We had an email from fotoLibra member Jon Lees this morning. He wrote

I uploaded a image (Belfast Duck Tours) on Tuesday, supplied several keywords and a title.  I ran a Google search on the subject matter this afternoon and did not get one fotoLibra hit. In fact there are not many hits on this subject matter, the business website and old newspaper article from an earlier press release, so in theory finding my picture should be very easy? Similarly searching for my photo of the Marine Current Turbine fails to generate any hits without mentioning fotoLibra in the search. Is this then a failing of the website, or are sales not generated this way?  I must admit I know little about how search engines generate their information, but surely there must be a way of raising the site profile?

Given that the Google search robot alone uses up 4 gigabytes of our bandwidth every month trawling the fotoLibra site, we asked our Technical Development Manager Damien to respond. This is what he wrote:

What we aim to do is to improve the fotoLibra user experience for both buyers and sellers.

Search engines have hundreds of thousands sites to crawl every day, and many of them have thousands of pages. Though we give them instructions to help them find pages or combinations they would not necessarily guess, and try to add and present as much as possible relevant and descriptive data — we recently added metadata, categories, and collection names, which they would not have guessed — we can’t force, or even suggest to them, what they should do.

So they sometimes crawl aimlessly. They simply visit, so to speak.

But they usually have something precise in mind, and will query terms they know they don’t have enough relevant data about. And they do the same on other websites. Once they’ve gathered data, they are the ones who finally “decide” what and who is to rank higher, using complex algorithms.

This process can take weeks.

We can’t force them. They might feel we’re trying to fool them.

So what is important is: relevant and correctly spelled keywords. Informative, concise captions AND descriptions, in good English. Search engines just can’t analyse what’s in a picture, yet. They rely on the surrounding text which is expected to be descriptive. And search engines “prefer” good English, and tend to ban what they call “keyword stuffing”.

Another crucial point: other sites have to link to fotoLibra. We have to be the site people talk about.

Photographers’ sites have to link to us. fotoLibra has a wide range of very professional and specific pictures: specialised sites have to link to pictures on this particular domain too.

And so should forums, blogs, articles or anything that deals with picture researchers.

Because people interested in the subject read them, and they’re crawled by the search bots as well.

So, if anyone wants his pictures to be more likely to rank higher, he also has to post on forums or blogs or whatever he likes and uses to share his passion, and to tell people that he’s got amazing, rare, high quality pictures on that given subject, and that they are available at this place which has got hundreds of great pictures of his: fotoLibra.

There is no magical receipe. We’re trying to present and emphasize what’s good in it. If it helps the user finding images he’s looking for, the search engine will feel it does too. We’re targeting the improvement of user experience rather than being rigidly search engine compliant, like a porn site. The pictures are getting increasingly accurately described. The Pro Search tool is always improving. Then it gets better for everyone. That’s part of our job.

We have great quality images that people won’t find because they’re poorly captioned or keyworded. That’s the responsibility of fotoLibra’s members.

Good images must be advertised and talked about. People have to spread the world around the web, and elsewhere.

And that’s a part of all our jobs.