Short version: If somebody has to create a new pool of keywords from scratch, but doesn't understand how to do it, what is a good way to teach them how to pick good keywords?
TL:DR version
My employer has an internal database of animals, and everybody is complaining that they can't find what they are searching for. This is understandable, as the item description rarely contains useful search terms.
We (= the inhouse IT department) implemented a keyword functionality. In the future, anybody entering a new animal should attach existing keywords to it (there is currently no way for users to add new keywords, but I suspect a manual process involving phoning the DB owner will establish itself).
The DB owner has access to an interface which lets her enter keywords as strings instead of choosing them. She now has to go through the existing ~2500 animals and assign keywords to them. These will be later available to users.
The problem is that the DB owner, while being very intelligent and a good professional, has an unfortunate anti-talent for structuring information. She is also uncomfortable with computers, and while she uses them, she prefers working on paper whenever possible.
This time, she did the first few dozens of animals, and we discovered that what she entered was not keywords, but lengthy descriptions. For example, if an animal has a mutation in the ptgs-1 gene, she enters a mutation of the ptgs-1 gene leading to disruptions in the COX-2 cycle, exhibits squamous cell hyperplasia
when the right thing to enter is probably a ptgs-1
keyword and a squamous cell
keyword. I tried to explain to her briefly that this is not how she is supposed to do it, but this didn't help.
Sadly, nobody else can do the tagging, as not even the other scientists have the necessary knowledge in biology. My boss scheduled a talk with her where we will try to educate her. We are thinking of creative ways to get her to understand how tagging with keywords and keyword-based search work, so she can somehow get the job done. Any ideas how to help her build the correct mental model? I have a vague concept of letting her attach paper labels to (copies of) animal descriptions printed on paper, but I am not sure how well can I build the analogy, and I am also afraid of coming across as patronizing if I start playing paper games with her.
*update in response to the current answers. *
- The keywords are really needed here. There is no full-text to be searched, just structured information, and it is not the information used for search queries. For example, a mouse can be good for researching obesity, but the word "obesity" will not appear anywhere in the record, so the idea is that the author will be able to add it as a keyword.
- I was serious about the "anti-talent" part. It is like being tone deaf, only for information structures. She is determined to do the job despite this being very hard for her, and I want to help her. But solutions have really be something which will work for a child, or for your favorite grandma who can't find her cookie recipe on the computer if the file isn't on the desktop.
a mutation of the ptgs-1 gene leading to disruptions in the COX-2 cycle, exhibits squamous cell hyperplasia
then searching forptgs-1
should still bring up the relevant record by searching in the keyword text. What's the issue..?