In 2016, Aaron Davis shipped the first version of MetaMask as a browser extension with one purpose: let developers interact with Ethereum without running a full node.
It worked. For developers.
By 2021, MetaMask had 30 million monthly active users and a valuation that made it one of the most successful crypto products ever built. It was also, by almost every UX benchmark, one of the worst.
How does the most-used wallet in crypto lose the design war? That question has a specific answer.
What MetaMask inherited
The architecture decision that shaped everything
MetaMask was built for a world where the user was a developer. That assumption was never a design choice — it was a technical default that calcified into product philosophy.
The result:
| What users saw | What it meant |
|---|---|
0x095ea7b3000000000... |
Token approval signature |
Gas limit: 21000 |
Minimum transaction fee |
Nonce: 47 |
Transaction sequence number |
Chain ID: 1 |
Ethereum mainnet |
None of these were explained. None were designed. They were surfaced because that is what the protocol exposed, and MetaMask surfaced the protocol.
Joseph Lubin has talked about ConsenSys, MetaMask's parent company, operating with a thesis that developers were the primary user. Build the infrastructure, developers build the experience. MetaMask was infrastructure.
The problem: 30 million people were using it as a consumer product.
The moment competitors understood what MetaMask missed
Phantom's founding assumption
When Francesco Agosti and the Phantom team started building in 2021, they made a single decision that produced a completely different product:
Treat the wallet as a product for a person who does not know what a seed phrase is.
That assumption had downstream effects on every surface:
MetaMask: "Secret Recovery Phrase"
Phantom: "Your Secret Recovery Phrase is like a master password..."
[Plain-language explanation of what this actually is]
[Why losing it means losing your funds permanently]
[What to do and what not to do]
MetaMask labeled it. Phantom explained it.
Same information. Different product philosophy.
Rainbow's founding question
Mike Demarais, Rainbow's co-founder, has been specific about why they built it: because existing wallets felt like infrastructure, not products.
Rainbow was built around a different question than MetaMask. Not does it work but would someone want to use this daily if they didn't have to?
That question produced animated transitions, a color-coded portfolio view, social features, and a design language that made a financial product feel personal.
The counterargument MetaMask would make
Here is the version of this story that favors MetaMask:
MetaMask was first. MetaMask shipped when no design playbook existed. MetaMask supported every chain, every token, every edge case that polished competitors have not yet encountered. And MetaMask still has more monthly active users than Phantom and Rainbow combined.
Dan Finlay, MetaMask's co-founder, has pushed back against the narrative that MetaMask "lost." His argument: most wallet comparisons are made by people who use Ethereum mainnet. MetaMask's complexity reflects actual protocol complexity, not design negligence.
This argument has real weight.
Phantom launched on Solana — a chain with faster confirmations, lower fees, and far fewer edge cases than Ethereum mainnet. Rainbow launched with a single-chain focus. MetaMask supports everything. Supporting everything is harder to make look simple.
The counterargument does not explain why MetaMask's onboarding flow never received the same investment as its protocol support. It explains the difficulty. It does not explain the choice.
What the autopsy actually shows
MetaMask did not lose the UX war because of technical constraints. Every design decision Phantom and Rainbow made was available to MetaMask.
It lost because of organizational philosophy.
| Factor | MetaMask | Phantom / Rainbow |
|---|---|---|
| Primary user assumption | Developer | First-time user |
| Design investment timing | After scale | Before launch |
| UX authority | Shared across teams | Core product function |
| Error message ownership | Engineers | Product + design |
| Success metric | MAUs | Daily return rate |
The technical capability was identical. The organizational priority was not.
The specific failure: error states
MetaMask's error messages, as of 2023, remained engineer-authored strings:
❌ "Transaction failed"
❌ "Nonce too low"
❌ "Intrinsic gas too low"
❌ "execution reverted"
Phantom's equivalents:
✅ "This transaction failed because the price moved too fast.
Try again with a higher slippage tolerance."
✅ "You don't have enough SOL to cover this transaction fee."
These are not hard to write. They require someone to care enough to write them. MetaMask had 30 million users and did not consider this a priority.
What this means for every product built on top of a protocol
MetaMask is not a cautionary tale about a bad company. It is a case study in what happens when infrastructure values get applied to consumer products.
Protocol values:
- ✅ Completeness (expose everything)
- ✅ Correctness (never hide a failure)
- ✅ Flexibility (support all cases)
Consumer product values:
- ✅ Clarity (explain what matters)
- ✅ Recovery (show the path forward)
- ✅ Opinion (hide what is irrelevant)
When a product team operates with protocol values on a consumer surface, users see raw protocol state. That is fine for developers. It fails everyone else.
The wallets that grew faster than MetaMask after 2021 did not have better protocol support. They had a clearer opinion about who they were building for.
What this means practically
Define the user before the technical spec. MetaMask defined the technical spec first. Every UX decision after that was constrained by it.
Error messages are product decisions, not engineering cleanup. The message a user sees when a transaction fails is as important as the transaction itself. Assign ownership.
MAUs are a lagging indicator. MetaMask's user count masked the design debt accumulating underneath it. The users who left for Phantom and Rainbow showed up in competitor growth before they showed up in MetaMask's churn.
Being first does not mean having a design advantage. It means having a user base before you understand those users. That is an opportunity — or a liability, depending on what you do next.
MetaMask is still the most widely integrated wallet in crypto. It is also a monument to what happens when thirty million people use infrastructure that was never designed for them.
The gap between those two facts is where the next generation of wallet products will be built.
I work as a Web3 creative director helping teams figure out which user they are actually building for — before 30 million people answer that question for them. somaryuu.xyz
