same disputes, and they will not exactly. Network enrolment differs, and a vendor connected to a program you are not on will show a different denominator. Treat 83% and 52% as the shape of the decision rather than a precise line.
There is also a definition problem hiding inside "deflected." A vendor billing on success decides what counts as success. Ask whether a deflection is recorded when the refund is issued or when the chargeback provably never arrived, because those produce different invoices. Ask the same of us.
And the largest caveat is the one I keep returning to: the fee is the smallest line in an alert program. At a $120 average order value, the difference between $24.00 and $29 per stopped dispute is about 3% of what that dispute actually cost you once the refund is counted. Getting the billing model right is worth doing. It is not worth doing before you have worked out whether refunding at your order value makes sense at all.
The measurement to run first: take last quarter's alerts and count how many ended in a refund issued inside the response window, divided by total alerts billed. That fraction is your d. If it is above 83%, buy on the per-alert card and keep optimising speed. If it is below 52%, buy the per-deflection model and stop paying for alerts you were never going to action in time.