Why Data Lakes Failed but Company Brains Won't

VISPAICO Journal — Issue 004

---

In 1799, French soldiers rebuilding a fort near the town of Rosetta uncovered a slab of black granite covered in three scripts. The stone was, in the most literal sense, information — dense, complete, and perfectly preserved. It sat in a museum for two decades, fully intact, entirely unread. Scholars could see every character clearly. What they lacked was the relationship between the scripts: the understanding that the same decree had been written three times, in three languages, and that one of those languages was already known.

It took Jean-François Champollion until 1822 to make that connection, and the moment he did, an entire civilisation's writing became legible again — not because more stone was found, but because someone finally understood how the pieces related to each other.

This is worth sitting with, because it describes almost exactly what has gone wrong inside a generation of companies that spent enormous sums accumulating information and remain, despite all of it, unable to answer their own questions quickly. They have the stone. They have never quite had their Champollion.

---

The Warehouse Without a Floor Plan

Picture a vast warehouse, built to hold everything a company has ever produced: every contract, every customer email, every meeting note, every spreadsheet, every slide deck from every strategy offsite for the past fifteen years. Nothing has been thrown away. Every item that entered the building is still there, somewhere, preserved.

Now imagine walking into that warehouse and being told to find the reasoning behind a pricing decision made three years ago. Technically, the answer exists inside the building. Practically, finding it requires knowing which aisle, which shelf, which box — and the person who could have told you that retired last spring. So you wander. You open boxes at random. Eventually, if you're persistent, you find something close to what you needed, though you can no longer be entirely sure it's the final version, or whether a more recent box somewhere else supersedes it.

This is not a hypothetical. It is a fair description of how most large organisations actually experience their own accumulated information. They have not failed to collect it. They have failed to organise the relationships between what they collected — which is a different problem entirely, and a far harder one.

---

The Library With No Catalogue

Jorge Luis Borges once imagined a library containing every possible book — every combination of letters that could ever be arranged, including, somewhere within its infinite shelves, the precise and complete truth about anything one might wish to know. The library was, in principle, complete. It was also, for all practical purposes, useless, because it contained no way of finding the one true book among the infinite volumes of nonsense surrounding it. Completeness without navigation is not a library. It is noise with the appearance of order.

Real libraries solved this problem centuries before anyone thought to build a warehouse of digital files. A library is not valuable because it holds books. It is valuable because of the catalogue — the quiet, unglamorous system that tells a visitor precisely where a specific idea sits, how it relates to adjacent ideas, and which edition supersedes which. Remove the catalogue and a library becomes a building full of paper, no more useful than the warehouse with no floor plan. The books haven't changed. The value has evaporated entirely, because value was never really in the paper. It was in the relationships the catalogue made visible.

Most corporate data initiatives of the past two decades built extraordinary buildings and never quite finished the catalogue. They solved the problem of storage — genuinely, impressively — while leaving the harder problem, the problem of relationship, essentially untouched. Executives were told, reasonably enough, that once the information was centralised, understanding would follow naturally. It rarely did. Storage and understanding turned out to be different disciplines entirely, and mastering one gave no particular advantage in the other.

---

Why Storing Is Not the Same as Knowing

There is a useful distinction, borrowed loosely from biology, between a genome and an organism. A genome is a complete, precise record of every instruction required to build a living thing. It is, in a real sense, the most compressed and total information a creature possesses. But a genome sitting in a sequencing machine is not alive. It has no metabolism, no behaviour, no way of responding to its environment. Life exists in the relationships between genes — in how they switch on and off in sequence, in how they respond to each other and to signals from outside — not in the raw sequence itself.

A company's stored information behaves the same way. A customer record is not knowledge. A contract is not knowledge. An email thread from a negotiation two years ago is not knowledge. Each is a fragment, precise but inert, until it is understood in relation to everything else the company knows: this customer's pricing history, that supplier's past reliability, the reason a similar deal fell through the previous year. Knowledge lives in the connective tissue between fragments, not in the fragments themselves — and connective tissue is exactly what most large-scale data efforts were never designed to build.

This is why so many companies can report, with genuine pride, that they have "all their data in one place," while simultaneously admitting that finding a straightforward answer to an everyday question — why did we lose this client, what did we agree with this supplier, has this exact problem happened before — still requires tracking down whichever employee happens to remember. The data is centralised. The understanding never was.

---

The Cartographer's Advantage

Before satellite imagery, mapmaking was less about collecting terrain and more about deciding what relationships mattered enough to draw. A skilled cartographer did not attempt to record every rock and tree; that approach produces something closer to Borges's library than to a usable map. Instead, the cartographer decided which features related to which — which river fed which town, which mountain pass connected which valleys — and represented those relationships clearly enough that a traveller, having never set foot in the territory before, could still navigate it with confidence.

The map's power was never in how much terrain it depicted. It was in how well it captured relationship: this connects to that, this leads toward that, this matters more than the ten features left off entirely. A map that tried to include everything would be unreadable. A map that captured relationships precisely could guide someone through country they had never seen.

Companies today possess an extraordinary amount of terrain and almost no maps. They can tell you, with reasonable precision, how many customer records exist in a given system. They struggle badly to tell a new hire, in under an hour, how the business actually works — which decisions depend on which others, which relationships with which clients carry hidden risk, which past initiatives quietly failed and why, so the same initiative doesn't get proposed again next quarter with genuine enthusiasm and no memory of the first attempt.

---

What a Company Brain Actually Does Differently

The idea of a Company Brain is easy to misunderstand as simply a better warehouse — faster storage, cleverer search, a more powerful engine for retrieving the same undifferentiated files. That misses the point almost entirely. The distinguishing feature is not storage capacity. It is the capacity to hold relationship: to understand that this contract connects to that client history, which connects to that pricing exception, which connects to the reason a particular account manager always handles this customer personally rather than routing them through the standard process.

This is closer to Champollion's insight than to a larger warehouse. The Rosetta Stone did not need more information to become legible. Every character was already there, fully preserved, for two decades before anyone read it correctly. What it needed was the relationship between what was already present — the recognition that this script and that script were saying the same thing, and that recognising the connection would unlock everything else.

A Company Brain, properly built, behaves the same way inside a business. It does not primarily add new information to what a company already has. It makes visible the relationships between information that already exists but has never been connected — the pricing decision and the client history behind it, the current operational exception and the three similar exceptions from years past, the new hire's question and the specific, particular answer this specific business has always given, for reasons that made sense at the time and mostly still do.

---

The Uncomfortable Question for Every Executive

The uncomfortable implication, for any leader who has invested seriously in centralising company data over the past decade, is that storage was never really the bottleneck. It felt like the bottleneck, because it was the part that could be measured, budgeted, and reported on with confidence at a board meeting. Terabytes are a satisfying metric. Relationship is not, which is precisely why it has been so persistently underinvested in, even by companies that spent generously on everything adjacent to it.

The right question for any executive to ask now is not "where is our data?" Most companies, whatever their frustrations, already know roughly where their data lives. The right question is closer to the one Champollion effectively asked in front of that stone: what connects to what, and why has nobody yet made that connection visible? It is a harder question, and a more valuable one, because the answer cannot be bought as easily as storage can. It has to be built, deliberately, out of the relationships a business already contains but has never had reason, or means, to make explicit.

---

The Real Lesson of the Stone

The Rosetta Stone sat in plain view for twenty-three years, read by dozens of scholars, all of whom saw exactly the same characters Champollion eventually decoded. What separated understanding from mere possession was never access to more information. It was the recognition of how the pieces already in hand related to one another — a recognition that, once made, could never again be unmade, and that changed, permanently, an entire field's ability to read a civilisation that information alone had failed to unlock.

Companies today are not short of stone. Most have accumulated more of it than any generation of executives before them ever had access to, spread across systems, folders, and archives that nobody fully remembers building. What they are short of is the relationship between the pieces — the quiet, patient work of understanding how one fragment of the business connects to another, and making that connection available to anyone who needs it, not only to the handful of people who happened to be present when the connection was first made.

The businesses that grasp this first will not be remembered as the companies with the most data. They will be remembered, quietly and probably not until much later, as the ones who finally understood what they already had.