We interact with search every day when we Google something. Google really was a transformational technology for the internet and has set the standard for search with its relevance and speed. With such a clear benchmark for search, how does onsite search (the search box that you have on your own website) differ and what makes a good onsite search experience? Let me walk through the different levels of site search and what you as a website owner should be aiming for.
There are two elements that fall into baseline expectations of search, the absolute minimum acceptable standard, but even some onsite search experiences fail to deliver on these.
The search results that are returned for a search query must be accurate and relevant. When I search for “study” do I get results that make sense and are the results in an order where the most relevant result appears at the top of the list. If we think about using Google, how often do we need to look past the first page of results to find what we are looking for? A good onsite search experience should be the same, the results should be relevant and we should rarely need to go past the first page.
Again I know I am harping on this but Google has set the standard for fast search. A search for the word “study” in Google will get me 5.5b results in less than half a second.
As a guide the “best in class” onsite search response time should be <200ms or 0.2 second (note most websites have less than 5.5b pages so we can give Google a pass on this one!). The search response time is the time taken for the search engine to run the query and not the time it takes for the user to see the results. The total time as seen by the user is highly dependant on the website and the complexity/size of the page.
If you are seeing a search response time of anywhere close to 1 second then there is definitely a problem with the search that needs to be investigated.
So the next level up is what I refer to as basic search. In reality this level is suitable for a very basic website with not a lot of content.
There is nothing more polite than a search engine who can correct your spelling without making you feel embarrassed. Below we can see how sharksmart.nsw.gov.au can handle an incorrect spelling of the word “drumlines”. It will display the correctly spelled word and allows the user to click on it to display the results for this word.
All too often however onsite search delivers an experience like this:
The user is now left to identify and correct their own error, to be able to proceed. If we are an organisation who is potentially providing information to a wide range of users who have different education levels or potentially be from an English as an additional language background this can present significant barriers for them. Spelling correction capability is fairly standard across search technologies and is a common expectation of your users.
Query completion refers to the suggestions that appear as the user types into a search box. Query Completion is great as it delivers a number of things for the user.
It makes it faster as the user doesn’t have to type in the full word to get their results.
It helps users get the spelling of words correct.
It promotes the most relevant results as the suggestions are filled with popular searches
It only suggests words that will have results.
Here we see an example of query completion on the tocal.nsw.edu.au website.
A synonym is a word or phrase that means exactly or nearly the same as another. When we work in a field we often get caught up in our own lingo. We can start to use words that a typical user might not. My favourite, or not so favourite one is “fine” a quick search on the WA police website for “fine” shows us a good example of the use of synonyms in search. From the example below the users has results for the word “infringement”, which is a synonym of the word “fine” that was typed in. This such a better user experience than trying to think up what the “correct” word would be to get the results needed.
If we have been able to get to this point, we already have a functional search experience for the user. The better category of search is going to deliver a really good search experience. This should be your target level if you have a larger website (100 pages+) or if you have an e-commerce or higher education website.
So “facets” is probably not a word you come across a lot and it really can be thought of as “filters”. For an e-commerce, higher education or very content heavy website, facets are critical. The image below shows how Amazon use facets (filters) down the left column to allow the customer to quickly get to the product that they are looking for.
In the Amazon example above we can see in a couple of clicks, the customer is able to filter results to show only hardcover books with a 4+ review. Another advantage of facets is that only options that have results are displayed, this ensures that a user will never be able to apply a filter that shows no results.
Rich Result Display
Rich Result Display refers to how each individual search result item is displayed to the user. An example of a basic display is shown below. We can see a clickable title and some basic information for each item. It’s functional, but it requires the user to read each result and take a little time to work out which item they need to click on.
A Rich Result Display uses images, buttons and formatted information to make it easy for a user to understand which result they need to click on. An example of a well formatted result is tocal.nsw.edu.au.
In the above example we can see publications, courses and web page results are all formatted differently to make the information much more usable and readable.
It’s even possible to display Call To Action buttons, like the Apply Now on a course to help drive conversions right from the search result, that’s the difference between a good and a great search experience.
So at this point we have a great search experience for our users. But there is something that can take our search to the best level. Just like a top quality wine, a top quality search experience should get better with time.
The leading search engines include the ability to learn from the interactions of users with the search results through machine learning. This video explains how they use machine learning.
Essentially based on how other users have interacted with search results for a particular term, the search engine will bump the popular results up the list, so over time the results get better and better.
The last item to make my most important features of search list is, analytics. Now this isn’t a feature that is exposed to your end user, but rather the data and information that you as the website owner have access to. It’s information that tells you how your search is performing and more importantly how your end users are using your search.
A leading search engine will provide you with a full suite of analytics, similar to Google Analytics, but specifically for your search pages. A good analytics interface will look something like this one below:
This report has a wealth of information and I will write a separate article about leveraging your search analytics in detail. My top three metrics that I recommend to start with are:
No result searches - this identifies where you are missing content or synonyms.
Search response time - ideally you want this at <200ms or 0.2 second.
Click rank - the position of the item that the user clicked on, ideally you want this to be 1-3.
The important point with your analytics, is to review these on a regular basis, so you are tracking the performance of search over time and actively improving it.
There are many other capabilities available in a leading search product that can and should be used on the best website, but I will save these advanced features for another article.
As web users we have the pleasure and often the pain of onsite search. Search is an amazing tool when used well and for e-commerce, higher education and content heavy sites, search can be the difference between an easy to use and completely unusable website. Like all things in life, choosing how much time and effort to invest in your search is a balancing act. This article has outlined the very simple baseline search that onsite search needs and how to prioritise investment to improve the performance, functionality and usability of your onsite search to achieve best in class results.