In the past few weeks, I’ve mentioned some “missing money” a few times – QuickBooks telling me I earned more this year than I could account for in my bank accounts. I think I finally found that today while going over things again and figuring out how July went (worse than May and June, better than April). Chargebacks weren’t being accounted for in one of the columns I thought they were included in on one of my merchant account statements. Totaling over $10,000 so far this year, that seems to account for the discrepancy. Still my best year yet, but it could’ve been better by that much.

I hate harping on this subject, but it’s still my primary problem. A chargeback isn’t just a lost sale but often comes far enough after the order is processed that I lose the cost of providing the services purchased as well. I already have MinFraud scrubbing every credit card transaction; those with a high score are thrown out before they hit the payment gateway, and those with medium scores get a look over before I accept them. I still miss that 7% of sales made with stolen cards somehow.

The only solution I’ve had posed when I ask people in the banking industry, and people running similar businesses, is to call every customer and verify the order before accepting it. That’s just not going to work for me.

• I won’t call them myself – if that level of involvement in the business is required, then suddenly the perceived margins drop way below acceptable levels for me, and I’d rather not be in that business at all. I value my time, specifically my free time, too much for that.
• There’s not enough margin to outsource calling customers on a daily basis to verify orders and still pay myself comfortably. The calls would be per order, while the profit per order isn’t substantial enough. I receive 25-35 relatively small orders on the average day, and it’s in aggregate that they become substantial.
• Automated telephone verification – the kind MaxMind and DialVerify offer – is not a solution I can live with. I’ve been on the brink of implementing MaxMind’s phone verification several times, and already coded up a gateway for adding it to my shopping carts. But I don’t like the user experience at all – trying to force out a “billing” phone number from each customer, making them sit at a phone to place an order so they can verify it, and then the expectation that if I’m requiring a working phone number, I’m also going to answer pre-sale questions over the phone. I’m not; my goal is to be hands-off as possible. My phone number is on billing statements, but I don’t want to be selling over the phone to potential customers, or answering followups from those that just ordered and weren’t sure exactly what they were ordering, and I think I’ll be setting that expectation.

I’ve yet to have anyone I reached out to provide other advice. But I’ve come up with some of my own. Stop accepting credit cards. Only accept PayPal, like I did back when I first started targetedvisitors.info in 2004. These days, PayPal is more appealing than ever — well over 100 million accounts, and the account optional checkout process was made extremely streamlined and sleek this year. There are four main benefits:

1. Except for customers in countries PayPal doesn’t support, there’s no real loss in payment options. With the account optional checkout being prominent and clear now, it’s simple to pay with a credit card. The only difference is the payment takes place on PayPal’s website instead of my own.
2. PayPal takes the burden of initial fraud scrubbing, and they do it well. My reversal rate with PayPal is probably under 1%, just where it should be.
3. If a chargeback does occur, it’s against PayPal, not me. My ability to accept credit cards on low risk sites like W3Counter (no chargebacks in its history) is kept safe, and PayPal helps fight chargebacks on my behalf since they have a stake in it too. I’ve had them win before; they can do it without a tangible item.
4. Fraud perpetrated through PayPal is usually noticed by either PayPal or the account owner much faster than the 1-2 months it takes for credit card charges to be noticed on a consumer’s statement. Usually it happens between 1-3 days, which means if the order slipped past us all and I already started providing service, I’ve only invested a small amount into that customer so far. I can recover most of my costs by canceling the order while it’s mostly unfilled.

I’m going to think this over some more, but I’m also going to start working on the code now. I don’t want to alienate my best customers, who repeatedly order on company credit cards, so I can’t drop it altogether. The goal would be to seamlessly provide the multiple payment method checkout when an existing customer is placing an order, and streamlined PayPal-only checkout for new customers. I don’t need to change any part of the website or leave any evidence there used to be choice – simply send the customer to PayPal at the payment stage as I do now when they do choose. There are a couple reasons I think this could actually improve the bottom line:

• The losses due to chargebacks are high enough that it would take a large increase in cart abandonment due to lack of payment options or being sent off-site to result in lower net profit.
• I expect the complexity, inconvenience, and surprise at automated phone verification of credit card payments may cause higher cart abandonment than offering only PayPal would.
• A streamlined checkout process with less steps, where billing information isn’t collected, and a payment option choice screen isn’t needed, means customers get from the purchase decision to payment faster, resulting in a lower abandonment rate.
• I’ll spend less time developing new fraud screening techniques, less time reviewing orders, and less time processing chargebacks due to fraud.

By capturing the customer’s name and address from the PayPal IPN postback after payment, I can maintain the integrity of the current database while doing this. That means if I were to test it, and saw an unacceptable drop in sales due to the change, I could revert to the old checkout process without any holes in the necessary data. I would still have my merchant account, so I don’t lose the benefit of multiple processors in case I were to have a problem with one. I think this may be worth trying.