The spreadsheet arrived on a Tuesday, forwarded without comment. Nine analytical dimensions. Forty-seven sub-metrics. A six-category risk matrix with probability, impact, and mitigation columns, color-coded in the way that signals seriousness to a client. Every single cell read the same three characters: N/A. Not pending data, not low confidence — just N/A, repeating down the page like a flatline. The pipeline had executed without error. The template had populated. Somewhere upstream, the source document had never loaded, and nothing in the system was built to notice. The output was structurally perfect, epistemically empty, and indistinguishable at a glance from a report.
I have been that spreadsheet. In 2017, at thirty-three, I sat in a Barcelona co-working space with fifty-one ICO whitepapers stacked in a browser window and a growing suspicion that the word "utility" was doing work the code could not. I published a critique of the utility-token fallacy and was called a cynic for eighteen months. The correction arrived anyway. What I learned then is what that N/A artifact confirmed last week: a framework with nine dimensions cannot report zero. It will always produce something. Whether that something is information is a completely separate question.
By 2025 I was writing about compliant decentralization and the institutional identity layer settling into regulatory clarity, and the work was genuinely interesting. But somewhere in that period the research itself changed shape. The memos got longer, the scoring rubrics got finer, and the distance between the page count and the actual knowledge inside it kept widening.
The industrialization of crypto research happened between 2019 and 2024, and it happened for a defensible reason. As institutional capital entered, allocators needed comparable output across dozens of protocols; you cannot diligence forty positions with forty bespoke memos. So the industry built templates — technical, tokenomic, market, ecosystem, regulatory, team, risk, narrative, supply-chain — and templates scaled. The cost was that coverage became the product. A fund's internal diligence deck now needs a slide for every dimension whether or not the underlying data exists, because a missing slide reads as laziness and a slide full of N/A reads as diligence, and that particular kind of diligence is expensive in a bull market and nearly worthless in the one we are in now.
We are in a bear market, and that changes the question readers bring to any piece of analysis. In 2021 they wanted to know which protocol would ten-x. Today they want to know whether the thing they are holding is bleeding. That is a narrower question, a more answerable one, and precisely the question most nine-dimension frameworks are worst at answering — because the answer rarely lives in the dimension labeled "narrative." It lives in three columns nobody enjoys filling: who pays, how much, and when do they stop.
Consider what those columns look like when you actually go and get the data. Take tokenized real-world assets, a story that is now three years old and counting. I traced one 2024–2025 pilot end to end: a tokenized money-market fund, permissioned ledger, transfers gated to a KYC allowlist, the cash leg settled on a private rail, and a public chain used as a timestamping notary and nothing else. The AUM, the NAV, the redemption terms, the counterparty exposure — all of it lived inside a custodian's database. "On-chain" was an adjective applied at the marketing layer. None of it will ever appear in a block explorer, which means any framework scoring RWA by on-chain metrics is scoring the shadow of the thing rather than the thing.
Take data availability, where an enormous share of the last two years of infrastructure capital went. Post-EIP-4844, a rollup's blob spend is measurable and small: most production rollups post a handful of blobs per hour, and the median blob carries far less than its target capacity. The economics of a dedicated DA layer require a customer base whose demand curve has not arrived, which is why so much DA capacity is being sold rather than bought. Capacity is a supply-side narrative. It always sounds like a roadmap and reads like a warehouse.
Take Bitcoin miners, now well past the April 2024 halving. The subsidy is 3.125 BTC. Hashprice compressed through 2024 and 2025; fee revenue is lumpy and event-driven, spiking on inscription waves and settling back to a low single-digit share of miner income. At six cents a kilowatt-hour, a sub-megawatt operation running older-generation hardware sits at or below gross margin. The measurable consequence is consolidation: in a given week, two or three pools routinely control a majority of blocks. Decentralization here is a distribution question with integers attached, and the integers are drifting.
Three sectors, three narratives, one identical analytical shape. In every case the survival question reduces to a handful of numbers that fit on a napkin, and in every case those numbers are absent from the template — because the template was built to describe technology while the question is about cash flow. That is the gap the N/A artifact exposed without meaning to. It had no place to put an empty field that mattered. So it stamped N/A into forty-seven places and let the reader draw whatever confidence they wanted from the shape of the page.
When I run this audit now — and I have run some version of it on roughly sixty protocols since 2017 — I allow myself three measurements and refuse a fourth. Protocol-attributable revenue, meaning fees paid by users who are neither the treasury nor the emission schedule. The unlock calendar in absolute tokens, mapped against the emission-funded demand that has been absorbing it. And the number of addresses that have used the thing twice in ninety days. Everything else — the throughput claims, the partnership announcements, the grant programs — is context for those figures, not a replacement for them. It took me a decade to learn that a short list of real numbers beats a long list of structured guesses, and the market has never once rewarded me for the longer document.
The DeFi Summer of 2020 taught me the same lesson from the other direction. I spent that year inside Uniswap's liquidity incentives, writing about the social contract between LPs and the protocol that hosted them. What the incentives revealed was that most of the liquidity was rented, and that the metric everyone quoted — total value locked — measured the size of the subsidy more accurately than the size of the demand. Six years later, the same measurement error reappears in every RWA dashboard and every DA utilization chart. We are still quoting the size of the warehouse and calling it the size of the market.
There is a name for this in behavioral economics: the illusion of explanatory depth. We mistake the existence of a framework for the existence of understanding. A filled template triggers the same recognition reflex as a filled bank statement — completeness reads as competence, even when the completeness is pure formatting. After the 2022 collapse I spent months in near-total retreat, auditing my own calls and writing about the toll of it. The hardest part was not the calls I got wrong. It was discovering how many of the calls I got right had been assembled from the shape of the page rather than the contents of the evidence.
Which is why I will argue the opposite of what my instincts want. The empty report is more valuable than the filled one. Give me a dimension marked N/A and I know a question was asked and the answer was unavailable; give me that same dimension scored three out of five on a hunch and I have been handed manufactured confidence complete with a color code. The industry's incentive gradient runs the wrong way here. Analysts are compensated for coverage, publishers for volume, funds for deployment — and none of those incentives pay for the sentence "I don't know." So the sentence rarely gets written, and the market prices the resulting silence as consensus. To hunt the truth, one must first bury the hype — and more often than anyone admits, the hype is nothing but a well-formatted spreadsheet.
Pay close attention to what N/A actually asserts. It is not the absence of an opinion. It is a positive claim: the question was defined, the sources were checked, and the answer was not there. That is a finding, and a difficult one to produce. Most reports skip the definition step and the check step entirely and go straight to the answer, which is why so many of them read as authoritative in January and evaporate in July on contact with a block explorer. An N/A in the right column is worth more than a paragraph of inference.
The next narrative in this cycle will not be real-world assets, or AI agents, or restaking. Every one of those sits downstream of a simpler unsolved problem: almost nobody can prove where a claim came from. Provenance — of data, of attestations, of the integers inside the deck — is the unmet need, and it is the one the bear market is quietly pricing. When the next nine-dimension report lands in your inbox, with its color-coded matrix and its confident scoring, ask one question before you read the conclusion. How many cells say N/A? That number will tell you more about the analyst than the analysis ever will.