The myth: turn it on and fraud chargebacks are the bank's problem
The standard pitch for 3D Secure goes like this: authenticate the customer, and fraud liability moves to the bank that issued the card. The pitch is half right, and the half it leaves out is where merchants get surprised. Stripe's own documentation, the most careful public source on the mechanics, says successful authentication does not guarantee liability shift. A merchant can expect the shift for successfully authenticated payments that later receive eligible fraud disputes, but it is not guaranteed for any specific payment or dispute. That single qualifier changes 3D Secure from insurance into a tool with conditions, and the conditions, plus a Canadian price tag the explainers never quote, are the subject of this piece.
What an attempt actually costs in Canada
Start with the number Canadian merchants are rarely given. Stripe's Canadian authentication page publishes CA$0.04 per 3D Secure attempt for accounts on custom pricing, checked November 3, 2026. On standard pricing, Stripe's Canadian pricing page lists its card rates with no separate per attempt 3D Secure line, so a standard pricing merchant should confirm in their dashboard or agreement whether attempts are bundled rather than assume either answer. Other processors price authentication inside gateway or risk bundles, or quote it. What no honest source supports is a single Canadian market price for an attempt, because the published record contains exactly one firm figure, and it belongs to one pricing tier of one processor.
Dated price check, November 3, 2026: Stripe Canada, CA$0.04 per 3D Secure attempt, custom pricing accounts (Stripe authentication page). Stripe Canada dispute received fee, CA$15.00 per dispute (Stripe Canada pricing page). Moneris chargeback fee, $25.00 (Moneris pricing page, incidental fees).
Four cents sounds trivial until it is multiplied by a checkout. Authenticate every order in a store taking 1,000 orders a month and the custom pricing bill is $40.00 a month, $480.00 a year, paid on good orders and bad alike, including attempts where the card turns out not to support the protocol at all. Authenticate only the riskiest fifth of those orders and the bill is $8.00 a month. The price per attempt is small. The policy decision, all orders or selected orders, is a forty to eight dollar decision, and it should be made with the dispute side of the ledger open.
What the liability shift covers, and what it does not
The shift, when it applies, covers fraud disputes on successfully authenticated payments. Stripe's documentation draws the boundaries clearly. If a card does not support 3D Secure, or an error occurs during authentication, the payment proceeds normally and liability does not generally shift, because no successful authentication took place. The shift can also fail or be downgraded in rare cases, for example when a merchant with an excessive fraud level is enrolled in a fraud monitoring programme. Some industries are outside it entirely: Visa does not support liability shift for wire transfer and money order businesses, for non financial institutions offering foreign or non fiat currency, or for stored value card purchases and loads.
Just as important is the dispute type boundary. The shift answers a fraud claim, the cardholder saying the purchase was not theirs. It does nothing for the other large family of disputes: goods that never arrived, services not as described, cancelled subscriptions billed again. Those follow the normal dispute process with the normal fees, authenticated or not. A store whose disputes are mostly delivery and description problems will buy very little protection with authentication attempts, and should spend the money on the checkout and shipping evidence that actually answers those disputes.
One mechanic works in the merchant's favour at dispute time. When a merchant counters a dispute on an authenticated payment, Stripe automatically adds the requested electronic commerce indicator and the 3D Secure authentication outcome to the evidence details. The authentication record is, in other words, filed for you, provided the authentication actually succeeded in the first place.
The break even math: attempts against dispute fees
Price one avoided fraud dispute on Stripe Canada. The merchant keeps the sale amount that would otherwise be reversed, and avoids the CA$15.00 dispute received fee. Fighting the dispute manually would add a CA$15.00 counter fee, returned if the dispute is won and kept if it is lost, so an unauthenticated fraud dispute the merchant contests costs the sale plus up to CA$30.00 in fees. On a $120 order, that is $120.00 to $150.00 of exposure. At CA$0.04 an attempt, $150.00 buys 3,750 authentication attempts. A merchant does not need authentication to work often for the arithmetic to favour it on risky orders. One shifted fraud dispute in every few thousand selected attempts pays for the programme.
| Policy, at 1,000 orders a month | Attempts | Monthly attempt cost at CA$0.04 | Fraud disputes that must shift to break even, at $135 each |
|---|---|---|---|
| Authenticate every order | 1,000 | $40.00 | About one every three months |
| Authenticate the riskiest 20% | 200 | $8.00 | About one every seventeen months |
| Authenticate nothing | 0 | $0.00 | None, and every fraud dispute lands in full |
The table uses $135 as the value of a shifted dispute, a $120 order plus the CA$15.00 received fee, and it is a worked example with those assumptions stated, not a promise. Merchants on other processors should substitute their own dispute fee: Moneris, for example, publishes a $25.00 chargeback fee, which raises the value of each shifted dispute and shortens every break even in the table. The conclusion survives either fee schedule. Selective authentication is cheap insurance. Blanket authentication is a recurring bill that needs a fraud rate to justify it.
A selective authentication procedure for a Canadian store
For a store that wants the protection without the blanket bill, the procedure is short.
- Confirm how your own account prices attempts. Look for a per attempt line in your pricing agreement or dashboard before turning anything on, because the published Canadian figure applies to one tier and guessing wrong in either direction distorts the math above.
- Set request rules instead of a global switch. Stripe notes that merchants on its Radar Plus tier can customise rules controlling when to request 3D Secure and how to handle each authentication outcome. Sensible triggers are high order values, first orders from a new customer, and mismatches between billing and shipping details. The triggers are your policy choice; the rule mechanism is the documented part.
- Record the authentication outcome on every selected order, so the result is on file if a dispute arrives months later.
- Answer dispute inquiries on authenticated payments rather than ignoring them, and counter fraud disputes so the authentication evidence is submitted with your case.
What success looks like: a test order that meets your triggers shows a completed authentication and an authenticated result in your processor records, your monthly attempt count is a small fraction of your order count, and the first fraud dispute on an authenticated order is answered with the authentication outcome already attached.
The honest summary is narrower than the pitch and more useful. 3D Secure in Canada costs four cents an attempt on the one published custom pricing figure, protects only fraud disputes, protects them only when authentication genuinely succeeded, and pays for itself at remarkably low success rates when it is aimed at the orders that need it.
— Dex Carter, Tech Writer
Sources, accessed November 3, 2026: Stripe, Optimised Authentication (Canada), CA$0.04 per 3D Secure attempt for accounts with custom pricing; Stripe Canada pricing (dispute received fee CA$15.00, dispute countered fee CA$15.00, returned on won disputes); Stripe documentation, Authenticate with 3D Secure (liability shift expectations, exceptions, and evidence handling); Moneris pricing (chargeback fee $25.00). Figures are processor published and change over time. Confirm your own agreement before relying on a figure. General information only, not financial advice.