Cloud storage cost reduction: a practical guide

Why cloud storage bills keep growing

Cloud storage is priced to feel cheap per gigabyte, which is exactly why it becomes expensive at scale. You never make a single big decision to spend more; instead thousands of small, invisible ones accumulate until the monthly invoice is a line item nobody can explain. Effective cloud storage cost reduction starts with understanding where that growth actually comes from, because you cannot cut what you cannot see.

There are four repeat offenders:

• Duplication. The same file is uploaded to three team sites, forwarded as an attachment, copied into a shared drive, and re-saved as "final_v2". Each copy is billed independently, and most estates hold far more duplicate bytes than anyone assumes.
• Version history. Platforms like SharePoint and object stores with versioning enabled keep every revision of every file by default. A frequently edited document can quietly hold hundreds of stored versions, each consuming quota.
• Dark data. Research from the Veritas Databerg report found that around 52% of stored enterprise data is "dark": retained but never used and of unknown value. You are paying full price to store, secure and back up data that delivers nothing.
• No lifecycle policy. Most data has no expiry date attached to it. Nothing ever moves to cheaper tiers or gets deleted, so every byte you have ever created stays on your most expensive storage forever.

Step one: audit what you actually store

Before you optimize anything, you need an inventory. A useful storage audit answers four questions for every location you pay for (cloud object stores, SharePoint and OneDrive, NetApp and NAS, and local drives):

• How much data is here, and how fast is it growing?
• When was each file last accessed or modified?
• How much of it is duplicated elsewhere?
• What is each category costing per month?

The last question is the one that changes behaviour. "You have 40 TB of files nobody has opened in three years" is easy to ignore. "Those files cost you a recurring amount every month, and here is the number" is not. A good audit expresses waste in the currency finance already tracks.

Identify ROT and duplicates

Most reclaimable storage falls into a category storage professionals call ROT: redundant, obsolete and trivial data. Redundant means duplicates and near-duplicates. Obsolete means files tied to projects, employees or systems that no longer exist. Trivial means data with no plausible business, legal or compliance value: the caches, temp exports, and personal media that drift into shared storage.

ROT is where the fastest wins live because removing it carries almost no risk. Nobody misses a duplicate of a duplicate, or a backup of a temp file from a project that shipped four years ago. The challenge is simply finding it across a sprawling, multi-platform estate, which is exactly the work manual audits never finish.

Lifecycle and tiering versus actually deleting waste

When teams finally confront a storage bill, the instinct is to reach for lifecycle rules and tiering: automatically move older objects to cheaper "cold" or "archive" storage classes. Tiering is genuinely useful, but it has an important limit: it makes waste cheaper to keep, it does not stop you keeping it.

If half your estate is dark data and duplicates, tiering all of it to cold storage still leaves you paying to store, index, secure and eventually retrieve data that has no value. You have reduced the unit price of waste, not the amount of it. Worse, retrieval fees on archive tiers can make occasional access surprisingly expensive.

The durable strategy is a sequence, not a single lever:

• Delete the ROT. Remove redundant, obsolete and trivial data first. This shrinks every downstream cost at once: storage, backups, replication and carbon.
• Tier what remains but is rarely used. Once the estate is clean, lifecycle rules on genuinely cold-but-valuable data are efficient rather than a way to hoard waste cheaply.
• Keep hot data hot. Leave active, frequently accessed data on fast tiers where retrieval is free and instant.

Delete first, then tier. Doing it in that order means the lifecycle rules you write apply only to data worth keeping.

Measuring cost per finding

The reason storage cleanups stall is that they are framed as risky chores with no visible payoff. Attaching a euro figure to every finding flips that. When each duplicate cluster, oversized file and stale library shows its own monthly cost, prioritization becomes obvious: you tackle the biggest numbers first, and you can report reclaimed spend back to finance in terms they already use.

The same measurement principle applies to carbon. Every gigabyte you store draws power for the drive, the cooling around it, and the replicas and backups behind it. Measuring cost per finding in both euros and CO2 turns an abstract housekeeping task into a defensible business and sustainability case, the kind that gets funded and repeated.

How TerraBytes automates the audit

Everything above is straightforward in principle and brutal to do by hand across a real estate. TerraBytes is a storage cleaner and optimizer built to automate exactly this audit. It is serverless and private by design: your data never leaves your environment, and everything is encrypted in transit and at rest.

One scan spans SharePoint, OneDrive, NetApp, NAS and local drives, and flags duplicate, outdated, oversized and risky files in a single prioritized view. Critically, each finding is costed in both euros and CO2, so you fix the biggest wins first, and nothing is ever deleted automatically. TerraBytes flags and recommends; you decide what to keep, tier or remove.

If you want to see your own numbers, the fastest route is to run the audit itself. Explore the storage cleaner and optimizer to scan your whole estate, or start with a focused SharePoint storage cleanup if Microsoft 365 is where your bill hurts most.