[michielvoo]

Always track interactions (1) with the carousel and evaluate with the client later. You could make a case for a different solution if the statistics show that interaction with 2nd, 3rd etc. content items in the carousel is below acceptable levels (and you've done your best to make it work).

This type of experience (case) can then be carried forward to new projects, possibly convincing designers and clients. Although it might turn out they work just fine, it probably depends on the audience.

[1] Use Google Analytics events to track interactions (next, previous, pause, resume etc.), and add 'campaign' parameters to links inside carousels to compare traffic to product pages originating from carousel links vs. regular menu links (if any).

[lukestevens]

Tracking interaction with events is a great idea. Using campaign parameters internally however doesn't seem like a good idea, fwiw: http://www.annielytics.com/blog/analytics/how-to-trash-your-...

[bshimmin]

This is really sound advice and thank you for posting it. But unfortunately empirical evidence doesn't really wash with clients who use words like "snazzy" and have seen carousels on their competitors' sites...

(No doubt those are the clients you should fire, but we aren't all Don Draper.)