Every ecosystem eventually builds a discovery surface. A page that tells you which apps are worth your attention, so you do not have to read every launch announcement yourself. Core, Avalanche's native wallet, ships one inside the wallet. App stores do it. So does every "top protocols" leaderboard you have ever scrolled.
They almost all rank the same way underneath: by usage. Active addresses, transaction counts, unique wallets this week. It is the most honest available signal, and onchain it has the rare virtue of being public and checkable by anyone.
There is one hole in it, and it is worth understanding before you trust any ranking built this way.
Usage is counted in addresses, and addresses are free
Creating an Avalanche address costs nothing. It is a keypair. You can make ten thousand of them on a laptop in under a minute, and no one can stop you, because there is nothing to stop. The chain has no concept of a person.
So when a protocol reports ten thousand users this week, the accurate reading is: ten thousand addresses touched this contract. Those two statements look identical in a dashboard and mean very different things. One describes a community. The other describes a number that someone with a script and a gas budget can produce on demand.
This is not hypothetical, and it is not rare. Wherever usage counts decide something valuable, someone manufactures usage. Points campaigns get farmed. Airdrop allowlists fill with wallets created the week of the snapshot. Quest platforms measure completions, and completions are cheap.
The ranking inherits the problem
Here is the part that matters for discovery. If a ranking is built on usage counts, and usage counts can be manufactured, then the ranking can be manufactured too. Not by breaking anything. By playing the metric exactly as designed.
An app with ten thousand fresh, empty, single-purpose wallets and an app with ten thousand people who have been onchain for two years produce the same row in the table. The surface showing you what is "most used" cannot tell them apart, because the data it reads does not contain the difference.
The uncomfortable version: the projects most willing to manufacture usage are, on average, the ones you would least want a discovery page to recommend. A usage-only ranking quietly rewards them.
The missing step is measuring the users
You cannot fix this by counting more carefully. No amount of care separates a real address from a manufactured one if the only thing you look at is how many there were.
What does separate them is the history behind each address. Not identity, and not a name. Just the record the chain already keeps:
- Age. When did this address first transact? A wallet created the week of a snapshot is not the same as one that has been active since 2023.
- Cost paid. Gas spent over a lifetime is the one signal that cannot be faked for free. It is the closest thing onchain has to proof that someone bothered.
- Shape of activity. A real wallet interacts with many unrelated contracts over time. A manufactured one usually does one thing, in a burst, alongside a thousand siblings doing the same thing.
- Graph anomalies. Farmed accounts tend to cluster. They follow each other, fund each other, and act in rings. Those patterns are visible without knowing who anyone is.
None of these is proof of personhood, and anyone who tells you their score proves a human is behind an address is overselling. What they give you is something more modest and more useful: a cost asymmetry. Making one address look established is easy. Making ten thousand of them look established simultaneously is expensive enough that it stops being worth doing.
Why the method has to be published
A reputation number that nobody can inspect is just another ranking asking for trust, and it fails in a specific way: the people with the most incentive to reverse engineer it will do so anyway, while everyone honest is left guessing what the rules are.
Publishing the categories, the weights and the design principles does not make a score easier to game. The signals that carry weight are the ones that cost real money and real time to produce, which is exactly why they can be published safely. Knowing that wallet age matters does not let you go back and start your wallet in 2023.
It also lets someone disagree with you in a concrete way, which is the only kind of criticism that improves a formula.
Where this leaves discovery
Ranking apps by usage is a good instinct. The question "what do people actually use, not just what is loudest" is the right question to ask, and the surfaces asking it are more useful than the ones ranking by announcement volume.
It just needs one layer underneath it. Before you can trust a count of users, someone has to measure the users. That is a separate job from ranking the apps, and it is the job SOCI4L does on Avalanche: scoring any C-Chain address from what it has actually done onchain, with the methodology published and the results readable over an API.
You can check any address yourself, read exactly how the number is put together, and disagree with the weights if you want to. That last part is the point.