The mental bias of not wanting to delete data, even when it's no longer needed
We all know that small resistance when faced with throwing something away. An old file, a forgotten folder, a 2018 presentation, an export that “might come in handy someday.” On a personal level, it can raise a smile. At the corporate level, however, this habit becomes a governance, cost, security, and productivity problem.
Data decluttering starts right here: not with technology, but with the relationship people have with data. Companies accumulate information for technical, operational, and regulatory reasons — but also for a more human one: deleting feels like a loss. Keeping everything feels prudent. Even when it isn’t.
The Human Reflex of "Just in Case"
In data management, one of the most dangerous phrases is also one of the most common: “let’s keep it, just in case.” It seems like a reasonable choice, because nobody wants to delete something that might prove useful. But if this logic is applied to every file, every database, every report, and every legacy system, the result is an increasingly heavy information ecosystem.
The Keanu Reeves video on YouTube Shorts can be a fun hook to illustrate the point: we all have a bit of an attachment to the things we’ve accumulated. The problem isn’t the attachment itself. The problem arises when this reflex enters business processes and turns obsolete, duplicated, or irrelevant data into something nobody dares touch anymore.
In the enterprise context, “just in case” becomes storage, backups, cloud costs, legacy repositories, privacy risk, and decision-making confusion.
Status quo bias: why we prefer to leave everything as it is
Status quo bias is the tendency to prefer that things stay as they are, even when changing would bring benefits. Samuelson and Zeckhauser described it as far back as 1988 as a disproportionate preference for the current option. Subsequent studies have confirmed its relevance across many real-world decisions.
Applied to data, this bias is very concrete. If an archive exists, it stays. If a folder has always been there, nobody questions it. If a legacy system is still running, it’s perceived as necessary. The choice not to decide becomes a decision: keep everything.
The short answer is: many companies don’t delete useless data not because it’s actually needed, but because the current state feels less risky than change. Yet when it comes to data, inaction also has a cost.
This also undermines compliance, but the deeper problem is coordination. Compliance cannot be managed centrally if business functions do not disclose what they are doing. Legal teams cannot review tools they do not know exist. Data governance teams cannot protect data flows they cannot see.
The short answer is: a group-level written policy is not enough if every department interprets AI use differently. The organisation needs a shared operating model that tells teams how to make AI decisions, not just what the final policy says.
Loss Aversion: the fear of losing something outweighs the benefit
Loss aversion describes the tendency to feel the weight of a loss more intensely than the pleasure of an equivalent gain. In data decluttering, this translates into a very clear fear: “what if we delete something we’ll need later?”
This fear is understandable. No organization wants to lose useful data, documentary evidence, a historical baseline, or information needed for an audit. The point, however, is that the fear of loss is often factored in, while the costs of retention are ignored.
Keeping useless data can increase the attack surface, complicate compliance, slow down decisions, drive up cloud costs, and keep legacy systems alive. The imagined loss is visible. The real cost of accumulation remains diffuse, and therefore less perceived.
Endowment effect: we place more value on what we already own
The endowment effect is the bias by which we tend to assign greater value to something simply because it’s already ours. In the physical world, it’s easy to understand: an object we own seems more valuable than it would to an outside observer.
In the world of data, something similar happens. An old dataset, a historical archive, or a system built years ago can seem important because it’s part of the organization’s history. “We’ve always had it” becomes almost synonymous with “it must have value.”
But the value of data doesn’t depend on its existence. It depends on its current usefulness, quality, purpose, retention obligations, risk, and management cost. Data decluttering helps separate real value from perceived value.
Useless Data is not neutral
One of the most common mistakes is thinking that unused data is harmless. If we don’t use it, it seems to do no damage. In reality, every piece of retained data continues to live inside an infrastructure: it takes up space, gets included in backups, can be indexed, must be protected, may contain personal information, can be duplicated, and may be pulled into a future migration.
This means that useless data is never completely inert. Even if nobody opens it, it keeps generating cost and risk.
In the FIT Academy white paper “Delete your data. Because it’s the best thing you could do today“, deletion is presented not as an extreme act, but as a strategic capability of data management. The intentional reduction of information assets brings clarity, efficiency, and control.
Role-based AI training helps translate standards into behaviour. It ensures that every function understands both what AI can do and what it should not do without oversight.
Why deleting feels riskier than keeping
Deleting requires accountability. Someone must decide, approve, document, and answer for the choice. Keeping, on the other hand, seems simpler: it requires no explicit decision and pushes the problem into the future.
This is one of the reasons many companies accumulate ROT data: redundant, obsolete, or trivial data. Not because someone strategically decided to keep it, but because nobody had the mandate, the method, or the organizational courage to eliminate it.
The paradox is that keeping everything to reduce risk can actually increase risk. More data means more exposure, more complexity, more ambiguity, and more points to govern.
The role of Governance: turning fear into method
The right way to overcome the bias against deleting is not to tell people to be less cautious. Caution has its place. The point is to turn caution into method.
Data governance helps make deletion a defensible decision. It defines owners, policies, retention rules, classification criteria, responsibilities, approvals, and traceability. This way, the organization doesn’t eliminate data based on gut feelings, but through a shared process, one that also aligns with regulatory requirements such as the GDPR’s storage limitation principle, which requires that personal data be kept no longer than necessary for the purposes for which it was collected.
To make the right choices, a few simple but rigorous questions are needed: does this data still have a purpose? Does it have an owner? Does it contain personal data? Is it subject to retention obligations? Is it duplicated? Is it being used? Is there an authoritative source? What risk does it generate if it stays?
Data Erasers: Professionals in letting go
The concept of Data Erasers, linked to the Forbes article “Data Erasers: The Strategic Team You Don’t Have Yet”, emerges precisely from this need. Companies have many roles dedicated to creating, integrating, and using data. Far fewer are dedicated to assessing when data has reached the end of its lifecycle.
Data Erasers are not improvised “deleters.” They are professionals capable of working across lifecycle management, architecture, compliance, legacy systems, costs, risk, and decommissioning. They help organizations move beyond the “let’s keep everything” reflex with a structured approach.
In this sense, their value is also cultural. They bring a new question into data-driven organizations: not just “what data can we create?” but also “what data can we responsibly stop retaining?”
How to start overcoming the Bias against deleting
The first step isn’t to delete. It’s to make the accumulation visible. Much of the resistance arises because nobody truly sees how much duplicated data, obsolete repositories, and legacy systems actually cost. An initial assessment allows organizations to map volumes, risks, owners, costs, and priorities.
The second step is to classify. Not all unused data can be deleted. Some must be retained for regulatory obligations, others archived, others anonymized, others consolidated. Data decluttering works when it assigns every information asset a coherent destination.
The third step is to measure. Reduced storage, decommissioned systems, avoided costs, mitigated risks, consolidated repositories, and assigned owners are all indicators that help transform decluttering from a subjective perception into a corporate program.
Deleting doesn't mean losing memory
One of the strongest fears is that deleting means losing corporate memory. In reality, good data decluttering does the opposite: it better protects what truly deserves to be remembered.
When everything is kept, nothing is truly valued. Important data remains buried in copies, obsolete versions, random archives, and systems that are hard to query. When an organization reduces the noise, useful information becomes more visible, reliable, and governable.
Digital maturity doesn’t consist of keeping everything. It consists of knowing what has value, what must be protected, what should be archived, and what can be let go.
Table of contents
Status quo bias: why we prefer to leave everything as it is
Loss Aversion: the fear of losing something outweighs the benefit
Endowment effect: we place more value on what we already own
Why deleting feels riskier than keeping
The role of Governance: turning fear into method
Data Erasers: Professionals in letting go
How to start overcoming the Bias against deleting
FIT Academy helps structured organizations transform “just in case” into a governed Data Decluttering process.
Why don't people want to delete useless data?
Because cognitive biases such as status quo bias, loss aversion, and the endowment effect come into play. People prefer to keep what exists, fear loss, and assign value to what they already own.
What is status quo bias in corporate data?
It’s the tendency to leave archives, repositories, and systems as they are, even when reducing or decommissioning them would bring benefits. In data management, it often results in ungoverned accumulation.
Is deleting data always a good idea?
No. Deleting data without a method can create risk. Data decluttering must assess legal obligations, retention requirements, business purposes, privacy, legal holds, ownership, and informational value.
How do you overcome the fear of deleting data?
With governance, clear criteria, assessments, classification, approvals, and traceability. The fear diminishes when deletion becomes a documented and defensible decision.
What is ROT data?
Redundant, obsolete, or trivial data. It can include duplicates, outdated versions, temporary files, unused archives, and content with no operational or regulatory value.
Does data decluttering mean losing corporate knowledge?
No, if done correctly. It helps protect and make more visible the data that is truly useful, by reducing the informational noise created by copies, useless archives, and obsolete systems.