Back in 2022, at the peak of NFT mania, I co-authored a paper for the International Conference on Blockchain Technology and Applications called “NFT Scoring: An Analysis of the Considerable Features.” The goal was simple: most people were buying NFTs based on floor price and vibes, and almost nobody had a repeatable way to actually score one.
The market has changed a lot since then. The PFP gold rush is over. Trading volume dropped roughly 37% year-over-year in 2025 to about $5.5 billion, according to a KuCoin market report, and by some estimates the large majority of collections from that era now sit untouched with no active trading at all. But the core question the paper tried to answer, “what actually makes one NFT worth more than another,” didn’t go away. If anything, it matters more now, because the NFTs still getting attention in 2026 are the ones tied to gaming items, ticketing, or membership access rather than JPEGs someone hopes will 100x.
So here’s the framework, stripped of the academic language, with a few notes on where I’d point it differently today.
The three numbers we proposed.
The paper split NFT valuation into three separate scores instead of one blended number, because “value” for an NFT actually comes from three different places that don’t move together.
Liquidity score. This measures how easily you could sell the thing, not what it’s theoretically worth. We calculated it from a collection’s floor price movement over time. A collection can look expensive on paper and still be nearly impossible to exit if trading volume is thin. In 2026, this is arguably the single most important number to check first, since most collections now trade in occasional bursts rather than continuously.
Utility score. This looks at what the collection actually does beyond ownership: does it unlock a game feature, a real-world perk, staking rewards, access to something ongoing? When we ran the numbers in 2022, this score barely mattered to most buyers. Speculation drove almost everything. That’s flipped. Multiple 2026 market reports point to utility-driven and access-based collections as one of the few segments still pulling in real activity, so this score has gone from a nice-to-have in our model to the one worth weighting the most.
Rarity score. This is the one most people already track informally through tools like Rarity Tools, based on how uncommon an NFT’s individual traits are within its collection. It’s the score most people default to because it’s the easiest to compute, but on its own it tells you almost nothing about whether you can sell the thing or whether it does anything.
Why splitting these apart matters.
The mistake we saw constantly in 2021 and 2022 was collapsing all three into a single “this NFT is good” judgment, usually based on rarity alone, because rarity is the one that’s visible on every marketplace listing. That’s how people ended up holding NFTs with rare traits, zero liquidity, and no function at all.
Keeping the three scores separate forces you to ask three different questions instead of one: can I sell this, does it do anything, and is my specific piece unusual within its set? A collection can score well on one and badly on the other two, and knowing which is which changes what you actually do with the asset.
What I’d change if I rewrote this today.
If I were running this study now instead of in 2022, I’d weight utility score much more heavily, probably as the primary filter rather than a third factor considered alongside the other two. Back then it was genuinely hard to find enough utility-driven collections to study; that’s no longer true.
I’d also add a fourth axis the original paper doesn’t cover: how dependent the “utility” actually is on a single company or platform staying alive. A lot of what gets marketed as utility today is really a promise, redeemable only as long as one studio or brand keeps supporting it. That’s a different kind of risk than illiquidity, and it deserves its own number.
The Practical Version
You don’t need a spreadsheet model to use this. Before buying, or before assuming something you already hold is worth what a marketplace listing says, ask the three questions in this order:
- Could I actually sell this at a reasonable price this week, or would I be waiting for a rare buyer? (liquidity)
- Does this NFT do something beyond letting me claim ownership, and does that something depend on one company staying in business? (utility)
- Is my specific piece meaningfully rarer than the rest of the collection, or just marketed that way? (rarity)
Most of the value questions people ask about NFTs collapse into one of these three, and mixing them together is usually what leads to a bad read on what you’re actually holding.
This is a simplified, practitioner-facing rewrite of the original peer-reviewed paper, “NFT Scoring: An Analysis of the Considerable Features” (ACM ICBTA 2022). The full paper includes the underlying formulas and statistical population used to test the model.
