How to Research a Crypto Project (DYOR, Properly)

"Do your own research" is crypto's most repeated advice and its least explained, a disclaimer wearing a helpful expression. This guide is the missing method: the sequence we'd run before putting money into an unfamiliar project, refined from years of writing protocol analysis on this site. Our old GMX and mStable deep-dives ran on exactly these bones: product, usage, revenue, token design, then team. Budget about an hour per project. Most candidates fail early, which is the process working. The hour's purpose is finding reasons to say no cheaply, before the market finds them for you expensively.

Step 1: Make the project explain itself in one sentence

Open the project's own site and docs and extract: what does this do, and who needs it done? A real project survives this compression. "A derivatives DEX on Arbitrum sharing trading fees with token stakers" is an answer. Warning signs at this stage: word-salad ("empowering community-driven synergies"), a roadmap that's all partnerships and listings, or a pitch that's entirely about the token's price mechanics rather than anything the protocol does. If ten minutes of reading can't produce the sentence, the project either doesn't have one or doesn't want you to notice. Same conclusion either way.

Step 2: Check whether anyone actually uses it

Claims are marketing; usage is evidence, and in crypto the evidence is public. Pull up the project on DeFiLlama (TVL and, crucially, fees and revenue), a block explorer (transaction counts, active addresses), and its live app. The questions: is there real activity, is it trending up or bleeding out, and does any revenue exist, meaning users paying fees, or does all "growth" trace back to token emissions renting temporary liquidity? Our old fundamentals work leaned hardest on exactly this cut. Fee revenue and daily users separated the protocols that survived 2022 from the ones that evaporated. A project with modest but real usage beats one with spectacular numbers that all come from incentive printing.

Step 3: Read the tokenomics like an adversary

Now the token itself, with the full tokenomics toolkit: supply distribution (who got what, at what price), unlock calendar (what hits the market in the next year, sized against daily volume), emissions (revenue-funded or dilution-funded), and the demand case, meaning what holding gets you and whether the market cap versus FDV gap means you're buying the top of someone else's vesting schedule. Then the mechanical safety checks from the scam gauntlet: verified contract, locked liquidity, sellable token, holder concentration on the explorer. Adversarial is the right posture here. Assume the design is optimised to transfer value from late buyers until the document proves otherwise, because that's the base rate.

Step 4: Weigh the team, the code's exposure, and the money behind it

Who builds this, and what happens if the code fails? Named teams with track records beat anonymous ones. Not disqualifyingly (Bitcoin's author remains anonymous), but anonymity plus treasury control plus short history is a stack of risk to price. Check the development pulse (public repos, shipping cadence) and audit status with appropriate scepticism: an audit is a review, not a warranty, and code that's held real value through real time is the stronger signal. Note admin powers: upgrade keys, pause switches, and who holds them. Investor names cut both ways. Serious backers do diligence you can free-ride on, and their tokens unlock on a schedule you just read in step 3.

Step 5: Write the thesis down before you buy

The step that separates research from browsing: write three sentences. Why this wins, what would prove you wrong, and when you'd exit. Written, not mental, because every bias in the book will redecorate your memory later, and a thesis you can't falsify is a mood. Then size the position to the volatility you just signed up for and let the data, the same kind we publish on our coin pages, tell you over time whether the thesis is tracking. And a closing honesty from a site that's been wrong before: research improves your odds, it doesn't guarantee them. We rated protocols highly that later stumbled, and recommended an exchange that failed outright. The process's real product isn't certainty. It's smaller mistakes, caught earlier, sized survivably. That compounds.

Frequently asked questions

What should I look at first when researching a crypto project?

Usage and revenue, because they're the hardest things to fake: DeFiLlama for fees and TVL, a block explorer for active addresses. A project people pay to use clears a bar most of the market never reaches, and you can check it in five minutes before reading a word of marketing.

How do you check who owns a crypto token's supply?

Block explorers list top holders for any token: look at what share the largest wallets control, separating exchange and contract wallets from insiders where labeled. Heavy concentration in a few unlabeled wallets plus unlocked liquidity is the classic rug-pull setup.

How long should crypto research take?

About an hour per project for a first pass covering product, usage, tokenomics, and team, and most candidates disqualify themselves in the first twenty minutes. Anything you're sizing seriously deserves a second, slower pass. Less than that isn't research; it's permission.