The idea that shoppers abandon shopping carts when they run into checkout friction has been said so often that it is approaching cliché status. The truth is much more nuanced and complicated. The level of checkout-friction-resistance changes—for the identical consumer—repeatedly during the merchant interaction.

Let’s consider that abandoned shopping cart consumer. They ran into some site stumbling block, got frustrated and bolted. Let’s further assume that the shopper is somehow reeled back in, most likely with a friendly-phrased text message. Whatever level of resistance/tolerance that shopper had before they abandoned, it’s now ten times more sensitive. A hassle that they would have tolerated before is now cause to run away—and they won’t be coming back.

But let’s tweak that scenario slightly. This time, that same consumer runs into some friction, which is that a price seems too high or the choice of color/style is too limited. That shopper then does some Google searching, visits a bunch of other sites and ultimately comes to the conclusion that your offering’s price/color/style is the best available. That consumer sheepishly comes back to complete the purchase. The situation is now flipped. That consumer’s resistance to checkout friction is now dramatically lower, perhaps ten times lower. Having discovered that your deal is the best they’ll get, they will put up with far more hurdles than they would have before they did that research.

Welcome to the world of payment facilitator BlueSnap. BlueSnap specializes in reducing friction at checkout and CEO Ralph Dangelmaier has become quite the expert at how friction resistance morphs for the same consumer through the shopping process.

As the shopping experience moves more and more often to mobile devices, the friction acceptance/resistance numbers calculation changes yet again. With the smaller screen size, the identical number of hurdles that might be quite acceptable to a shopper when interacting via a desktop could be cart-abandonment-level-frustrating on a smaller mobile device.

And different causes for abandonment also impact that acceptance/resistance number. A very common reason for abandonment, Dangelmaier said, are coupons don’t work. In those cases, the shopper leaves solely to find a better coupon, ideally one that hasn’t expired. In other words, those consumers haven’t really abandoned. They have simply left temporarily while seeking a way to get a better price. Those shoppers will come back on their own and buy. Of course, to the merchant who owns that site, it sure looks like that shopper might have abandoned for good.

But when those shoppers do come back on their own, they come back with a far higher checkout friction tolerance. “That shopper is way more tolerant. They’ll painfully go through the check process and then complain,” Dangelmaier said.

The problem, of course, is that there’s no way to know ahead of time the risk tolerance. And even if a merchant could know the friction tolerance of every one of its shoppers, that merchant would still need to design a checkout that catered to the shopper with the lowest tolerance. Hence, they need to build checkout that way anyway.

And as more purchases go mobile, this problem is going to just get worse. “Tolerance on the phone is very low,” Dangelmaier said.

Mobile purchases can be handled in two ways: Via a mobile browser; or within a merchant app. “If you have app, your mobile checkout can deliver very little friction. If you don’t have an app, that’s where you get into a lot of friction,” he said. To be nitpicky, it’s not that the mobile app delivers less friction per se. It’s simply that it shifts that friction to the initial site setup, which is generally completed when the consumer has the time to fill out a lot of forms, ostensibly having to do it only once.

In short, the app shifts the friction to an activity where it will be tolerated much better—app setup—and away from one where friction is hated more, the checkout for a specific purchase. Amazon’s app is a good example of an app that, once setup, delivers a lightning-fast and effortless checkout experience.

Much of the problem with mobile web purchases is at the software level. “It’s a technical problem. You have to deal with different devices, different versions. You have to set up your API to recognize the IP of the country and the device,” he said. The mobile app can be downloaded based on those initials in the initial setup. Again, it’s shifting the problem to an activity where it’s better tolerated.

Although it might seem that mobile apps are the answer, trends suggest that such apps may eventually go away. After all, few consumers want to have to download an app for every merchant they might use. That is why Dangelmaier argues that it’s going to be mobile wallets that will eventually solve the checkout friction issue. “I believe that (a mobile wallet) is the solution for folks. In the meantime, mobilely optimize your site for checkout or create an app and drive everyone to the app,” he said.

Dangelmaier is a firm believer that his company’s move to a payment facilitator model has made these issues so much easier for his merchant customers to resolve. Being a PF “lets the merchant be the merchant,” he said.

The CEO cites a recent personal example, where he was trying to rent a place for a vacation. When he was going to pay, the site redirected him to PayPal and it displayed a new name that Dangelmaier didn’t recognize. “Who is this company? What is going on?” he asked.

When the CEO later spoke with an executive with the rental operation, the executive said that “nobody ever checks out on the first time.” When Dangelmaier asked why he tolerates that from his site, the executive said that he didn’t know that was any alternative.