We Measured Bitcoin Confirmation Times Across Fee Levels

Bitcoin confirmation times are quoted as ten minutes and experienced as anything from two to ninety. We measured it across fee levels over two weeks. Our reference for the published-terms case was a provider that lets a business accept Bitcoin payments.

The method

Payments at three fee levels relative to the prevailing rate: at the recommended rate, at half, and at a quarter. Recorded time to first confirmation.

The results at the recommended rate

Median eight minutes. Ninetieth percentile twenty-six minutes. Longest observed forty-one minutes.

Consistent with expectations and highly variable, which is the point. A customer told to expect ten minutes and waiting twenty-six has a poor experience even though nothing went wrong.

At half the recommended rate

Median twenty-two minutes. Ninetieth percentile ninety-four minutes. Two payments took over four hours.

At a quarter

Median over two hours. Three payments were still unconfirmed after twelve hours during a congested period.

What this means for checkout

The payment window has to be longer than for fast networks. Twenty to thirty minutes is reasonable; ten is not.

And the fee matters enormously. A customer whose wallet set a low fee has paid, and your system shows nothing, for hours. On the merchant side the equivalent is crypto acquiring for businesses, which we tested separately.

This is not fraud and it produces the same support conversation. Providers that flag an underpaid fee rather than leaving the payment silently pending handle this much better, and only one of the four we use does it.

The confirmation policy question

Accepting at zero confirmations for small amounts removes the wait entirely and carries a small replacement risk.

We tested whether our providers absorb that risk. Two do below a stated threshold, using replacement signalling detection and fee assessment. Two do not and pass the risk to the merchant.

Worth knowing which, because it determines whether a tiered policy costs you anything.

What we changed after measuring

Payment window extended to twenty-five minutes. Confirmation tiers by amount rather than a flat policy. And a line on the payment page explaining what the wait is for.

That last change had the largest measurable effect on completion, more than any technical adjustment. Telling someone what is happening while they wait is worth more than shortening the wait. Reproduce this yourself. the published coverage list quotes without an account, which makes the comparison a five minute exercise.