We’ve been talking about micropayments for a few decades. Why haven’t they taken off?

A micropayment is small digital payment. There is no common definition for exactly how small, but a range of anywhere from a fraction of a cent to 50 cents is often used. Micropayments have been touted as a way for media companies to make the transition from print to digital. Instead of having to pay a large monthly subscription for a newspaper or being bombarded with ads, readers would pay a dime to consume each article. Artists and other content creators would be able to monetize their creations by charging a few cents to view each blog post or video.

There are two theories for why we don’t have micropayments yet. The first theory is that payments networks are expensive to run, thus pricing out small purchases. Take a writer who wants to make some money by selling a short story. She sets the price at 25 cents, with the expectations of selling 1,000 views for a net return of $250. Sounds great, right?

But let’s look more closely at the transaction costs involved. The Visa network charges 0.05 percent plus $0.21 on each “small ticket” debit card transaction. That works out to a 21 cent fee per eyeball. Given that the writer has set the price for her story at 25 cents per view, almost all of her revenues are consumed by card fees.

The ability to make cheap micropayments has improved over the years. In the blockchain space, tools like Changetip, XRP Tip Bot, and Tippr can all expedite small amounts at a fraction of the cost of the card networks. Non-blockchain micropayments providers manage to reduce card network fees by aggregating many small payments into one lump-sum payout.

This brings us to the second theory for why micropayments haven’t taken off. Nick Szabo, a legal theorist/monetary theorist, has postulated that there is a “mental accounting barrier” to micropayments. For larger amounts, say $20, the seconds spent puzzling out whether to make a $20 purchase is time well spent. Making the wrong decision means wasting the $20. But no one wants to waste brain cycles estimating whether 2.3 cents is too much to pay for short story. Micropayments are a nuisance—and so we don’t make them.

It’s not entirely true that micropayments don’t exist. For small subset of the population, they do. A micropayment is around 10 cents or 0.06 percent of the average U.S. household’s daily income. But for a billionaire, 0.06 percent of daily income equates to $60. Think Bill Gates buying groceries.

Gates consumes plenty of goods that can only be purchased with a billionaire’s version of a micropayment—his box of cheerios, his toothpaste, and his sweater. But he has a limited amount of brain cycles to devote to shopping. The more he worries about the prices of small ticket items, the less time he has to devote to big ticket items in the rage of the thousands, millions, or billions of dollars.

To get around the mental hassle of living in a world of micropayments, Gates has several options. One is to shop for subscriptions. This allows him to make a once-a-year comparison between different subscription plans rather than frequent comparisons between the individual components of a plan.

But for many goods and services, these plans don’t exist. Grocery stores don’t offer grocery subscriptions, nor do clothes stores. Gates has another option: hire assistants to do his micropayments for him. A good assistant gets to know Gates’ preferences for food and other necessities and tries to satisfy them within the budget that Gates sets out. Trust is key. After all, the assistant will get to know Gate’s most intimate information. He doesn’t want this to be leaked.

Could computers one day replicate the task performed by Bill Gates’s human shopping assistants, but for the rest of us? These agents could learn about us over time, and then take on the mental burden of shopping for small ticket items.

Two recent micropayments—Flattr and Brave—seem to be making efforts to knock down the mental accounting barrier. In the original version of Flattr, which debuted in 2010, users had to manually click a Flattr button that had been embedded on a website in order to reward the content creator with a few cents. Flattr never really took off, no doubt due to the inconvenience of having to think about pushing the button.

The new version of Flattr, which debuted in late 2017, automates the process of making small payments. After users download a browser extension, Flattr algorithms monitor how much attention the user gives to each website that is visited. This includes not only time spent, but also whether the page has been recently scrolled or the mouse pointer moved. Using the attention metrics it has gathered, Flattr automatically distributes micropayments from the user’s monthly subscription across all of the websites visited.

Brave is similar to Flattr. Users download the Brave web browser and activate Brave Rewards. The browser tracks the user’s content consumption and automatically distributes a pro-rata reward to the sites that have been visited based upon attention scores generated by an algorithm. Whereas Flattr distributes fiat currency, Brave divvies up a user’s fixed amount BAT, or Basic Attention Tokens.

Bill Gates would only hire a shopping assistant who he trusts will guard his privacy. Likewise, Brave and Flattr go out of their way to assure that user data such as the amount of time devoted to each site stays locked in the device. The Brave blockchain and Flattr only receive enough just information to expedite the payment.

Will these ‘assisted’ micropayments platforms get any traction? That remains to be seen. It could be that the mental accounting barrier to tiny payments is just too difficult to for humans—or our machines—to scale.