Relevant Search In Your Application
Every application that provides some data storage capability has to have a search feature. It allows customers to find the required data quickly and with minimum effort. The following article contains several best practices on how to do it properly and make your customer’s life easier.
Order By Importance
Some information may be more important for your customers than the other. It is crucial to understand which information is important and show it at the top of the found results. Relevance calculation is a separate topic, but in general, every data record usually has several fields used most of the time (e.g., name or email). Application has to boost search in such fields and rate them higher compared to other less important fields.
Try To Predict
It is also crucial to understand why this information is important, so the application can predict the customer’s behavior and show him the most relevant results in advance. Prediction is usually context-specific. It means that it has to behave differently depending on the page used by the customer. For example, if a customer is browsing his income emails and uses a global search, then there is a good chance that he is looking for another email or a person. Alternatively, a customer may need to find some information related to the open email content, so the application may try to analyze it, get some keywords, and use them to predict what the customer may be looking for.
Less Is More
Sometimes a customer may look for general information, and so the application may find too many relevant data records. What to do in this case? First, the application can show a hint that there are too many results and propose to make a more explicit request. Then the application may ask to fill some filters to decrease the result set and make it easier to process. Finally, the application may intentionally hide results with low relevance and display a button to show the full result set.
Ask Your Customers
You should also ask your customers from time to time to provide feedback about the application search. There is a best practice to collect statistics related to search requests performed by the customer. Having this information, you may ask a customer whether he is satisfied with the results and if he is not — what was the issue. After that, you can analyze the results, find the problem, and fix it.
Customers may also give some hints on new features they may need. Just ask what they would like to improve n search behavior, group results by functional areas, and then which of improvements will help most of the customers.
Rinse And Repeat
It would be best if you did not stop there. You have to continually analyze customer needs and demands, see how you can help them, and bring new solutions to life. It is an iterative process — each iteration may start from requirements analysis, followed by the implementation. Finally, customers’ feedback can show you results and give hints for the following iterations.