What Your Company Chooses to Forget
Neuroscientists call it active forgetting. While you sleep, your brain prunes memory traces on purpose — a targeted, mechanistic process with a name and a body of research behind it. LinkedIn calls the same instinct “radical prioritization.” Only one of these can point to a mechanism.
Data
Management
“We're a data-driven organization,” the slogan goes. So everything gets saved. Every version, every thread, every meeting gets a transcript, and nothing is ever deleted, on principle.
![](data:image/svg+xml,<svg display="block" id="565267359" role="presentation" viewBox="0 0 3000 1376" xmlns="http://www.w3.org/2000/svg"/>)
FORGETTING IS A FEATURE
Hyperthymesia is a neurological condition that lets a person recall an extraordinary share of their own life in vivid, dated, near-total detail. Researchers at UC Irvine first documented it in 2006; fewer than 100 people have been formally diagnosed since.
It certainly sounds like a superpower.
But, ask the people who actually have it, and the story flips. Solomon Shereshevsky, the Russian mnemonist studied by Alexander Luria for decades, could recall entire pages verbatim years later. But he struggled his whole life to hold a job, follow a plot, or read a sentence for its meaning instead of drowning in the sound of every word.
Total recall didn't sharpen his mind. Nothing was ever filtered, so nothing stood out, and everything competed for attention at once, forever.
The brain deletes memory traces on purpose, every night, in a process researchers call active forgetting. This controlled mechanism prunes synaptic connections the same way it strengthens the ones you'll actually use again. In folk psychology, forgetting is what happens when memory fails. But in reality, forgetting is what happens when memory keeps working. To recall a fact, your brain needs to actively suppress the competing, adjacent facts that would otherwise blur it. Sharpness, focus — these depend on the brain's radical deprioritization of noise.
And yet, forgetting is scary. We associate it with loss. And there's a subtle kind of comfort one gets from subclinical hoarding (you just don't know when x would come in handy). This individual attitude carries over to organizational settings. A company needs to deliberately scaffold an active forgetting mechanism to keep its focus. Otherwise, it will run on Shereshevsky's brain: able to produce anything on demand, unable to tell you what any of it means.
(We've argued elsewhere that when a company's tools genuinely extend a shared Mind, that's the whole point — worth reaching for. This is the failure mode running in the other direction: not too little extension, but too little pruning of what already got extended.)
There's a job whose entire function is deciding what an archive should discard.
UNASSIGNED RIGOR
Archivists call it appraisal, and it's the actual discipline behind every functioning institutional memory, from a national library to a hospital's patient records. The archival theorist Theodore Schellenberg gave the field its still-standard test back in the 1950s: does a record hold primary value — the reason it was created, a contract, a decision, a transaction — or only secondary value, the kind historians or auditors might want later? Primary value expires. Most of it, fast. Only a fraction of anything a company produces was ever going to be worth keeping past the reason it got made.
And we're ready to bet that role doesn't exist in your company. IT owns the storage. Legal owns the exceptions. No one owns pruning, so the mind never gets sharpened.
“We're a data-driven organization” was supposed to mean something like Schellenberg's test, applied constantly: which of this actually informs a decision, and which of it just accumulated because deleting felt riskier than keeping.
Instead, it became the opposite — a slogan that makes indiscriminate hoarding sound like discipline. Being data-driven and being unable to throw anything away are not the same skill. One requires judgment. The other just increases your storage bill.
This is a fairly new kind of conundrum.
STORAGE GOT CHEAP
In 1956, IBM would rent you the first commercial disk drive — five megabytes of it — inside a system that ran $3,200 a month, about $36,700 in today's money. That same monthly bill would rent you roughly a petabyte and a half on S3 now: three hundred million times the data, for the same line item. Back then, keeping something was a recurring cost somebody had to defend every single year. As storage got cheaper, the immediate need to make that decision disappeared, regardless of the long-term consequences of letting it slip.
(The same logic that says whatever gets cheap gets big — cheaper photographs mean vastly more photographs, cheaper letters mean vastly more email — runs just as reliably on the way in. Cheap keeping floods the archive the same way cheap making floods the feed.)
The volume of data a typical company stores climbs every quarter, dependably, on a chart that only goes one direction. Set next to it, the share of that data anyone actually opens in a given year stays close to flat, and close to zero. The gap between those two lines is a type of cost that never makes it in front of the board.
Slower search, bogged-down processes, longer onboarding, security exposure. Blurred vision. Judgment used to be free, bundled into the cost of storage. But then storage got cheap, and we (ironically) forgot to substitute for the vital function of active forgetting.
What a deletion policy actually looks like.
THE FOUR PARTS
Not “let's clean up the drive” as a one-off January ritual, and not a blanket ban that panics everyone into keeping more out of fear of guessing wrong. A real policy has four moving parts, and you can audit your own company against all four in an afternoon.
A retention schedule, by category — not “documents” as one bucket, but contracts, customer records, internal drafts, meeting notes, and system logs, each carrying its own answer to how long it stays useful and to whom. Legal has probably already set this for the categories it cares about. But what about the rest?
A named owner — a person or a function whose job includes saying no to keeping something, not just yes to storing it. If deletion isn't anyone's job, it isn't going to happen, no matter how many slides mention “data hygiene” going forward.
A cadence — appraisal on a schedule, not triggered by the drive filling up. Quarterly for anything that moves fast, annually for anything that doesn't.
Solid criteria — usage, relevance, and risk, asked in that order, by someone whose job it is to ask, on a schedule that doesn't depend on anyone remembering to.
Some of what your company keeps really is untouchable, and we have to grant that to stay honest.
BEING HONEST
Tax records, employment files, anything under active litigation or regulatory hold, certain financial data by law. These are, of course, not judgment calls. The schedule simply says so and moves on.
But, the seven-year-old export nobody's opened? The seventeen versions of a deck? The entire drive of a project that shipped and closed two years ago? Ask why any specific file is still there and “we might need it someday” arrives immediately, confidently, and almost never followed by for what. That instinct is there to justify a decision nobody wants to make.
Forgetting is just what makes remembering possible.
THE MOVE
When the “someday” your company needs to remember something inevitably comes, what allows a proper response is a focused retention strategy, rather than monolithic raw data.
Your brain runs this discipline every night without asking permission. Archivists built a whole profession around running it for institutions. Your company, in all likelihood, has assigned it to no one.
“Data-driven” was supposed to describe an organization that knows the difference between a record worth keeping and one worth releasing. Right now, for most companies, it describes an organization that's simply not brave enough to press delete.
The businesses that get this right won't be the ones with the biggest archive. They'll be the ones who can tell you, for anything in it, why it's still there.
Ask your team who owns deletion. If the answer takes more than a second, you'll probably find use in a conversation with us.
Featured Case Studies
Featured Case Studies
What Your Company Chooses to Forget
Neuroscientists call it active forgetting. While you sleep, your brain prunes memory traces on purpose — a targeted, mechanistic process with a name and a body of research behind it. LinkedIn calls the same instinct “radical prioritization.” Only one of these can point to a mechanism.
Data
Management
“We're a data-driven organization,” the slogan goes. So everything gets saved. Every version, every thread, every meeting gets a transcript, and nothing is ever deleted, on principle.
FORGETTING IS A FEATURE
Hyperthymesia is a neurological condition that lets a person recall an extraordinary share of their own life in vivid, dated, near-total detail. Researchers at UC Irvine first documented it in 2006; fewer than 100 people have been formally diagnosed since.
It certainly sounds like a superpower.
But, ask the people who actually have it, and the story flips. Solomon Shereshevsky, the Russian mnemonist studied by Alexander Luria for decades, could recall entire pages verbatim years later. But he struggled his whole life to hold a job, follow a plot, or read a sentence for its meaning instead of drowning in the sound of every word.
Total recall didn't sharpen his mind. Nothing was ever filtered, so nothing stood out, and everything competed for attention at once, forever.
The brain deletes memory traces on purpose, every night, in a process researchers call active forgetting. This controlled mechanism prunes synaptic connections the same way it strengthens the ones you'll actually use again. In folk psychology, forgetting is what happens when memory fails. But in reality, forgetting is what happens when memory keeps working. To recall a fact, your brain needs to actively suppress the competing, adjacent facts that would otherwise blur it. Sharpness, focus — these depend on the brain's radical deprioritization of noise.
And yet, forgetting is scary. We associate it with loss. And there's a subtle kind of comfort one gets from subclinical hoarding (you just don't know when x would come in handy). This individual attitude carries over to organizational settings. A company needs to deliberately scaffold an active forgetting mechanism to keep its focus. Otherwise, it will run on Shereshevsky's brain: able to produce anything on demand, unable to tell you what any of it means.
(We've argued elsewhere that when a company's tools genuinely extend a shared Mind, that's the whole point — worth reaching for. This is the failure mode running in the other direction: not too little extension, but too little pruning of what already got extended.)
There's a job whose entire function is deciding what an archive should discard.
UNASSIGNED RIGOR
Archivists call it appraisal, and it's the actual discipline behind every functioning institutional memory, from a national library to a hospital's patient records. The archival theorist Theodore Schellenberg gave the field its still-standard test back in the 1950s: does a record hold primary value — the reason it was created, a contract, a decision, a transaction — or only secondary value, the kind historians or auditors might want later? Primary value expires. Most of it, fast. Only a fraction of anything a company produces was ever going to be worth keeping past the reason it got made.
And we're ready to bet that role doesn't exist in your company. IT owns the storage. Legal owns the exceptions. No one owns pruning, so the mind never gets sharpened.
“We're a data-driven organization” was supposed to mean something like Schellenberg's test, applied constantly: which of this actually informs a decision, and which of it just accumulated because deleting felt riskier than keeping.
Instead, it became the opposite — a slogan that makes indiscriminate hoarding sound like discipline. Being data-driven and being unable to throw anything away are not the same skill. One requires judgment. The other just increases your storage bill.
This is a fairly new kind of conundrum.
STORAGE GOT CHEAP
In 1956, IBM would rent you the first commercial disk drive — five megabytes of it — inside a system that ran $3,200 a month, about $36,700 in today's money. That same monthly bill would rent you roughly a petabyte and a half on S3 now: three hundred million times the data, for the same line item. Back then, keeping something was a recurring cost somebody had to defend every single year. As storage got cheaper, the immediate need to make that decision disappeared, regardless of the long-term consequences of letting it slip.
(The same logic that says whatever gets cheap gets big — cheaper photographs mean vastly more photographs, cheaper letters mean vastly more email — runs just as reliably on the way in. Cheap keeping floods the archive the same way cheap making floods the feed.)
The volume of data a typical company stores climbs every quarter, dependably, on a chart that only goes one direction. Set next to it, the share of that data anyone actually opens in a given year stays close to flat, and close to zero. The gap between those two lines is a type of cost that never makes it in front of the board.
Slower search, bogged-down processes, longer onboarding, security exposure. Blurred vision. Judgment used to be free, bundled into the cost of storage. But then storage got cheap, and we (ironically) forgot to substitute for the vital function of active forgetting.
What a deletion policy actually looks like.
THE FOUR PARTS
Not “let's clean up the drive” as a one-off January ritual, and not a blanket ban that panics everyone into keeping more out of fear of guessing wrong. A real policy has four moving parts, and you can audit your own company against all four in an afternoon.
A retention schedule, by category — not “documents” as one bucket, but contracts, customer records, internal drafts, meeting notes, and system logs, each carrying its own answer to how long it stays useful and to whom. Legal has probably already set this for the categories it cares about. But what about the rest?
A named owner — a person or a function whose job includes saying no to keeping something, not just yes to storing it. If deletion isn't anyone's job, it isn't going to happen, no matter how many slides mention “data hygiene” going forward.
A cadence — appraisal on a schedule, not triggered by the drive filling up. Quarterly for anything that moves fast, annually for anything that doesn't.
Solid criteria — usage, relevance, and risk, asked in that order, by someone whose job it is to ask, on a schedule that doesn't depend on anyone remembering to.
Some of what your company keeps really is untouchable, and we have to grant that to stay honest.
BEING HONEST
Tax records, employment files, anything under active litigation or regulatory hold, certain financial data by law. These are, of course, not judgment calls. The schedule simply says so and moves on.
But, the seven-year-old export nobody's opened? The seventeen versions of a deck? The entire drive of a project that shipped and closed two years ago? Ask why any specific file is still there and “we might need it someday” arrives immediately, confidently, and almost never followed by for what. That instinct is there to justify a decision nobody wants to make.
Forgetting is just what makes remembering possible.
THE MOVE
When the “someday” your company needs to remember something inevitably comes, what allows a proper response is a focused retention strategy, rather than monolithic raw data.
Your brain runs this discipline every night without asking permission. Archivists built a whole profession around running it for institutions. Your company, in all likelihood, has assigned it to no one.
“Data-driven” was supposed to describe an organization that knows the difference between a record worth keeping and one worth releasing. Right now, for most companies, it describes an organization that's simply not brave enough to press delete.
The businesses that get this right won't be the ones with the biggest archive. They'll be the ones who can tell you, for anything in it, why it's still there.
Ask your team who owns deletion. If the answer takes more than a second, you'll probably find use in a conversation with us.
Featured Case Studies
Featured Case Studies
What Your Company Chooses to Forget
Neuroscientists call it active forgetting. While you sleep, your brain prunes memory traces on purpose — a targeted, mechanistic process with a name and a body of research behind it. LinkedIn calls the same instinct “radical prioritization.” Only one of these can point to a mechanism.
Data
Management
“We're a data-driven organization,” the slogan goes. So everything gets saved. Every version, every thread, every meeting gets a transcript, and nothing is ever deleted, on principle.
FORGETTING IS A FEATURE
Hyperthymesia is a neurological condition that lets a person recall an extraordinary share of their own life in vivid, dated, near-total detail. Researchers at UC Irvine first documented it in 2006; fewer than 100 people have been formally diagnosed since.
It certainly sounds like a superpower.
But, ask the people who actually have it, and the story flips. Solomon Shereshevsky, the Russian mnemonist studied by Alexander Luria for decades, could recall entire pages verbatim years later. But he struggled his whole life to hold a job, follow a plot, or read a sentence for its meaning instead of drowning in the sound of every word.
Total recall didn't sharpen his mind. Nothing was ever filtered, so nothing stood out, and everything competed for attention at once, forever.
The brain deletes memory traces on purpose, every night, in a process researchers call active forgetting. This controlled mechanism prunes synaptic connections the same way it strengthens the ones you'll actually use again. In folk psychology, forgetting is what happens when memory fails. But in reality, forgetting is what happens when memory keeps working. To recall a fact, your brain needs to actively suppress the competing, adjacent facts that would otherwise blur it. Sharpness, focus — these depend on the brain's radical deprioritization of noise.
And yet, forgetting is scary. We associate it with loss. And there's a subtle kind of comfort one gets from subclinical hoarding (you just don't know when x would come in handy). This individual attitude carries over to organizational settings. A company needs to deliberately scaffold an active forgetting mechanism to keep its focus. Otherwise, it will run on Shereshevsky's brain: able to produce anything on demand, unable to tell you what any of it means.
(We've argued elsewhere that when a company's tools genuinely extend a shared Mind, that's the whole point — worth reaching for. This is the failure mode running in the other direction: not too little extension, but too little pruning of what already got extended.)
There's a job whose entire function is deciding what an archive should discard.
UNASSIGNED RIGOR
Archivists call it appraisal, and it's the actual discipline behind every functioning institutional memory, from a national library to a hospital's patient records. The archival theorist Theodore Schellenberg gave the field its still-standard test back in the 1950s: does a record hold primary value — the reason it was created, a contract, a decision, a transaction — or only secondary value, the kind historians or auditors might want later? Primary value expires. Most of it, fast. Only a fraction of anything a company produces was ever going to be worth keeping past the reason it got made.
And we're ready to bet that role doesn't exist in your company. IT owns the storage. Legal owns the exceptions. No one owns pruning, so the mind never gets sharpened.
“We're a data-driven organization” was supposed to mean something like Schellenberg's test, applied constantly: which of this actually informs a decision, and which of it just accumulated because deleting felt riskier than keeping.
Instead, it became the opposite — a slogan that makes indiscriminate hoarding sound like discipline. Being data-driven and being unable to throw anything away are not the same skill. One requires judgment. The other just increases your storage bill.
This is a fairly new kind of conundrum.
STORAGE GOT CHEAP
In 1956, IBM would rent you the first commercial disk drive — five megabytes of it — inside a system that ran $3,200 a month, about $36,700 in today's money. That same monthly bill would rent you roughly a petabyte and a half on S3 now: three hundred million times the data, for the same line item. Back then, keeping something was a recurring cost somebody had to defend every single year. As storage got cheaper, the immediate need to make that decision disappeared, regardless of the long-term consequences of letting it slip.
(The same logic that says whatever gets cheap gets big — cheaper photographs mean vastly more photographs, cheaper letters mean vastly more email — runs just as reliably on the way in. Cheap keeping floods the archive the same way cheap making floods the feed.)
The volume of data a typical company stores climbs every quarter, dependably, on a chart that only goes one direction. Set next to it, the share of that data anyone actually opens in a given year stays close to flat, and close to zero. The gap between those two lines is a type of cost that never makes it in front of the board.
Slower search, bogged-down processes, longer onboarding, security exposure. Blurred vision. Judgment used to be free, bundled into the cost of storage. But then storage got cheap, and we (ironically) forgot to substitute for the vital function of active forgetting.
What a deletion policy actually looks like.
THE FOUR PARTS
Not “let's clean up the drive” as a one-off January ritual, and not a blanket ban that panics everyone into keeping more out of fear of guessing wrong. A real policy has four moving parts, and you can audit your own company against all four in an afternoon.
A retention schedule, by category — not “documents” as one bucket, but contracts, customer records, internal drafts, meeting notes, and system logs, each carrying its own answer to how long it stays useful and to whom. Legal has probably already set this for the categories it cares about. But what about the rest?
A named owner — a person or a function whose job includes saying no to keeping something, not just yes to storing it. If deletion isn't anyone's job, it isn't going to happen, no matter how many slides mention “data hygiene” going forward.
A cadence — appraisal on a schedule, not triggered by the drive filling up. Quarterly for anything that moves fast, annually for anything that doesn't.
Solid criteria — usage, relevance, and risk, asked in that order, by someone whose job it is to ask, on a schedule that doesn't depend on anyone remembering to.
Some of what your company keeps really is untouchable, and we have to grant that to stay honest.
BEING HONEST
Tax records, employment files, anything under active litigation or regulatory hold, certain financial data by law. These are, of course, not judgment calls. The schedule simply says so and moves on.
But, the seven-year-old export nobody's opened? The seventeen versions of a deck? The entire drive of a project that shipped and closed two years ago? Ask why any specific file is still there and “we might need it someday” arrives immediately, confidently, and almost never followed by for what. That instinct is there to justify a decision nobody wants to make.
Forgetting is just what makes remembering possible.
THE MOVE
When the “someday” your company needs to remember something inevitably comes, what allows a proper response is a focused retention strategy, rather than monolithic raw data.
Your brain runs this discipline every night without asking permission. Archivists built a whole profession around running it for institutions. Your company, in all likelihood, has assigned it to no one.
“Data-driven” was supposed to describe an organization that knows the difference between a record worth keeping and one worth releasing. Right now, for most companies, it describes an organization that's simply not brave enough to press delete.
The businesses that get this right won't be the ones with the biggest archive. They'll be the ones who can tell you, for anything in it, why it's still there.
Ask your team who owns deletion. If the answer takes more than a second, you'll probably find use in a conversation with us.
Featured Case Studies
Featured Case Studies