Why you need a search strategy for your EPiServer site
While most EPiServer sites have some kind of search feature, many site owners and developers underestimate the importance of defining a clear search strategy.
What is a search strategy?
It's the sum of decisions you make that allow visitors to search for your content. A search strategy usually involves a bit of content/navigation analysis, SEO, usability and technical implementation. Your site may have the coolest design, the sleekest navigation and the smartest copywriters ever, but if your visitors still don't find your content, you have a problem.
You just might be guilty of having a bad search strategy.
In this blog post I'll focus mainly on internal search (or on-site search), and what part it plays in your search strategy.
When do you need a search strategy?
Consider the following scenarios:
You're building a new site and content from scratch. You want your content to be searchable and rank well in external and internal search.
You're moving content from an old site to a new site. You don't want to lose any of your hard-earned link juice or ranking, and you don't want old links to give a 404.
You have a commerce site listing lots of products or services. You want visitors to discover items to purchase by cross-searching categories, comparing specs and getting suggestions on related items.
You have an enterprise site where the content comes from many different sources and systems, both public and internal.
You have lots of archive content which is kept online for historical reference, but new versions are published regularly. You don't want outdated content to overshadow new content in search.
Your site relies heavily on media files like pictures, audio or videos. Most of the important information about your content is stored as metadata.
You expect certain keywords to rank well in your internal search, but they don't.
If any of the above apply to your site, you need a search strategy.
Why search is often neglected
Having implemented a lot of EPiServer sites for a wide range of customers, we've seen many different attitudes towards search. Recognize any of these?
THE UNICORN: The customer has a very clear vision about how content and search should overlap, right down to preference of on-site search engine.
THE PUPPET: The customer has no strong opinions on the matter, but has been told by consultants what search strategy is about.
THE BLISSFULLY UNAWARE: The customer has a non-existant search strategy, or a random one at best.
The alarming part: The last scenario is probably true for at least 50% of web projects. Why? Because many assume search will magically work great out of the box. Sadly, most default CMS search don't.
In some cases, the topic of search doesn't even come up until a developer is about to start coding the search page template. A designer may have drawn a basic sketch of how the search page might look, but there will often be no in-depth description of the thinking behind it or variations for different modes.
Or worse: A designer has made detaileded sketches of The Ultimate Search Page, without checking whether that solution is possible with the money and technology available to the project. The customer now thinks they're going to get "Google on steroids".
Any developer can make assumptions and decisions about search. But those decisions should have been made long before you started coding. As an experienced developer you are expected to give good advice on which parts of EPiServer can be modified and extended, or which add-ons to buy, to support the search strategy.
Site owners and designers don't need to know the details of how to implement a search feature. But they do need to know why they want a search feature in the first place.
How important is internal search for your website?
For enterprise (organizational/governmental) sites, search and navigation usage will often be pretty evenly divided (and equally important).
Visitors will recognize the navigation since these sites have quite similar content structure, and also a large number of visitors will be people familiar with the organization. But because of the mass of information, most users will use search sooner or later.
For commerce (product/service/knowledge-driven) sites, the percentage of visitors that use search is quite high. Think webshops, job/property listings, user forums.
Search will be the visitor's preferred way of finding content rather than navigating. These sites often rely heavily on presenting and cross-linking metadata. A bad search experience might lead directly to lost revenue or a dented reputation, which is why most commerce sites spend a lot of resources on their search strategy.
For non-commerce (informational) sites however, visitors actually rarely use the internal search. Research has shown that the usage may be as low as 2-5%. For this reason, a few sites decide to focus all their efforts on content and navigation optimization, and give up on internal search entirely.
Don't go abandoning your search feature just yet, though. 2% might sound low, but if your site has 500,000 visitors, then 2% equals 10,000 visitors. Can you afford to ignore their needs?
Search is often a useful fallback option when visitors don't find your content through navigation. You should always have great navigation and content, but a well-designed search feature, even a simple one, will help your content reach those extra visitors.
How much should you expect to invest in search?
For an enterprise site, you will almost certainly need a commercial search engine specializing in enterprise. Your content will come from web pages, databases, e-mail, file systems, document management systems etc, and you need a search solution that will index and unify content across all these integrations. These are usually the most expensive search engines.
For a commerce site, your success depends on a great search experience. You will probably want advanced features like stemming, synonyms, best bet, adjustable weighting, etc. In these cases you should probably go for a commercial search engine instead of building the features yourself. There are several suitable search engines available at a reasonable price.
For a non-commerce site, a simple search feature will be adequate, but if you're planning to use EPiServer's default search, there are several (cheap or free) plugins available that will improve both the indexing and the output of the standard search. I'll talk more about these in an upcoming blog post.
Will your internal search automatically be awesome if you buy an expensive, feature-packed search engine? No, not necessarily. Unless you clearly define your search strategy and understand your goals and limitations, throwing money at the problem will not magically fix it.
Which option you choose will always be a matter of Customer's Needs vs Cost vs Value. In an upcoming blog post I will try to give some pointers on how to choose the best search solution for your website.