A Window Is Not a Witness: Reading the Blockworks OlympusDAO Dashboard
Four information points. That is the entire payload.
A third party built a public dashboard that tracks OlympusDAO's books. Two of the four points are somebody's opinion about what that means. One is a source note. One is the fact itself. No contract address. No data schema. No publication date. No treasury figure. No supply number. No audit reference.
I have spent ten years reading crypto announcements, and the ones that matter almost never arrive this thin. The ones that don't matter usually do.
Charts lie. Liquidity speaks. And this chart has not spoken yet.
So let me do the thing that the press release did not do. Let me separate what shipped from what was claimed, and what was claimed from what can actually be verified. Because there is a specific failure mode in this genre of news, and it is not dishonesty. It is something quieter and more expensive.
It is the substitution of visibility for verification.
Visibility is cheap. Verification is expensive. Every cycle, a project confuses the first for the second, prints a press release, and waits for the market to re-rate it. The market does not re-rate visibility. The market re-rates cash flows, collateral, and exit liquidity. Everything else is narrative.
This article is about a dashboard. It is also about a pattern. The dashboard is the small story. The pattern is the one that keeps taking money from people who read headlines instead of reading code.
What Actually Shipped
Strip the adjectives and here is what exists.
Blockworks, a crypto media and research firm, released a public dashboard that tracks OlympusDAO's accounts. That is the hard fact. It is a data visualization product. It indexes some set of Olympus-related numbers and renders them in a browser.
The rest of the message is framing. The framing says this improves transparency. The framing says this may increase institutional trust and investment. Both of those are claims about consequences, not descriptions of mechanisms.
Now, the interesting part.
A dashboard is a rendering layer. It sits on top of data. It does not create data, validate data, or guarantee data. It takes numbers from somewhere and draws them. The question that decides whether this announcement is meaningful or decorative is a single question, and the message does not answer it.
Where do the numbers come from, and can I reproduce them independently?
If the answer is "on-chain, from public contract state, with an open query anyone can re-run," then this is a genuine utility. It lowers the cost of diligence. It is a small, real good.
If the answer is "from our own internal index, with our own definitions, updated on our own schedule," then this is a marketing surface. It is not necessarily deceptive. It is simply not verification. It is a nicely typeset opinion about a balance sheet.
A dashboard is a story with axes. The axes make it feel like physics. The story is still a story.
Based on my audit experience, this distinction is where most retail capital gets mispriced. People treat a public number as a settled number. The number is only as settled as its provenance. And provenance is invisible on a screen. You have to go find it.
The Two Subjects Problem
The message collapses two technically separate subjects into one narrative. That collapse is where the confusion starts.
Subject A is the OlympusDAO protocol. Bonds. Rebases. Protocol-owned liquidity. Lending facilities. Live contracts on a mainnet, with admin keys, upgrade paths, and treasury assets.
Subject B is a Blockworks data product. An indexer, a database, and a frontend. A Web2-shaped tool that happens to display Web3 numbers.
These two things have almost nothing in common technically. One is a set of smart contracts that custody and deploy capital. The other is a piece of software that reads and draws. Lumping them together lets a claim about Subject B borrow credibility from Subject A, and vice versa.
A dashboard cannot change a protocol's trust assumptions. It cannot reduce the power of a multisig. It cannot make an upgradeable proxy immutable. It cannot make an illiquid treasury asset liquid. It cannot audit a contract it merely reads.
What it can do is make the existing state legible to more people. That is not nothing. But legibility and safety are different axes, and the announcement sells them as one.
I have watched this exact move before. In 2022 I spent months inside Lido's staking mechanics, mapping the validator set and the governance surface, while the market was busy reading a headline about staking yields. The headline was legible. The validator concentration was the risk. Nobody drew a dashboard for the thing that could actually hurt you.
The most dangerous number is the one that is easy to display. Easy-to-display numbers get displayed. Hard-to-display risks get ignored. That asymmetry is the whole game.
Olympus, Briefly: How a Bond Became a Bank
To understand what a treasury dashboard can and cannot prove, you have to understand what Olympus was actually building. And you have to understand it in sequence, because the project's meaning changed more than its code did.
Olympus launched in March 2021 on Ethereum. Its central mechanism was bonding. Instead of selling tokens for cash in a public sale, the protocol sold OHM at a discount to buyers who paid in assets β stablecoins, ETH, or liquidity provider tokens β and received OHM after a vesting period. The protocol kept the assets. The buyer kept the discount.
On paper, this is elegant. It converts speculative demand into permanent capital. It is a machine that turns hype into a balance sheet.
It is also a machine that turns hype into a balance sheet, which is the same sentence with a different stress pattern. The capital is only as good as the demand that keeps arriving.
The second mechanism was rebasing. OHM holders received new OHM every eight hours. The advertised yield was enormous. At the peak, the annualized figure was not in the double digits. It was in the thousands. For brief windows, higher.
And the third mechanism was protocol-owned liquidity. Olympus owned its own OHM trading pairs instead of renting liquidity from mercenary farmers. It was, in the language of the time, a central bank with a treasury and a currency and a monetary policy.
The phrase "DeFi 2.0" got attached to it. Dozens of forks copied the template within months. Wonderland on Avalanche. Klima for carbon. Redacted Cartel. Hector. A small industry of imitators, each with a rebase curve and a promise.
The design was genuinely novel in 2021. I will not pretend otherwise. The bonding-plus-POL combination was a real architectural contribution, and it reshaped how later protocols thought about liquidity ownership. There is a version of this story where Olympus is a footnote that became a foundation.
But the market did not price the architecture. The market priced the rebase.
FOMO is a tax on the unobservant. The 2021 OHM trade was, for most participants, a tax paid in the name of yield.
The Rebase Illusion
Here is the mechanic that mattered, and the one a dashboard is least equipped to fix.
Rebase yield was denominated in OHM. If you held one OHM and the protocol printed more OHM, your OHM count went up. Your share of the network stayed roughly flat. Your claim on the treasury per token drifted.
The number on the screen went up. The value of the claim did not necessarily follow.
This is the oldest trick in monetary design, and it predates crypto by centuries. Denominate the reward in the unit you control, and the reward can be as large as you want. Yield measured in a currency you can mint is not yield. It is dilution wearing a yield costume.
The people who understood this exited into strength. The people who did not stayed for the number.
A dashboard that shows "backing per OHM" is a partial corrective to this. It reframes the holder's position in terms of treasury assets rather than token count. That is a genuinely useful reframe. If the dashboard shows backing per OHM clearly, prominently, and with a reproducible methodology, it does one real thing: it moves the conversation from the numerator to the denominator.
But it only works if the denominator is honest. And the denominator depends entirely on how treasury assets are valued.
Which brings us to the part of the balance sheet that never renders well on a screen.
Protocol-Owned Liquidity: The Self-Referential Balance Sheet
Protocol-owned liquidity is the crown jewel of the Olympus design, and it is also the hardest asset to value honestly.
Here is the problem in plain terms. Olympus owns liquidity in OHM pairs. The value of that liquidity depends on the price of OHM. The price of OHM depends on the depth of that liquidity. The two are not independent. They are the same variable, viewed from two angles.
A treasury dashboard will list these positions. It will assign them a value. That value will be computed from current market prices. On a good day, the number looks solid. On a bad day, the number was always theoretical and the market simply revealed it.
Book value is a promise. Realizable value is a fact. A dashboard shows you the first and calls it the second.
This is not a criticism unique to Olympus. It is a structural property of any treasury that holds its own token or its own token's pairs. It is a circular reference in a spreadsheet, and the circularity does not disappear when you render it in a nice font.
I learned a version of this lesson the hard way in 2020. I deployed five hundred dollars into an arbitrage bot between two automated market makers. The math was clean. The execution was not. A slippage error cost me twenty percent in under an hour, because I had priced the trade against the pool's stated reserves and not against what the pool would actually pay me when I hit it.
The reserves were a number. The fill was a reality. The gap between them was my loss.
A treasury dashboard is the reserves. The market is the fill. Anyone who has ever tried to exit a position in a thin pool understands the distance between those two things. It is not a rounding error. It is the entire trade.
So when I read that a dashboard tracks Olympus's accounts, my first thought is not "transparency." My first thought is: which valuation methodology, and does it mark POL to market, to cost, or to hope?
Backing Per OHM: A Floor Made of Fog
The single most consequential number a treasury dashboard can display for a reserve-backed token is backing per unit. It is the number that anchors the entire investment thesis. It is the "floor."
And it is, almost always, softer than it looks.
Backing per OHM is computed by dividing some measure of treasury assets by some measure of supply. Both halves are contentious. The numerator depends on asset valuation and on which assets count. The denominator depends on how you treat vested, locked, and pending supply.
Change either assumption and the floor moves. Change both and you can produce a number that ranges from reassuring to alarming, from the same balance sheet, on the same day.
This is why the provenance question is not academic. If Blockworks publishes a backing figure with a transparent, reproducible methodology, they have done the ecosystem a service. If they publish a backing figure with an opaque methodology, they have created a new anchor for a narrative that may not survive contact with a liquidation.
There is a specific risk here that I want to name clearly, because it is the one I would flag in any internal review.
If the dashboard publicly displays a "backing per OHM," it may strengthen the market's belief in a hard floor. But if the treasury's assets are illiquid or self-referential, that floor is a spreadsheet artifact. The dashboard would then be doing the opposite of what transparency is supposed to do. It would be converting an uncertain claim into a confident-looking graphic.
Transparency theater. The lights are on. Nobody is home.
Information Presentation vs. Information Verification
This is the core distinction, and I want to make it as sharp as I can.
There are two very different activities that both get called "transparency."
The first is presentation. Taking data that exists and making it readable. Charting it, labeling it, publishing it. Presentation reduces the cost of looking. It does not increase the reliability of what you are looking at.
The second is verification. Independently establishing that a claim is true. Reproducing a number from primary sources. Confirming that a contract holds what it says it holds. Auditing the code that moves the money.
A dashboard is presentation. An audit is verification. A proof-of-reserves attestation is verification. A block explorer is a form of verification, because anyone can re-derive the state from the chain.
The message about the Olympus dashboard describes presentation and implies verification. That gap is the whole story.
And it matters more for Olympus than for most protocols, because Olympus's entire pitch was always about its balance sheet. Other protocols sell throughput. Other protocols sell yield. Olympus sold the treasury. If the treasury is the product, then the credibility of the treasury's reporting is not a nice-to-have. It is the product.
So a dashboard is either a serious instrument or a serious liability. There is very little middle ground when the balance sheet is the thesis.
Three Questions Every Dashboard Must Answer
I have a short checklist I run on any data product before I let it into a decision. Three questions. If a product cannot answer all three, I treat its numbers as promotional.
Question one: is the data source singular or plural? If the dashboard reads from one indexer or one internal database, it inherits that source's errors, omissions, and incentives. If it reads from primary chain state and lets you cross-check against a block explorer, it is more trustworthy.
Question two: are the metric definitions published? "Treasury value" is not a definition. "Treasury value, mark-to-market at the close of the last eight-hour rebase epoch, excluding assets under vesting, valued at the time-weighted price on a named venue" is a definition. The difference is the difference between an argument and a number.
Question three: can a third party reproduce the result? If someone with a node and a query can arrive at the same figure, the dashboard is verification-adjacent. If not, it is a broadcast.
Olympus has a governance structure with multisig execution. Its contracts have historically been upgradeable. These are not accusations. They are the standard risk surface of a DAO of this type, and they are exactly the things a dashboard does not touch. A dashboard reads the state. It does not constrain the admin.
So here is the honest summary. A dashboard that answers all three questions is a real tool. A dashboard that answers none of them is a press release with a user interface. The announcement does not tell us which one shipped.
The Missing Date
There is a structural flaw in the source material that I cannot paper over, and it changes the meaning of everything above.
The message has no publication date.
For most protocols, a missing date is an inconvenience. For Olympus, it is a category error. Olympus is a project whose narrative lifecycle is extreme. The difference between 2021 and 2024 is not a difference of degree. It is a difference of kind.
In 2021, Olympus was at the center of the market's imagination. A treasury dashboard then would have been a major event, read by tens of thousands, capable of moving sentiment and price. The audience was enormous and primed.
In 2022, the reserve-backed DAO sector unraveled. Forks collapsed. The narrative broke. A dashboard then would have been a stabilization attempt, a way to say "the balance sheet is still here" while the market was asking whether the balance sheet meant anything.
After 2023, Olympus settled into a much quieter existence. The audience shrank. The narrative faded. A dashboard then would be a research product for a niche, not a market event.
These three readings imply completely different conclusions about impact, about intent, and about what the announcement is really for. Without a date, I cannot tell you which one you are reading. Neither can anyone else who only has the four information points.
I will go further. A transparency announcement without a date is not incomplete. It is uninterpretable. The date is not metadata. For a project like this, the date is the thesis.
What Blockworks Gets Out of This
I want to be fair to the builder here, because the cynical read is not the only read, and it is not the most likely one.
Blockworks is a media and research operation. Research operations build data products. They build them by picking visible, well-known protocols as showcase cases. Olympus is a perfect showcase case: it is famous, it has a genuinely complex balance sheet, and it has a story that is interesting to explain. A dashboard that explains Olympus is a demonstration of capability. It is a portfolio piece.
That framing changes the incentive analysis. The dashboard is not primarily a favor to Olympus. It is a product for Blockworks, and Olympus is the raw material.
Which is fine. That is how research firms work. But it means the dashboard's design will optimize for clarity, engagement, and narrative coherence, because those are the things that make a data product successful. It will not necessarily optimize for adversarial robustness, because adversarial robustness is not what makes a data product sell.
The failure mode that follows is predictable. When a dashboard is designed to explain, it tends to resolve ambiguity. Ambiguous inputs become clean outputs. Clean outputs become confident narratives. And confident narratives about a treasury are exactly the kind of thing that gets people hurt when the ambiguity was the risk.
There is also the editorial question. If the dashboard is published by a media organization rather than an auditor, its data definitions may pass through an editorial process. Selection bias is not a bug in editorial work. It is the job. But selection bias in a treasury dashboard is a feature you did not ask for.
None of this makes the dashboard bad. It makes it a specific kind of thing. Know which kind you are reading.
The Institutional Trust Claim
Now the claim I find hardest to accept.
The framing says the dashboard may increase institutional trust and investment in Olympus.
I have spent the last two years of my career pitching institutional capital. In 2025 I helped pitch an institutional client on an AI-driven execution system, and the entire conversation was about latency, slippage, custody, and reporting. The dashboard was the last slide, not the first.
Institutions do not buy transparency. They buy verifiability, and they buy it through instruments that a public dashboard cannot substitute for.
Here is what an institution actually needs before it allocates to a treasury-backed token. A legal wrapper that defines what the token represents. Custody arrangements that survive a governance attack. An audit of the contracts that move the money. A reporting standard that matches their internal compliance. Liquidity analysis that tells them how much they can exit and at what cost. Counterparty documentation.
A dashboard touches exactly one of those β reporting β and only partially. It does not create the legal wrapper. It does not change custody. It does not audit the code. It does not create exit liquidity.
So the claim that a dashboard "may increase institutional investment" is not a finding. It is a hope dressed as an inference. The causal chain is missing a dozen links, and the announcement presents the first link as if it were the last.
Institutions do not need to see the number. They need to be able to prove the number to a regulator. Those are different requirements, and only one of them is satisfied by a chart.
Retail vs. Smart Money on Transparency News
This is where the pattern gets expensive, and it is the part I most want readers to internalize.
Transparency news is a specific category, and it has a specific behavior profile. It is structurally positive-sounding. It is hard to falsify. It generates a clean headline. And it almost never functions as a short-term catalyst, because it does not change any cash flow, any supply, or any liquidity.
The retail response to transparency news is to read it as validation. "They are being transparent, therefore they are trustworthy, therefore I am safe." The chain feels logical. It is not. Transparency and trustworthiness are correlated but not identical. A liar can publish a dashboard. A dashboard cannot make a liar honest.
The sophisticated response is to ask what changed. And the answer, almost always, is: nothing changed in the protocol. What changed is the cost of looking. If you were already doing the work, the dashboard saved you time. If you were not doing the work, the dashboard may have given you a false sense of having done it.
The market does not reward the arrival of information it already had. It rewards the arrival of information it did not have. A dashboard that repackages on-chain state is, for anyone with a node, information you already had.
This is why I expect the price impact of this announcement to be small and short-lived, if there is any impact at all. The news is real. The information content is thin. Those two facts coexist comfortably, and the market knows the difference even when the commentary does not.
FOMO is a tax on the unobservant. And the specific unobservant behavior here is mistaking a new way to look for a new thing to see.
The Forks Tell the Story
There is a cleaner way to understand Olympus's position than reading any dashboard, and it is comparative.
The reserve-backed DAO sector was defined by its forks. Wonderland, Klima, Redacted, Hector, and dozens more copied the bonding-and-rebase template within a single cycle. The template was portable. That portability was the tell.
When a design is genuinely defensible, forks struggle. They lack the liquidity, the integrations, the network effects. When a design is primarily a narrative engine, forks thrive, because the narrative is free to copy. The template spreads faster than the value accrues.
Most of those forks are now dead or effectively dormant. That outcome is not a coincidence. It is the market discovering that the mechanism was not the moat. The mechanism was the marketing.
Olympus survived. That survival is meaningful. It means there was something underneath the narrative β a treasury, a community, a governance process, and a willingness to iterate. The project did not simply evaporate like its imitators.
But survival is not scale. Olympus today is not competing with the DeFi blue chips. Lido, Aave, Uniswap β these protocols generate real revenue from real usage at real volume. Olympus's value capture is different in kind. It rests on treasury backing and on the demand to borrow against it, not on fee flows.
That distinction determines everything about how to read a treasury dashboard. For a fee-generating protocol, a dashboard is a nice addition. For a treasury-backed protocol, the dashboard is close to the entire external interface. The balance sheet is the product, so the reporting of the balance sheet is the marketing, the diligence, and the risk disclosure, all in one artifact.
That is a lot of weight to put on a chart.
Cooler, gOHM, and the Parts a Dashboard Cannot Show
Olympus iterated past the original design. The V2 era introduced gOHM, a wrapped, non-rebasing representation that made the token easier to use across chains and in DeFi integrations. Later, the project added lending facilities, most notably the Cooler-style loans, where the token could be used as collateral against the treasury on defined terms.
These are the parts of the architecture that matter most, and they are precisely the parts a treasury dashboard is least equipped to illuminate.
A lending facility creates a second-order balance sheet. When holders borrow against their position, the protocol's health depends not just on what the treasury holds but on what has been borrowed against it and on the terms of repayment. A dashboard can show the treasury. It is much harder to show the reflexive risk of the borrowing layer in a way that a reader will correctly interpret.
The gOHM wrapper changes the supply accounting. Rebase mechanics interact with wrapped supply in ways that are easy to misstate. If the dashboard reports a supply figure without specifying whether it is OHM, gOHM, or an aggregate, the backing-per-unit calculation is wrong before it starts.
I have seen this class of error repeatedly in my own work. When my team built a mean-reversion strategy for Layer 2 tokens in 2024, the hardest part was never the signal. It was the accounting. Token supply across bridges, wrapped representations, and vesting schedules is a minefield, and a single wrong assumption in the denominator will produce a confident, well-formatted, completely wrong result.
Data products fail at the accounting layer far more often than at the visualization layer. The chart is never the problem. The definition behind the chart is the problem.
The Liquidity Question
The dimension I would weight most heavily, and the one entirely absent from the announcement, is liquidity.
A treasury has a value. A position has an exit. These are different numbers, and the gap between them is where most losses live.
For a token like OHM, the relevant questions are not about the treasury's headline figure. They are about depth. How much can be sold at the current price before the price moves? What is the slippage on a position of meaningful size? How concentrated is the liquidity across venues? What happens to the market maker inventory if sentiment turns?
These questions have answers. The answers live in order books and pool reserves, not in a treasury dashboard. And the answers change fast. A liquidity profile that looks adequate in a quiet market can evaporate in hours when everyone tries to leave through the same door.
I care about this more than most, because my first real loss came from exactly this misunderstanding. I priced a trade against reserves and got filled against reality. The reserves were on the screen. The fill was in my P&L.
So when a transparency product is announced for a treasury-backed token, my instinct is to ask about the exit before I ask about the entry. What does the liquidity look like? Who is providing it? What happens to it under stress? A dashboard that answers those questions would be genuinely valuable. A dashboard that answers the treasury question while leaving the liquidity question untouched has solved the easier half of the problem.
Charts lie. Liquidity speaks. The chart is now available. I am still waiting to hear the liquidity.
What I'd Check Before Sizing Anything
Concrete. This is what I would actually do if I were evaluating a position in a treasury-backed token after a transparency announcement. No theory. Just steps.
Pull the contract addresses and confirm the dashboard's numbers against primary chain state. If I cannot reproduce a headline figure from a node and a query, I treat the dashboard as unverified and I discount it accordingly.
Read the metric definitions in full. Not the labels. The definitions. What counts as treasury? How are POL positions valued? What price source, what timestamp, what treatment of vesting and locked supply?
Check the supply accounting. OHM versus gOHM. Circulating versus total. Rebasing versus wrapped. Confirm the denominator before trusting any per-unit figure.
Map the admin surface. Who holds the multisig? Is the contract upgradeable? What is the timelock? A dashboard that shows me a healthy treasury attached to an unconstrained admin key has shown me a healthy treasury that could be redirected tomorrow.
Measure exit liquidity directly. Depth at the current price, slippage at my intended size, venue concentration, and what the order book looks like at minus twenty and minus forty percent. I want to know the shape of the downside before I consider the upside.
Finally, I check the date. Always the date. Because the meaning of every number above depends on when I am reading it, and a message without a date is a message I cannot fully use.
This is not glamorous work. It is the work. And the reason a dashboard is valuable is that it can make this work faster β not that it can replace it.
Scenarios, Not Forecasts
I do not publish price targets. I publish conditional structures, because the market does not pay you for being right about direction. It pays you for being positioned for the right scenario.
Here is how I would frame the Olympus dashboard across three cases.
If the dashboard is genuinely reproducible, with open data sources and published definitions, it is a modest, durable positive. It lowers the cost of diligence for sophisticated participants and it creates a public record that makes future claims checkable. The impact on price is small. The impact on the quality of the market's information is real. Over time, better information attracts better capital.
If the dashboard is a closed product with opaque definitions, it is a marketing surface. It will generate a headline, a small sentiment bump, and no durable change. In that case, the correct response is to ignore the dashboard and read the chain.
If the dashboard displays a confident backing figure built on self-referential or illiquid assets, it is worse than nothing. It manufactures a floor that cannot be defended and invites capital into a position it cannot exit. That is the scenario that should worry anyone holding size.
I cannot tell you which scenario is live, because the announcement does not contain the information required to distinguish them. That absence is itself the finding. A transparency product that does not disclose its methodology has not delivered transparency. It has delivered the appearance of it.
Takeaway
The dashboard is a window. It is not a witness. Windows let you see. They do not testify.
Blockworks has built something that may genuinely help people understand one of the more complicated balance sheets in DeFi. That is worth something. I am not dismissing it. I am locating it. It is a presentation layer over a set of numbers that were already there, and its value depends entirely on whether those numbers can be independently reproduced.
The part of the announcement that should raise an eyebrow is not the dashboard. It is the claim that a dashboard can create institutional trust. Institutions do not buy windows. They buy attestations, wrappers, custody, audits, and liquidity. A public chart is none of those things.
And the part that should raise both eyebrows is the missing date. A story about a project whose meaning changed completely across three distinct eras, delivered without a timestamp, is not a story. It is a fragment.
The broader lesson is the one I keep relearning. Every cycle produces a new instrument for making information look settled. The information is rarely settled. The instrument is just well lit.
So here is the question I would leave with anyone holding a treasury-backed token after reading this news. When the exit gets crowded and the floor gets tested and the dashboard is still drawing a clean line β what will you actually be able to sell, and to whom, and at what price?
That number is not on the dashboard. That number is the trade.
Appendix: Reading the Source Honestly
The source material behind this analysis contained four information points. One described the dashboard's existence. Two were opinions about its effects. One was a source note. Nothing in it addressed technical specifications, token data, team, funding, regulation, or market metrics.
I want to be explicit about that, because the discipline of separating fact from inference is the discipline that keeps traders solvent. In this article, I have marked what is claimed, what is known, and what is inferred. Where I have supplied background on Olympus β its history, its mechanisms, its forks, its iterations β that background comes from my own reading of the protocol's public record and should be verified independently before it informs a position.
Where I have supplied judgment β about transparency theater, about institutional requirements, about liquidity β that judgment comes from my own experience running strategies and pitching capital. It is not a fact about Olympus. It is a frame for evaluating it.
A reader who wants the dashboard's methodology should not read this article. They should read the dashboard. That is the whole point. The announcement gave us a window and asked us to trust the view. The work is to check the glass.
Trust the data, ignore the discord. But only after you have confirmed the data is data, and not a story someone decided to draw.
The window is open. Someone should look through it properly. And then someone should walk around to the other side of the wall and check what is holding it up.