I am working on a project for an ecommerce website where there are multiple personas. Each persona will have a different mental model of the system so the IA is proving especially hard.

My original comment: "As well as allowing users to review certain products, one could allow users to tag products, this way the taxonomy will better match users’ mental models of the system." -

and stakeholder comments: "is there evidence to show how many users actually do this? Anecdotally I’ve observed that whilst folksonomy was trendy about 5 years ago, not many users actively tag. This means that for the effort required to set up a folksonomy infrastructure, there is very little benefit. Secondly folksonomy requires a very large number of people to tag so that the tags can be weighted to be effective."

my question: I want to propose to stakeholders that allowing users to tag certain products may be good practice but i am looking for best practice / evidence for support.

  • Allowing users to specify tags seem to have been working for Twitter. May be a bit "far out there" in comparison to what you're trying to achieve. But a stakeholder may still listen to that ear. Commented Jun 18, 2012 at 10:04
  • It's basically a crowdsourced version of cardsorting then. As a hypothesis I'd say that in the long-term it will improve the IA as more people use and tag stuff the IA should grow out of that, but in the short-term nothing beats the structure of an organised and scheduled card-sorting exercise.
    – JonW
    Commented Jun 18, 2012 at 10:21
  • I agree, but I came into the project late and none of this had been done, so I need to make recommendations for the IA...without any research to back what I say up...]
    – colmcq
    Commented Jun 18, 2012 at 10:48
  • I work in E-Commerce. On roundabout 500 shops of clients of ours, not a single user has ever tagged a product.
    – vzwick
    Commented Jun 18, 2012 at 11:02
  • hmmmm interesting.....
    – colmcq
    Commented Jun 18, 2012 at 11:05

1 Answer 1


Ideally you would be able to have a bit of both.

A robust, expertly designed structure based on user research. Plus an ongoing folksonomy that allows for ongoing improvement and validation.

However the comment above that few users take the time to add a tag is generally correct. When adding a profile of themselves or uploading their products for sale, users have motivation to contribute. However if it is an e-commerce site they are buying from they expect this to be done. Why should they help? You need to look for the motivational trigger. There are a few things to look into:

  1. Integrate the act of 'tagging' into other functions on the site. (Searching, users organising saved content/bookmarking etc)
  2. Use something like Amazon Mechanical Turk to quickly get a large set of human generated tags. May not be as accurate as real research but it could provide a cheaper starting point.
  3. Gamify the act of tagging to reward users for helping you out.

One thing to note with user generated tags is that unless you get a large sample set you can get distortions in the data. However a good UX persion should be able to help with this (that's why they are experts).

Also, any user generated tags requires the site to be operational and used for some time. This means you need something to start with and you wouldn't want to start with a site that frustrates or looses you customers.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.