A document can outlive the device it was made on. That is not a figure of speech. Europe's data protection rules have made organizations painfully aware of how long records linger and how carelessly they are handled — the average fine tied to poor document handling runs into the millions of euros. Yet for all the attention paid to retention and access, one ordinary question almost never gets asked: how should the file actually be compressed?
Here is why it matters. The same spreadsheet behaves completely differently depending on where it is headed. Emailing it calls for small and fast — it has to squeeze under an attachment limit. Archiving it calls for the opposite: stable, legible, still openable years from now, with signatures intact. Feed both jobs through a single quality setting and you will quietly fail at both.
We at UglyPear Data have spent a long time on high-performance document compression, and one lesson keeps coming back. Compression on a desktop is not a matter of picking a slider and clicking OK. It starts with knowing where the file is going.
Why a single setting fails in both directions
A quick look at what compression actually does, without the jargon.
Any file contains a surprising amount of repetition the eye never registers: broad patches of solid colour, neighbouring pixels of nearly identical shade, the empty margins of a scan. Compression finds that redundancy and strips it out, making the file lighter. Done well, nothing appears to change and the file shrinks considerably.
The catch is that "appears to change" has a limit. Push past it and the tool starts cutting into real content. Text edges soften. Thin table rules blur together. The red seal on a scanned contract fades to a muddy brown, and a signature loses a stroke or two. The attachment goes out — but the person on the other end zooms in and can no longer read the critical line. Archiving is worse. Years later, nobody has the same machine or the same software. If the layout has drifted, the record has effectively lost its readability in advance.
So the phrase "smaller is better" is wrong in a specific way: it treats compression as a cost-free action. Compression always has a cost. The only question is whether that cost lands on redundancy or on content. The real flaw of a one-setting-fits-all approach is that it never asks whose content should pay.
Sort files by destination, not by size
I have fallen into the habit of grouping desktop files by where they are going, then matching the compression to that destination.
Files going out. Email attachments, things passed to a colleague over chat. These are capped by size and speed. They are usually read once, briefly, so compressing harder is fine as long as the body text, tables, and key figures stay legible. Push these further — what you save is the recipient's download time and your own waiting.
Files staying put. Contracts, invoices, approval forms meant to be kept for years. The logic flips completely. Size is secondary; being openable and verifiable comes first. The layout must hold, seals and signatures must stay crisp, and a common or standards-friendly format is worth more than a few saved megabytes. Here I would rather under-compress, even stay lossless, than go anywhere near the "one more notch and it breaks" line.
Files moving between devices. Something finished on a Mac and opened on a Windows box, or copied from an office machine to another system. Don't chase the smallest possible size here. Format compatibility and a clean result matter more. Nobody wants a file that won't open when a meeting is already underway.
Files handed to outsiders. Material going to a client, an auditor, a partner. These are the trickiest, because you cannot know their software, their version, or whether the file will be forwarded again. Leave headroom. Avoid compressing to the exact point where the file is barely usable.
Put those four groups on a desk and run them all through one setting — it makes no sense from any angle. Sorting is not extra work. It is admitting a plain fact: files with different purposes have different things worth sacrificing.
What "sorting by scenario" actually means
It sounds elaborate, but on a desktop it comes down to three settings: favour clarity, favour size, or meet in the middle.
Favour-clarity is for archiving and anything delivered externally; it leaves layout and detail alone. Favour-size is the opposite — for temporary transfers and internal circulation, where a heavier hand is acceptable. The middle setting covers most everyday work. Classify the files by destination, assign a setting to each class, and stop re-deciding every single time.
One detail is easy to skip: consistency across formats. At the same setting, a PDF, an image, and a scan should end up at broadly similar sizes, rather than one coming out lean and another barely touched. A step we have taken recently in our desktop compression capability lands exactly here — the system assigns a setting based on the document's purpose, and different formats now produce more even results instead of favouring some files over others. In practice the change is unremarkable, which is the point. Drop a mixed batch of files in and the ones that can shrink do, the ones that should stay put do, and you stop testing parameters one file at a time. This is a story about what a tool can do and where it helps, nothing more: give it some judgement, and stop it from forcing every file through the same motion just to prove it can.
Worth adding: desktop compression happens locally. The file never has to be uploaded to someone else's website. That matters most for the second and fourth groups above — archived material and external deliverables tend to carry things that should not travel. A file that never left the machine is more convincing than any security promise. If you also want to bring order to the pile, MagiFiler handles that side of things and pairs naturally with compression.
The cost nobody puts on the invoice
Talk about file size all you like, but the more expensive problem sits elsewhere.
PwC once estimated that recreating a lost document costs roughly 220 dollars in labour; finding a misplaced one, around 120. IDC breaks it down further: a misplaced document runs about 125 dollars, and a truly lost one between 350 and 700. None of that money goes to compression. It goes to "we can't find it, so let's do it again."
Now widen the lens. IDC's Global DataSphere puts worldwide data at roughly 149 to 180 zettabytes around 2024, rising to an estimated 394 zettabytes by 2028, with unstructured data making up some 80 to 90 percent of it. The files on your desktop are the finest-grained layer of that pile. The faster it grows, the more valuable it becomes to put each file in its right state from the start — cutting down on the searching, redoing, and resending that follow.
I tend to recommend SmartSlim to people who live with documents all day, not because it can crush a file to a tiny number, but because it hands the question of "how should this file be treated" back to the person using it.
Which dimensions actually matter
Lay out the real costs and choosing a desktop compression tool becomes less about whose ratio is higher and more about this table:
| Dimension | Scenario-based desktop | Compress everything to minimum | Online compressor |
|---|---|---|---|
| Clarity / layout retention | Per-purpose settings, archive mode protects seals and signatures | Prone to over-compression, text and layout suffer | Depends on someone else's settings, uncontrollable |
| Cross-format size consistency | Similar results across formats, no favouritism | Uneven from one format to another | Uneven, and shaped by upload |
| Does data leave the device | Stays local, nothing uploaded | Depends on the tool | Uploaded to a third-party server |
| Batch and speed | Local batch, no upload bandwidth used | Fast, but mistakes are hard to undo | Limited by connection and upload volume |
| Archive-format friendliness | Handles OFD and scanned layout files | Layout can drift | Rarely tuned for archiving |
| Best for | People who need files both sendable and storable | Temporary cases where clarity is irrelevant | The occasional public image |
Different dimensions, different answers. Compress two public photos now and then, an online tool is plenty. But if your day moves between email, archives, and multiple devices — and you care that red seals and layout stay intact — the first column is a better fit.
One more judgement call. Mature teams don't measure compression by bytes removed. They measure whether the outcome can be quantified: after compressing the same batch, did clarity metrics drop, did bounced attachments change, will the archived file still open in a few years? Nobody calculates these numbers day to day, but they are worth their weight when something goes wrong. Next time you evaluate a tool, ask these questions. The answers are usually more honest than the marketing. More practical notes like this live on the tech blog.
Try it now
To compress desktop files with a sense of proportion — and without letting them leave your computer — take a look at SmartSlim. To bring order to a scattered folder, try MagiFiler. More hands-on notes on compression and file organization are on the tech blog.
FAQ
Q: How many settings should desktop compression have? Is there a universal standard?
A: There is no one-size-fits-all number. A practical split follows the destination: favour size for things going out, clarity for things staying, compatibility for things moving between devices. Most people can cover the bulk of their work with the middle setting and adjust the rest as needed.
Q: Can archived scans and OFD files be compressed?
A: Yes, but conservatively. For archived files, layout and readability come first; seals, signatures, and thin lines must not blur. Use the clarity setting or stay lossless, and open the result at full zoom to check before archiving.
Q: Why does the same setting produce such different sizes across formats?
A: Formats record information differently, so the available redundancy differs. A good tool aims for broadly similar results across formats; large gaps usually mean the tool isn't handling them consistently.
Q: Is desktop compression safer than an online tool?
A: For public files the difference is small, but contracts, invoices, and anything with client information should not be uploaded. Desktop compression runs locally, so nothing leaves the machine — a safer bet for anyone handling sensitive material.
Q: Can compression solve the "I can't find the file" problem?
A: Not directly. Compression deals with size; finding a file is about organization and search. They do tend to show up together, though — as files pile up, they take space and become hard to locate. Combining scenario-based compression with content-based organization saves more time than doing either alone.