Suggestive search limit

Question

So I'm currently having a debate with people I work with around a suggestive search box. The debate is around what's the best UX for displaying large amounts of content. As an example:

If a user types in at least 3 letters it will drop down with a list of suggestions based on that. Since our database is large for the sake of the example lets say theres a possible 200 possible outcomes. One side of the room thinks that showing as many as possible from those 3 is the best outcome because it's better to show the wide range on offer. To make this work they want to put a scroll bar on it... and if you don't like the results you type more

On the other side we have people who think it should be limited to 2-3 per category so as not to have the list too long and to get a better result you type more characters in so as not to be confused and bombarded with too many options when you've only typed 3 letters in.

Is there a standard practice in this? (using Bootstrap Typeahead with solar)

Answer 1

You could show a reduced set of matches in relevant categories and add a method for the user to explore more if it's of interest.

This way, you show the wide range on offer, and limit the display to only a few.

The user has less information to filter themselves and so it's easier to decide what's relevant - i.e. what to explore or what not to explore and whether they should keep typing.

It also shows which categories are the most relevant - you could order the categories.

Answer 2

Roger has already covered the wide range problem, and I terribly agree with his proposed solution.

The brain has limits

What perhaps important to emphasise to the first half of the room is that the brain has limitations, and long list are not its friend.

As a very rough figure, one research has shown that around 70 of similar items (search results) is where people give up.

Information foraging

Seeing as this is a typeahead related question, one theory that comes to mind is the information foraging theory, which while observed first in nature, was then demonstrated to hold with online information seeking behaviour:

In short, and within UX context, the marginal value theorem asserts that users will take action to depart from the current state based on:

• The time/effort it takes to get to the next state.
• The prospect of the next state.

In your case, the time/effort factor is a keystroke - an ignorable price to pay; and the prospect is a refined search with better hits and less results.

This suggests that long lists are likely to be ignored - users shall much prefer to refine their search since the cost is low and the benefit is high.

The twist

Now do be aware there can be an awkward twist here - if you do want people to refine their search so the matches are more appropriate, showing long lists can actually work well since they will be ignored, while the more inviting short lists might cause users to delay on scanning them. Now whether or not this will actually work is anyone's guess. I'd do A/B testing if I could.