Most firms price their document problem wrong. They treat it as a storage line item — buy more disk, archive to cold storage, move on. The storage number is real but it's the smallest part. The actual cost hides in two places that don't show up on any invoice until something breaks: the hours people burn looking for things, and the risk of a file going somewhere it shouldn't.
Start with the hours. IDC has a survey where knowledge workers spend about 2.5 hours a day hunting for the information they need — roughly 30 percent of the working day. Sit with that for a second. In a firm of any size, nearly a third of the payroll is functionally spent relocating information rather than using it. A partner doesn't feel that directly; an associate does, scrolling shared folders at 11 p.m. trying to find the executed version of a clause they swear they saw last week.
Then there's the volume problem underneath it. IDC also forecasts that China's unstructured data will climb from 51 ZB to 129 ZB over the next five years, at a compound annual rate somewhere between 30 and 40 percent. Case files, contracts, due-diligence reports, court rulings — almost all of it is unstructured. It doesn't shrink on its own. Every new matter adds more, and the search problem gets worse not because people are disorganized but because the pile grows faster than any naming convention can keep up with.
The leak risk is the part nobody wants to price. IBM's 2025 Cost of a Data Breach report puts customer PII at 53 percent of all stolen data, employee PII at 37 percent, and intellectual property at 33 percent. Legal documents are dense with exactly those things. A divorce agreement, a merger term sheet — once it leaves the building, it's an incident, not a thought experiment. Confidentiality in this field isn't a feature you opt into; it's written into the professional duty itself.
So when a firm goes shopping for a "knowledge base," the framing is usually backwards. The question isn't "how do we search faster." It's "how do we stop paying the hidden tax." Three moves tend to matter more than any search feature.
The first is deciding, once, who owns the real version. The classic mess — same contract in five folders, named final, final2, FINAL_final — isn't solved by a better search bar. It's solved by a human marking one file as the one that counts. That single decision removes most of the "which version is current" questions, and it costs nothing but a moment of attention.
The second is keeping sensitive material off tools that can't guarantee where it lands. Plenty of teams quietly paste client details into public chat tools to summarize something, because internal search is so bad that's the only way to get an answer. That's how a client's information walks out the door without anyone meaning to. The fix isn't a stricter policy memo. It's a system that flags and scrubs client IDs and account numbers before anything enters a search index, and does it on your own network where you can actually govern it.
The third is shrinking the files themselves. Scans and PDFs are often enormous, and searching across giant files is slow and expensive to store. Running a back catalog through a local compression pass — everything stays on your machines, nothing uploaded — drops the volume substantially. Searches get faster, backups get cheaper, and sending a file to a colleague stops taking five minutes. The principle matters more than the brand: compress before you index, and do it on your own hardware.
What surprises people is how much of the day comes back once these are in place. Not because any one tool is magic, but because the three hidden costs — search time, version confusion, leak exposure — compound each other. Fix one and the others loosen.
If you're measuring this for your own shop, resist the urge to start with a vendor scorecard. Start by timing a typical "find me X" request end to end, and count how many of those answers depend on a file leaving the network. Those two numbers tell you more than any feature comparison.
UglyPear Data builds the pieces behind this approach — document compression and retrieval-augmented generation that run entirely on your own infrastructure, with sensitive information detected and redacted before it ever reaches an index. The point isn't the tooling, though. It's that the cost was never really about storage.
Try it now
To see how private compression and retrieval would land in your own matter files, start at UglyPear Data's product page https://www.uglypear.com/en/products/ and the RAG data engineering base overview https://www.uglypear.com/en/ . The engineering blog at https://www.uglypear.com/en/blog/ carries more on deployment and redaction order.
FAQ
Q: Is the 2.5-hours figure specific to law firms?
A: No. It's a general knowledge-worker survey number. Legal work tends to run higher because a "document" is usually several files plus email threads plus scans, so the search tax is heavier, not lighter.
Q: Doesn't cloud storage already solve the volume problem?
A: Storage is cheap; the cost isn't the disk. It's the search time, the version confusion, and the exposure when sensitive files move off-network. More storage doesn't touch any of those.
Q: We're a small firm — is this overkill for us?
A: Smaller firms feel the search tax per person even more, because there's less redundancy. The local-compression-and-redaction step scales down well; you don't need a large architecture to start.
Q: How do we measure success without a big project?
A: Time a few real "find me X" requests before and after, and count how many answers currently depend on files leaving the network. Those two deltas are your business case.