Last month, I bought my teammate a mouse on Amazon Singapore. He sent me the link. I clicked “Buy Now.” The purchase was automatically charged to my American credit card, and the payment was seamlessly converted from USD (U.S. dollars) to SGD (Singapore dollars). Best of all, the process took all of 30 seconds.

The hard part came when he paid me back. “I’ll take cash or USDC-SOL,” I said.

Grace “Ori” Kwan is the cofounder of Orca, a decentralized exchange on Solana. This article is part of CoinDesk’s Payments Week series.

Being colleagues at a decentralized finance (DeFi) startup, the latter seemed like the obvious choice. With the super-speed and low fees of Solana, it’d be easy, right? If neither of us cared about privacy, and the network was feeling cooperative that day, it could have been.

But here’s what the process looked like:

I opened my Solana wallet Phantom, generated a fresh address and sent it over via Discord. He logged into the crypto exchange FTX, exchanged some SOL for USDC and “withdrew” $0.01 to my address for the customary test transaction. FTX sent an email confirming that the withdrawal had been processed – but 15 minutes later, I still hadn’t received anything.

Maybe the network was backed up? Who knows. We went home for the day.

The next day, my $0.01 had come through. Hurray! He sent the rest of the tokens, only to realize that he’d forgotten to account for the exchange rate between Singapore dollars and USD (128 SGD = ~96 USD). No problem – I could just send back the extra to the same address, right? Wrong – since he’d sent it through a centralized exchange, the receiving address would be different. Sigh.

By the time we got the payment sorted, the mouse ordered on Amazon had already arrived.

See also: Why Haven't Crypto Payments Taken Off?

We live in a golden age of capitalist convenience. Crypto is supposed to be the next frontier. So why are crypto payments so damn hard? And how might we make them better?

As an engineer-turned-designer, my answer is: human-centered design.

If you haven’t heard of human-centered design, or HCD, it’s a simple concept: Observe the behaviors of your target users (aka real humans), learn what they need and design a product that caters to those needs.

In contrast, many crypto products embrace what one might call technology-centered design: Observe some underlying data and translate it directly as possible into the user interface. If data is stored as a table in the database, display it as a table in the UI, for example.

Often, the result of this tech-focused design process is an interface that contains all the necessary information, yet appears totally unfamiliar to first-time users.

So are we just talking about new interfaces? All new interfaces – that’s innovation, right? Not quite. But alas, “new” does not mean “better,” at least when it comes to design. Web 2 giants have spent years designing, developing and battle-testing the easiest ways to send money on the web. If we hope to invite the wider world into crypto, we must offer experiences that are at least as good as what’s already out there.

In many cases, that may mean black-boxing all the messy details that crypto apps leave naked for everyone to see.

Think of the last time you made a payment. Who were you paying? The first thing that comes to mind is likely a person or a business. It’s probably not a string of seemingly random letters and numbers.

And yet, crypto wallets and exchanges are largely address-centric. Sure, you can save someone’s address in most wallets, but that’s the same way my mom stored contacts back when all we had was a landline phone.

Web 2 solved this problem long ago through social profiles. In WeChat, I can pay a friend directly through the same interface we use to chat. No need to scan a QR code, trade a phone number or ask for any other identification. And yet, in Web 3, we’re stuck with the address book.

So, my controversial advice? Make Web 3 more like Web 2.

“But no!” the crypto purists are sure to cry. For many, Web 3 means decentralization, which in turn represents freedom from the ever-watching eyes of giant social networks. By attempting to emulate the user experience of Web 2, wouldn’t we sacrifice those ideals?

Yes and no. In some cases, users will happily trade away some degree of anonymity in exchange for a better user experience, which is a trade they should feel empowered to make. In other cases, the tradeoff itself is a fallacy: An experience more like Web 2 can be achieved through decentralized means, such as the Ethereum Name Service. (Similarly, the tagline for the DeFi protocol I co-founded in 2021 Orca is “The DEX for people, not programs.”)

For crypto payments to become mainstream, they must offer additional benefits without sacrificing good design.

See also: What Web 3 Can Learn From Steve Jobs | Opinion

In this sense, new approaches to on-chain payments seem promising. Published images of Solana Pay feature a familiar QR code-based payment interface. (Not my favorite payment UI – tap-to-pay interfaces like Suica and payWave win that contest – but at least there’s no guessing how to use it.)

The exact amount of payments and the identity of payers are not publicly viewable on the blockchain, offering a degree of the privacy users naturally expect from a payment platform.

Plus, it promises merchants lower fees and new approaches to build loyalty with customers through non-fungible tokens (NFTs). However, the protocol has not yet been tested at scale, and the reliability of the network is still a concern. For Solana Pay to succeed, these new features can’t come at the cost of the basics: security, reliability and simplicity.

The golden age of payments is already here. Let’s not make it harder.