Sixty-three cells. That is the count I landed on after opening a nine-dimension Web3 due-diligence report that crossed my desk this week — a framework with every table intact and every substantive field empty.
Technical positioning: N/A. Token type: N/A. Supply distribution: N/A. Team: N/A. Jurisdiction: N/A. Composite risk grade: "unable to rate — insufficient evidence."
The skeleton was immaculate. A Howey test grid with all four prongs laid out. A chain-propagation diagram with arrows running from mining hardware through protocol layers into DeFi and out to traditional finance. A risk matrix with seven rows — contract exploit, bridge failure, key management, securities exposure — and every probability column reading N/A.
This is not a failure of analysis. It is the analysis. In a market where the average "research report" is a press release with a spreadsheet stapled to it, a document that refuses to fill blanks is doing the one thing the industry has quietly stopped doing: it is marking exactly where the evidence ends.
The template economy outlived the bull market it was built for. Between 2021 and 2023, research became a deliverable rather than a process. Every exchange listing desk, every fund's two-person "research arm," every paid Telegram group needed a nine-dimension framework because the format itself conferred authority. Technical, tokenomics, market, ecosystem, regulatory, team, risk, narrative, value-chain. Nine boxes. Fill nine boxes, ship the PDF, invoice.
The projects changed. The templates didn't. So we now run 2026-era assets through 2021-era checklists, and the checklist keeps asking questions the source material cannot answer — because the source material was never a technical disclosure. It was a thread. It was a roadmap graphic. It was a founder's podcast appearance.
That is the structural condition this document exposes. The pipeline is: source article → information-point extraction → nine-dimension scoring → conviction. When extraction returns a scaffold instead of facts, the scoring stage faces exactly two options. Fabricate, or refuse. Most refuse silently by writing "strong fundamentals" and moving on. This one wrote N/A sixty-three times and published the count.
Start with the technical dimension, because that is where blank fields do the most damage. Whether an asset is L1, L2, application-layer, or pure infrastructure determines everything downstream — capital intensity, revenue model, competitive moat, regulatory surface. Without it, the four-row evaluation grid (innovation, maturity, security assumptions, performance) has nothing to compare against.
Here is what a real technical read looks like, and why you cannot template it. Last quarter I sat with an operator team running a ZK rollup and pulled three months of proving logs. The number that decides whether that business exists is cost-per-proof against amortized calldata savings per batch. When proving overhead exceeds the fee revenue the batch generates, every block the sequencer produces is a small loss. Scale does not fix it — it multiplies it. In the current fee environment, a meaningful share of general-purpose rollups are running that equation negative and funding it from treasury.
You cannot infer that from a whitepaper. You cannot infer it from a funding announcement. And you certainly cannot infer it from a document where the performance row reads N/A. A technical section without proving costs, sequencer decentralization status, and admin-key scope is decoration, not diligence.
Tokenomics is where honesty usually dies first. The standard four-row supply table — team, early investors, community/liquidity, treasury — is not complicated. It is public in most cases, or it should be. When all four rows return N/A, you have not learned that the distribution is fair. You have learned that nobody has published it, which is a different and more useful fact.
I have audited enough of these to recognize the tell. A filled table showing "35% community" with an empty unlock column is worse than a blank table. The percentage is marketing; the unlock schedule is the trade. Vesting cliffs, insider allocation, and treasury spend rate are the three variables that convert a supply chart into a dilution forecast. Remove them and the FDV/revenue ratio becomes a number with no denominator.
Governance blankness deserves its own paragraph, because the neutral reading is the wrong one. A governance row marked N/A does not mean governance is unknown. In practice it usually means governance exists, functions, and produces outcomes nobody wants to publish. On-chain voter turnout across major DAOs has sat under five percent of eligible supply for years. Proposal quorum is routinely met by four to nine addresses. Delegation concentrates further, because delegation is free and attention is not.
So when a report cannot confirm "whether the governance mechanism is genuinely decentralized," the honest answer is not "unknown." It is "probably not, and the data to prove it is on Snapshot in about eleven minutes of clicking." Neutrality in a governance field is almost always a choice, not a limitation.
The market and ecosystem dimensions are where the retention cliff hides. Daily active users, next-day retention, seven-day retention, thirty-day retention, organic retention after incentives end. That fifth metric is the only one that matters and the only one rarely disclosed.
Every incentive program produces a retention curve that looks like a cliff face. In my experience documenting these events block by block — I spent seventy-two hours in 2022 tracking oracle feeds through a peg collapse because I had capital sitting in an aggregator that froze — the pattern is consistent. Activity holds while emissions flow, decays by half within two weeks of the final distribution, and stabilizes at whatever genuine demand exists. That stabilized number is the protocol's actual user base. It is usually between four and eleven percent of peak.
Cross that against developer signal — contributor count, commit frequency, contract deployments — and you have a real ecosystem read. Cross nothing, and you have a logo wall.
Regulatory fields are the most commonly mistranslated. Four Howey prongs, all N/A. The instinct is to read that as "no securities risk identified." Wrong direction entirely. An unassessed prong is an open prong.
Last year, working through the post-ETF custody landscape, I spent two weeks translating a Wall Street compliance officer's framework into something a retail reader in Jakarta could act on. The lesson from that conversation: securities exposure rarely lives in the code. It lives in the marketing. A token with no promise of profit and no centralized promotion is hard to classify. A token with a foundation, a roadmap, a revenue share, and a founder on podcasts promising ecosystem growth is a much easier case to make. The compliance question is not "is it decentralized" — it is "what did the team say, where, and to whom."
The risk matrix, oddly, is the most functional section of the whole document. Seven rows, all rated "unknown," with an explicit note that unknown does not mean low. That distinction is the difference between a risk register and a risk theater. Every report that grades a bridge "medium" without reading the validator set is lying to its reader. Every report that writes "unknown — insufficient evidence" is telling the truth about its own limits.
Now the narrative dimension, where the blank field usually gets filled with someone else's story. A protocol with no disclosed technical differentiation often has an extremely well-developed narrative. In this cycle those narratives cluster around putting asset issuance on chains that were never built for it. Issuance layers bolted onto a settlement layer optimized for one thing — value transfer with maximum censorship resistance — produce exactly the throughput profile you would expect from loading freight onto a vehicle engineered for a completely different job. The vehicle is diminished and the cargo arrives late. The narrative survives anyway, because narrative does not require a benchmark.
Here is the contrarian read, and I hold it firmly: the sixty-three N/A document is the highest-value artifact in this market. Not because blankness is good, but because blankness is scarce. The industry currently pays for confidence. Analysts who produce confident tables get retained; analysts who produce empty tables get replaced. The incentive gradient runs directly toward fabrication, which means every filled cell in your average report carries an unknown discount rate. A document that voluntarily discounts itself to zero is the only one you can price.
There is a second layer. The absence of verifiable data is itself a signal. When a team has shipped, the data exists — GitHub, audit reports, mainnet state, Dune dashboards, Snapshot history. When the data does not exist, the shipping did not happen. Sixty-three blanks are not sixty-three mysteries. They are approximately eight identifiable gaps where the work has not been done, and one giant gap where the disclosure has not been made.
That is tradeable. Not as a short, necessarily. As a filter. The projects that survive this drawdown will be the ones whose founders can fill in fifteen of those sixty-three cells from memory, with links, in under an hour. The ones that cannot are the ones that will need the narrative to hold a little longer than the treasury can.
The question I am sitting with is not whether the report should have contained more. It is why the framework accepted sixty-three N/A values without raising a single error. A due-diligence template that cannot fail is not a diligence tool. It is a container for whatever the analyst is willing to assert. The next version of this pipeline will be judged by one field and one field only: whether every filled cell ships with a source. Watch the source column. That is where the next eighteen months get decided.