design:search

Author: Michelle

Searching

When you search, you start from a blank screen (like Google) and enter a search term to return a list of results based on that search term.
This is a more general way of finding data as the results returned will display all data that match your search query.
If your design conflicts with a user's preconceived mental models for searching, they will have a hard time interacting with your system.

A Search Feature is made up of 3 components:

Search should typically be found on the top right of the screen and should be in the form of a search box, not a link.
The search box should be wide enough to accommodate a search term otherwise this affects usability.

Search lets users bypass the intended navigation of a site. In most cases, a user simply wants to find the thing they are looking for. This is the paradox of the active user in action.
The paradox of the active user is basically that the user never reads the manual, they would rather get started using something immediately.
This is a paradox because the user would end up saving themselves lots of time in the long run if they spend some time at the start reading the manual or watching training videos.
With this paradox in mind, it's important for designers to design for the way users actually behave; i.e. in an irrational way.

Search is also used as an escape option for a user if they get lost on your site. This is why it's important to have a search feature on every screen.

Typical users are not great at query reformulation. If they don't get the results they are looking for on the first try, chances are they will give up and go somewhere else. This is why it's important to increase the chances that a user will find what they're looking for on the first attempt.
Users will almost never scan beyond the first page of web results. This is why prioritizing the search results is important to consider. Examine the most common queries in your logs to determine the best strategy for this.
Offer advanced search only if the user needs it. “Didn't find what you were looking for? Try Advanced Search”.

When you are filtering, you are refining/reducing the number of items from a full list of data
This is a more accurate way of finding data as the filters are selected by the user from an existing list of filters which helps to refine the data especially if the user doesn't know the dataset very well.
Filters eliminate results that do not specifically match the selected criteria.

This depends on user intent and page speed:

By being able to determine the user's intent when looking for something, you can determine the most efficient type of filter to use (interactive or batch filtering). This creates a smooth experience and allows the user to find what they're looking for.
Page speed is also a factor in determining which type of filter to use.

Interactive filtering occurs when you select one filter and the list of results immediately updates. This is good for users who are in an exploratory mindset and want to learn about the search space/data available to them.
If used incorrectly i.e. if the user has a different mindset as in they know what they want to filter for or if the site is slow to retrieve the data, then this can become disruptive for the user as the results will refresh each time a filter is selected.
A disadvantage of interactive filtering is that continuous page updates can become visually distracting. To overcome this, dim the results set and show a progress indicator before displaying the new results.

Batch filtering occurs when you select a number of filters before the list of results updates. This is of benefit to users that have a clearly defined set of criteria.
This is also good for slow sites where data retrieval may take some time.
This is the best option to implement on mobile devices.
Batch filtering is the safest option as it allows the users to tell you when they're done selecting filters.

Another way of implementing filters is by allowing the system to wait 1-2 seconds and if there has been no user input in that time, then refresh the results.
With that in mind, the system should still indicate with a progress indicator that the first input has been received (immediate feedback to the user).

Imagine you are looking for a pair of Bose Noise Cancelling Headphones.
You will probably go to Amazon and type in “Bose Noise Cancelling Headphones” into the search bar. This is performing a Search.
The results returned will match Bose, Noise Cancelling, and Headphones. The first page or so will return the most accurate products.
However as you progress through more pages, you'll find different brands of headphones other than Bose, or headphones that aren't noise cancelling.
If you were to perform a filter based on the category headphones, the brand Bose, and select Noise-Cancelling as a feature, you'll find a more accurate set of results that match those filters.