Freelance Keyworders Plus and Minus

With the downturn in the economy, there has reportedly been a surge in numbers of freelance keyworders.  This pool of labour can fill an important gap, but there are potential problems too.

There is a clear advantage in being able to add such people to your staff quickly, and dispense with their services when they are no longer needed.  Often they have good skills, having keyworded for a photo library or libraries before.  At the very least they will have knowledge of the picture business if they have come from sales or the photography side of a photo library.  Ideally then, freelance keyworders can be a pool of skilled labour which can be called upon whenever the need arises.


So what could possibly be wrong with that?

To start with, the ease with which freelancers can be employed is very much a double-edged sword.  As easily as you can dispense with their services, they can decide to work elsewhere.  This is particularly so if the work you offer them is sporadic.  It is likely that in time they will find more regular work elsewhere, and the next time you call them they will already be booked.

There is also a tendency for the overall pool of freelancers to expand and contract based on economic conditions.  Those keyworders who are working freelance may well be doing so out of necessity, and when a permanent position comes up they are likely to take it, especially if it’s their old job.

Freelancers, for all their skills, will always need some level of training.  If your system has a controlled vocabulary, that might be considerable.  If you’re likely to keep that person a long time that’s great, but if they are likely to leave in a short time, then it might not make economic sense.

Because freelancers are not on staff, you have less control over what they do.  Should a bad mistake be made with keywords – perhaps even libellous or with other legal consequences – you may not be in much of a position to call them to account.  Day-to-day supervision may be a problem if the freelancer wants to work from home.

So before hiring freelance labour, read through these tips:

  1. Find out early about their work intentions – Is this a temporary state of affairs, or are they likely to be loyal (proxy) staff members?
  2. Ask who else they work for – If there is a big company using their services, you are likely to run into availability problems.
  3. Get references from previous employers – Just because someone says they are a keyworder, it doesn’t mean they are good or even have experience.  Even if they have experience in one type of keywording, it may not be transferrable to what you need.
  4. If possible, use them for work requiring little training – That will keep down costs and make them easier to replace.
  5. Think carefully about whether to have freelancers work from home – It is far harder to train them and give timely feedback about their work when they aren’t in your office
  6. Don’t be seduced by the relative cheapness – quality of keywording is a major consideration, so don’t fall into the trap of sacrificing quality just to save a little money.
  7. Get them on a contract if possible – Try to get them to commit to a regular amount of work to solve the problem of availability, provided that would fit with your workflow – a regular commitment may not work for you.
  8. Before going down the freelance route, consider very carefully other option – It may well be better to hire them permanently or use an outsourced service which has just as much flexibility but is far more reliable.
  9. Count the full cost – Before hiring, work out exactly how much a freelancer will cost, including the cost of them being in the office, being trained and supervised.
  10. What if they up and left? – Consider how badly your business would be affected if the freelancer up and left without warning.

Your In-house Keywording Is Just Fine – Or Is It?

Keywording in-house, rather than using an outsourced solution, is often a good fit.  To make it even better, consider having your work audited.  It will keep your keywording in line with industry trends, and highlight areas for improvement.

Whilst in-house keywording teams have the advantage of learning the exact idiosyncrasies and nuances of the system being used, there is a danger that the system itself can be left behind by what is happening elsewhere in the world of metadata, or can become too well suited to people inside the organisation, rather than the general public.

In celebrity keywording, for instance, there is a trend towards including designer labels in clothing descriptions.  If your keywording doesn’t follow that trend, you may find yourself being eclipsed by competitors.  With stock keywording there has been a trend towards more conceptual words, and tighter keywording to better-focus search results.



It should be remembered that as an image or video library grows and changes, so do its keywording and metadata requirements.  A library of 10,000 items is a totally different beast from one which is 200,000 in size, where more elaborate systems are called for.  Customers’ requirements will also change over time, as will the standard of keywording being offered by your competitors – sales can suffer dramatically if you don’t keep up with these changes.

You may also find that at some point you want to submit to larger image and video aggregators such as Getty Images, Alamy, Corbis or Framepool.  If so, the in-house system you are using may be well off the mark for meeting the requirements of these companies – in fact that’s almost guaranteed.

With particularly small teams of keyworders – down to a single part-timer for instance – there is a danger of capture and keywording drift.  What that boils down to is that the person keywording, left to their own devices, decides entirely what should be done for better or worse and that standard may not be what was originally envisaged.  Worse, there can be drift in the way keywording is carried out, the vocabulary being used and so on, so that keywording in older parts of the archive is totally out of step with newer parts.

In all of these cases, the best way to stay on track, and get the keywording you need, is to get someone from the outside to look at what you are doing, sample various parts of the archive and report back.  If everything is humming along nicely then it’s good to know that.  It’s even more important to know about the problems so you can address them before throwing more money down the drain on inadequate work.

And because the critique is coming from outside the organisation, it is likely to be an easier message to deliver to colleagues who you also regard as friends.

Perhaps you already know of a metadata professional who can audit your keywording for you.  If not, try talking to keywording outsource companies and see if they will do an audit for you.

Try The No Keywords Test

If you want to put yourself in the shoes of your customers for a while, and find out the value of keywording, try the "no keywords" test.

Take a selection of images or videos and remove all keywords (if they have any).  Add them to your photo/video library as duplicates.  Then carry out a search to see how easily you can find those images/videos using your quick search window - this is the window used by the vast majority of researchers.


To make it realistic, see if you can find the images/videos using search terms other than the literal name of what you are searching for.  You can try colours, concepts, themes, the type of object (eg if there is a picture of an apple, try searching using "fruit"), and see what results you get.

If the keyworded images/videos are coming up more quickly and more readily, then you'll get a fair idea of how much better your keywording is than no keywords at all, and also whether metadata you use other than the keywords field is of much use.  Record the time taken to complete each task if you want to be scientific about it.

You can use the same method to evaluate different keywording methods and standards.

Keyword Health = Photo Library Health

If you want to know if a photo library is thriving, there's a very good way to find out without (fruitlessly) trying to get hold of sales figures and balance sheets - look at the keywording.

Whilst this is hardly a precise scientific measure of success, you can learn a lot from a company's keywords, like whether you should be contributing your images or videos to them, whether they have good cashflow, and whether they are as professional as they should be.



And here are three reasons why:

1.  Keywording is a habit shared by the successful - and they must know something . Getty Images, Splash News, Corbis, Frame Pool -  emphasis on good-quality keywording is just part of the game for such market leaders.  So if it's worthwhile for them, it should be worthwhile for any other successful and/or well-organised photo library.  They must have the figures to prove it, or they wouldn't be doing it.

2.   If you have the money for keywording, then you are likely to be putting money into other important areas such as advertising and sales.  A full-on marketing effort is likely to produce good sales and healthy cash flow.  Poor keywording, or rudimentary keywording is a sign that the library is struggling to free up staff, or purchase outsourced keywording for this basic function.  If they aren't doing that right, what else aren't they doing to market images and get sales?

3.   Big is beautiful in the image business, so if the library isn't big enough to need keywording, then perhaps the library isn't a substantial player.  It's easy to understand how a "mom and pop" operation would tend to be small and disorganised.  If there really aren't enough images or videos to make keywording necessary, then customers are likely to be going elsewhere.

So if you are planning to deal with a photo or video library, take some time to have a look at the keywording and if it isn't up to scratch, ask why.

Rubbish In, Rubbish Out

There is an old saying in computing: "rubbish in, rubbish out".  The same goes for keywording and captioning. If your keywording goes awry, perhaps the real reason is in the poor metadata you supplied in the first place. 

Generally speaking, keyworders aren't mind readers, whether they are on staff or in outside companies.  That means they are unlikely to be able to identify every species of bird, or know every city or building in a particular country.  So rather than risk them guessing wrongly, or not including the information at all, make sure all metadata supplied is as full and accurate as possible.


If research is going to be required, you should let the keyworder know where to find the best sources of information, preferably on the internet.

You can also save time and money by getting the language, grammar and spelling correct, particularly spelling of names.  There is usually no spell check for people's names, even those of celebrities, so the onus has to be on the photographer/videographer or photo library to get such details correct in the first place.


If you feel that photographers/videographers can't get it right in the first place, you may find that having someone to collate the information and fix up basic errors will produce a better result than waiting until the keywording is halfway through.

In terms of formatting, try to supply metadata in a spreadsheet, or better still in the IPTC fields of the images. This will make it easier to ingest the information into keywording software, and avoid being charged for extra admin time.  Handwritten notes or unclear photocopies (even in PDF form) are bound to cause problems, so avoid whenever possible.