Obiex Weekly Pulse

Worth knowing

• Needs action

  No limit order has gone through in at least 8 weeks. Every attempt failed: 8 last week and 48 since late August. 41 of the 48 hit the $1,000 minimum, which people only find out about after they try. The rest did not have enough balance. Showing the minimum before someone types an amount would fix most of this. Visits to the limit order page also fell to 134, the lowest in 8 weeks (from 182).

• Needs action

  Transfers are failing about 5 times more often than in early September. 4.0% failed last week (236 transfers), down a little from 4.8% the week before but well above the 8-week average of 2.5% and the 0.8% of the week of Aug 31. The top reason, people sending with an empty PIN, actually grew (75 to 111) while failed 2FA codes fell (106 to 63). The app seems to let people press send before the PIN is entered, so this is likely a design fix, not an outage.

• Watch

  Fewer people are starting verification, and fewer of those who start are finishing. KYC 1 completions fell to 51 (from 71), BVN to 22 (from 29), KYC 2 to 47 (from 54) and KYC 3 to 12 (from 21). Every step is at its lowest in 8 weeks and has fallen two weeks in a row. It is not only fewer people at the top: 69% of people who started BVN finished it, down from 83% in early October. Signups also hit an 8-week low (82), though our signup count is known to undercount.

• Watch

  Withdrawal failures are up since Oct 12, all of them on web, and we cannot see why. 68 failed last week and 76 the week before, compared with almost none in the six weeks before. About 36 people a week are affected, all on web, with almost no failures on iOS or Android. 62 of the 68 came with no reason attached. This may be a real web problem or web starting to record failures it did not record before. We need engineering to log the reason so we can tell.

• Check

  Web numbers changed on Oct 12, and much of it is probably tracking, not new people. Weekly web users jumped about 70% (1,404 to 2,351) and stayed there. Over the same two weeks, web logins fell from about 175 a week to 11, and web depositors stayed flat (133 last week vs 145 before the change). Real growth would push logins and deposits up, not down. Together with the withdrawal failures above, this points to a web release that changed what we track. Worth confirming with web engineering and growth before anyone counts it as growth. Separately, Android users have dipped three weeks running (1,338 to 1,153), now the lowest in 8 weeks.

• Good news

  Core money movement is steady. About 1,520 people deposited, 1,370 swapped and 1,490 withdrew, all within a few percent of the last two months. Transfer users are up 40% since late August (439 to 613). Swap completion improved to 69% from 67%, recovering from a dip to 60% in early October, though still below the 75% we saw in September.

• FYI

  A few small numbers to keep an eye on, and some blind spots. Account deletions rose for a second week (1, then 3, then 6), the highest in 8 weeks, and only 2 people sent a referral invite (from 8). Both are small numbers, so not a trend yet. Bill payments (data, electricity, cable) are being used but stopped sending tracking data months ago, and card spending has never been tracked, so they do not appear here. That is a gap in our data, not a lack of use.

What customers are telling us

From escalations in the support complaints channel, Oct 19 to 25 compared with Oct 12 to 18. That channel only holds cases support escalates, so these counts are a floor, not total ticket volume.

• USDT withdrawals stuck on processing (11, up from 6). Most were pushed through by the team within hours, one was rejected and one turned out to be a funding issue. One customer got stuck twice in a day. The data agrees on timing: failed USDT withdrawals went from none for six weeks to 25 and then 24 a week from Oct 12, the same week the web withdrawal failures above started. That is still small next to about 3,900 USDT withdrawals a week, but it is the thing customers feel most. Next step for engineering: pull the stuck USDT withdrawals since Oct 12, find what they have in common (network, provider or review queue) and log the failure reason.
• A coin we don't support (1). A customer sent wrapped bitcoin on the Ethereum network, which Obiex doesn't support, and asked for it back. We don't track this in our data. Next step for design: make the supported coins and networks unmissable on the deposit screen, and agree a standard recovery process with support.
• Bank hold on a naira payout (1). The customer's bank held the money and asked us to confirm the transaction. It is still being reviewed. Our data can't see this, and naira withdrawals look normal (about 7,700 completed, no failures recorded). Next step for ops: agree a fast turnaround for bank information requests.
• Push notifications not arriving (1). Support suggested checking notification settings and staying logged in. We can't check this, because notification opens aren't tracked. Next step for engineering: add notification tracking so we can see delivery problems.
• Last week's one-offs. A naira withdrawal stuck on processing that the customer asked to cancel, a customer who said money left their wallet without them (escalated for review), and a customer who couldn't sign in because password reset was switched off on their account. That last one took over a week of follow-ups. Password recovery overall fell (21 to 10 people starting it), so it looks specific to that account rather than a wider problem. Next step for engineering: let support re-enable password reset on an account without waiting on an engineer.

People using Obiex

Verification

Is it working?