The greenest file is the one you deleted

Most of what any organisation stores is never opened again after it is written, and it is powered continuously for years regardless. Deleting it is the only intervention in this whole subject that is large, free, and entirely within your control — and it is the one nobody schedules.

The shape of the problem

Storage is a continuous cost. A file sitting on a drive consumes energy every second it exists — the drive powered, the cooling running — for as long as you keep it. Transferring the same file costs energy for a few minutes and then stops.

So the footprint of your data is dominated by how long you keep it multiplied by how many copies exist. Not by how you sent it, and not by how well it compresses.

Now put that next to what is actually in most storage. Surveys of enterprise data consistently find that a large majority is never read after being written. The exact proportion depends on who is counting and how, and the methodology is usually loose — but nobody who has audited their own storage disputes the direction.

That combination is the argument. Continuous cost, applied to data with no readers.

Where it accumulates

Every version of everything. Six exports of the same video, each a full copy. Nine PSDs called final-v3-actualfinal.

Working files kept at deliverable priority. The RAW footage, the project files, the intermediate renders. Enormous, and almost never opened after sign-off.

Backups of things that should not exist. If you never deleted it, your backup faithfully preserves it, often several generations deep.

Automatic sync that nobody configured. Three services holding the same folder because each was set up separately.

Shared drives with no owner. The folder someone made in 2019 that everyone is slightly afraid to remove.

For anyone doing media work the numbers get large quickly: an hour of ProRes is around 300 GB, so a year of regular work is many terabytes, most of it unused footage.

Why this is the cheapest lever available

Compare the options honestly:

Compressing a file by 20% reduces its footprint by 20%, forever, at the cost of a re-encode and possibly some quality.

Switching to a greener provider reduces the carbon intensity of the energy, which is real, and does nothing about the quantity.

Deleting the file reduces its footprint by 100%, immediately, at the cost of a decision.

And deletion pays three ways at once, which is unusual:

• Energy. The file stops costing anything.
• Money. Storage bills scale with volume.
• Risk. Data you deleted cannot leak. If it contained personal data, holding it with no current purpose was a compliance problem anyway — storage limitation is a principle, not a suggestion.

There is no trade-off to manage. That is rare enough to be worth acting on.

100%

Making it happen

The reason this does not get done is that deletion requires a decision and nothing forces one. Storage never fills up dramatically; it just gets a bit more expensive each year.

Four things that work:

Separate the categories. Deliverables, working files, personal data, financial records. Different lifetimes. Applying one rule to all four is how people end up either hoarding or deleting the wrong thing. Choosing periods.

Write one sentence per category. "Working files and RAW are deleted six months after final delivery." That is a policy.

Put it in the calendar. Twice a year. Deletion that depends on someone feeling like it does not happen.

Prefer defaults that expire. For anything that fits, use a service where deletion is the default rather than a chore. Free transfer tiers delete within 7 to 15 days, and that accidentally-correct behaviour is doing you a favour — even when it is annoying.

Check the backups. Deleting the working copy while three years of backups retain everything achieves nothing. Ask your provider how long deleted data persists.

The counter-argument, taken seriously

There are good reasons to keep things. Clients come back. Disputes arise. Some records have statutory retention periods. A photographer deleting a wedding archive at six months will regret it.

None of that is an argument for keeping everything. It is an argument for keeping the deliverables and deleting the 700 frames that did not make the cut — which is the same amount of care and a fraction of the volume.

The failure mode is not "we deleted too much". It is "we never decided, so we kept it all", which is not a decision at all.

Perspective

For an individual this is a small share of a personal footprint, and anyone telling you that clearing your inbox is meaningful climate action is overselling it. The scale is worth keeping honest.

Where it is material: at organisational scale, in media work, and as an operating discipline that happens to reduce cost and legal exposure at the same time. Those are enough.

If you are also choosing a provider, what actually makes one greener covers it — but do that second. The vendor choice is worth ten minutes; the deletion habit is worth more than any vendor choice available.