Images usually get added in a hurry. The page is nearly finished, the hero slot is empty, so you search, find something you like and drop it in. The trouble is that this sequence contains no step where the source is checked.
Most of the time nothing happens. Then one day a letter arrives, and by that point the image is on several pages and nobody remembers where it came from.
“I found it in a search engine” is not a source
Image search collects pages an image appears on; it tells you nothing about rights. A good share of what appears there was itself taken from somewhere else by whoever runs that page. Passed from a second hand to a third, the original source disappears and the file becomes “a picture that was going around”.
The test is one sentence. If you cannot state in a single sentence where your right to use this image comes from, do not use it. “The standard licence at stock library A” is such a sentence. “I found it online” is not.
Four lines to check in the licence
Free image libraries have licences, and they differ from one another. “Free” does not mean “use it however you like”. Open the link next to the download button once and look for four things.
Attribution is the line missed most often. Using an image that requires credit without crediting it is a breach, and adding the credit afterwards is awkward. If you are going to use images that need attribution, decide at design time where the credit goes — a caption slot, or a credits list at the foot of the page.
Check modification too. Licences that forbid cropping and colour correction are rare, but conditions restricting composites with your logo, or overlaid text, turn up regularly.
Photographs of people carry a second condition
This is the genuinely risky part. The copyright belongs to whoever took the photograph; the rights in the likeness of the person in it belong to that person. They are separate, and the licence only settles the first.
Stock libraries flag this as a model release — documented consent from the subject. Images without one are frequently limited to editorial use, and putting them into advertising creates a problem regardless of the licence. Placements that imply “this person is our customer” or “this person uses our service” are the clearest example.
For the same reason, watch for trademarks, signage and distinctive buildings inside the frame. A photograph with another company’s logo clearly visible in the background is best kept out of your own promotional material.
Keeping a record is the last step
However carefully you check, without a record you are, a year later, in the same position as if you never had. By then the person who added the image may have left.
It need not be elaborate. Four columns per image — source URL, licence name, date obtained, required credit — is enough. In WordPress you can also write it into the media library’s description or caption field, which keeps the record travelling with the file rather than drifting away from it.
And when the position is unclear, not using the image is always cheaper. There are plenty of alternatives: photograph it yourself, build the screen from shapes and typography, or draw the illustration in code. That last option carries no rights question at all and usually suits the brand better.
Optimising and serving images responsively is covered in the images and media series, and privacy and rights topics are collected in the security archive. If you would like a whole site assessed at once, a diagnostic is part of our optimization program.