Card-to-crypto checkout — is anyone still getting approvals worth the 4%?

Running the card-input flow where the customer pays Visa or Mastercard and I settle in USDT. All in it is roughly 3.5% on the card side plus 0.5% gateway, so call it 4%, which I was fine with when it was converting.

2,614 attempts since February:

Month	Attempts	Approved	Rate
Feb	288	205	71.2%
Mar	401	279	69.6%
Apr	512	337	65.8%
May	549	341	62.1%
Jun	517	308	59.6%
Jul (to 22nd)	347	201	57.9%

Same offer, same lander, same traffic mix, same descriptor. Thirteen points of approval gone in five months and I cannot find the variable.

If you are on this flow, post your rate and your month. I do not want theories yet, I want to know whether the curve is mine or everyone's.

Mine is flat at around 64% since March, but the aggregate hides the thing that matters for me: it splits almost entirely by issuer country. SG and MY cards clear in the high seventies, ID and PH sit under forty and have all year.

If your traffic mix shifted geo at all between February and June, that alone would produce your curve without anything changing on the gateway side.

58% is still 58 free percent though!! Card processing at all in this space is a miracle, I'd take it

no. it's 42% of traffic he already paid for and will never see again

Split the declines before you go looking for a cause. Soft versus hard is two different problems and the aggregate rate can't tell them apart.

Also: same descriptor since February, or same descriptor since you last checked what the descriptor renders as on a statement?

Fair challenge, so I pulled the split for June and July.

Reason class	Jun	Jul
Soft (insufficient funds, velocity)	94	61
Hard (do not honour, restricted)	115	85

Hard declines are the majority and growing as a share, which rules out most of the things I can fix. Descriptor is unchanged and I checked it against a real statement in May, not against my own config.

Geo mix moved four points toward one market between February and June, which is not nothing but it is not thirteen points either. Best I have is that it's issuer-side and getting worse across the board rather than anything specific to me.

Is not issuer. Is category they file you under on their side.

I have same problem in spring, i change the business description on the account to something more boring, same offer same page nothing else different, and approval go up maybe fifteen point in two week. Issuer see different code, issuer behave different. Anyway try it before you accept that number.

Over how many transactions?

Fifteen points on a two-week window is also what you get from a normal weekly cycle if the before-period ended on a weekend. I am not dismissing it, I am asking for the denominator.

Enough of them. I not keep spreadsheet like you keep spreadsheet.

You can test it in one afternoon or you can wait for someone to bring you a table. Up to you.

A structural point that I suspect explains why the lever you are looking for does not exist on your side of this.

On that flow you are not the merchant of record. The gateway is. That is precisely the feature you are buying — it is why nobody asked you to complete merchant acquiring — but it also means the approval decision, the risk scoring and the relationship with the acquirer all sit with them and none of it is visible to you or adjustable by you.

The corollary is the part I would plan for: the chargebacks land on them too, which is why gateways in this arrangement suspend whole categories at short notice rather than negotiating with individual merchants. A declining approval rate and a category suspension are often the same event at different stages.