Placing a logo watermark that still looks right after resizing
A watermark sized in pixels breaks the moment the image is resized. Relative sizing, margins and opacity are what make it survive.

A watermark has one job: to stay on the image, readable, without ruining it. Doing that across a batch of photos with different dimensions requires choosing the right unit of measurement.
Size in pixels is the wrong unit
If you place a 200-pixel-wide logo on a 4000-pixel photo, it occupies 5% of the width. On a 1000-pixel thumbnail, the same 200 pixels occupy 20% — five times more visually dominant.
This is why professional watermark tools size the logo relative to the image width. Set 15% and it stays 15% whether the photo is 800 pixels or 8000. Doculven uses relative width for exactly this reason, within a 5–50% range.
The same logic applies to position: a margin specified in pixels drifts visually as dimensions change. A margin expressed as a percentage of the shorter side stays proportional.
Choosing the position
Five standard positions — four corners and centre — cover most needs:
- Bottom-right is the conventional choice. It reads as a signature without obstructing the subject in most compositions.
- Bottom-left works when the bottom-right is visually busy.
- Top corners are less common but useful when the subject sits low in the frame.
- Centre maximises protection and maximises visual damage. Use it only for proofs you do not want reused.
For e-commerce, a small corner mark is usually enough. Buyers find a large centred watermark intrusive, and it can trigger marketplace rules about overlaying product images.
Opacity: visible but not distracting
Full opacity looks like a sticker pasted on. Very low opacity disappears after one recompression.
A practical range is 40–70% for a light-coloured background and 60–85% for a dark one. Test on one image at the actual display size before applying to the batch.
Remember that JPEG compression at low quality can blot out a faint watermark entirely. If the image will be compressed afterwards, err on the side of higher opacity.
Logo preparation
The source logo matters as much as the settings:
- Use PNG with transparency. A logo on a white rectangle creates a visible box.
- Use a high-contrast version. A logo that nearly matches the photo’s tone will vanish at moderate opacity.
- Keep it simple. Fine details in a small mark become mush after resizing and compression.
- Mind the aspect ratio. The tool scales by width; a very tall logo may become too small to read.
A common mistake is exporting a logo at 4000 pixels and letting the tool scale it down. Better to provide it near the size it will actually be used, so that downscaling artifacts do not compound.
What a watermark does and does not do
It is honest to be clear about this: a visible watermark is a deterrent, not protection. It can be cropped, painted over, or removed with varying degrees of effort. Anyone determined enough will remove it.
What it does well:
- Discourages casual reuse of preview images.
- Establishes provenance when an image is shared without attribution.
- Signals “this is a proof” for client review.
What it does not do:
- Prevent theft.
- Survive a determined edit.
- Replace proper licensing or a visible copyright notice.
Doculven’s own documentation makes this point rather than overselling the feature.
Batch workflow
- Test on one image first. Position, size, and opacity interact with the content.
- Apply to the batch and review the first three results.
- Watch for edge cases — images where the chosen corner is occupied by the subject.
- Export as PNG if you need transparency preserved, or JPEG for smaller files.
The tool reports each file’s result individually, so a single failure does not block the rest of the batch.