If you're in the midst of implementing a new site search project or just starting your research on the topic, this blog post is for you. We'll walk through a brief primer on site search best practices including examples for how you can implement site search.

What is site search?

Site search is a search that’s limited to a specific website or related domains. This can include public websites or internal site search such as an organizational knowledge base or intranet. Unlike a web search engine such as Google, Bing, or DuckDuckGo that traverses the internet, site search is customized to a specific domain(s). 

Google is synonymous with search, so why doesn’t everyone simply use Google search for site search, too? Google’s site search products are ad supported and offer limited customization, so most customers have turned to newer, ad-free site search engines that can be customized to a business’ website and audience. Google is terrific for web search, but a provider like Sajari is better for building custom site search experiences. 

The importance of site search engines

Well-executed site search can not only improve your visitor's experience, it can actually increase conversion rates, lower support costs, and build customer loyalty.

When one company simply made their on-site search more visible…

• Site searches climbed by 439%
• Visitors stayed on the site 110% longer when they used search
• Page views increased by 45%
• Goal completions increased by 242%
  

However you define site conversion — sign ups, on-site sales, product ratings, lower customer support costs, etc. — on-site search can contribute by helping visitors find what they need faster. 

Search is well-suited to sites with large catalogs such as online retailers. But even smaller sites or different use cases — such as government search shown above — can benefit greatly from custom site search. 

Before we dive into examples of site search, let’s look at your options for implementing a search engine on your content management system or custom site. 

SaaS search solutions

Software as a service (SaaS) search is popular because it can typically be set up quickly and requires no infrastructure. SaaS is an attractive option for website owners because the service provider handles search engine maintenance. SaaS solutions are built on elastic infrastructure that can scale with the business’ needs. 

For developers, SaaS services typically include a robust set of search APIs or plugins for integrating search with your content, have built in search metrics for analyzing success, and are easy to extend across domains and search collections. 

See our Site Search Buyer’s Guide for 2021 for what to look for in a SaaS search provider. 

Cloud search solutions

Cloud search solutions include native tooling developed and delivered through AWS, Google Cloud Platform (GCP), Microsoft Azure, and other cloud providers. They offer great elasticity and flexibility, but they also demand higher maintenance and support, and non-technical users will not be able to easily optimize search results. Cloud providers typically don't include crawling, so you would need to work out the data sync and processing yourself. Advanced features such as machine learning or personalization are typically sold as separate services that require custom engineering.  

There’s valuable performance upside to co-locating your search engine with an app or site in the same hosting region. If you have the engineering capacity and you're building site search for an application already running on a cloud provider, this is a good option to consider.

Open source search solutions

Open source solutions (OSS) offer a good deal of flexibility for engineers to build custom solutions or tightly coupling search with an application. Open source is free to use, but not cost-free: it requires the most setup, management, and maintenance of all the hosting options. For instance you may need to configure your own web crawler. PDF and DOC indexing may require yet another service. Upgrades can be infrequent depending on your team’s availability or engineering policy. With OSS, major releases may not be backwards compatible. 

The bottom line whether you choose SaaS, cloud, or open source is that your search solution should be highly customizable and help you improve search relevance and results.

Crawlers, APIs, and indexing

The first step to adding site search is site indexing. There are two ways sites are indexed: crawlers and APIs.

• Crawlers build an index of your site periodically and are great for simple sites.
• API indexing is more suitable to applications where data structure and immediacy is of great importance, such as e-commerce sites.  

If you make edits to your content, you want your search engine to reflect those changes immediately. However, some search engines can take more than 24 hours to update a site index. A best practices is to select a search platform that can handle instant updates to the index.

Site search engines should also have a way to omit content. For example, you may want to exclude certain pages or certain content types to drive more relevant results. 

Custom site search examples and best practices

Search overlay

A search overlay displays results on the same page the user is searching from.

Unity has built an overlay that searches and returns results from across multiple domains including their website, support, and community pages. Unity also uses search suggestions (also called autocomplete) that appear as the search term is typed in the search box. 

Search results highlighting

Highlighting the query phase in search results is a nice touch for searchers. Bonus: in this site, Heap provides filters beneath the search box for visitors to limit results to their Developer Guides, API Reference, or other web pages. 

Filters and faceted search

The LA Times has a massive database of articles. To help users navigate them, they have included filters on the search results pages.

Companies can apply filters and facets to search. Facets are similar to filters insofar as they narrow results, but they’re dynamic and can change depending on the context of the results. Dynamic faceted search is shown in the example below:

Instant search with visual results

Instant search and search suggestions — displaying results as a user types — is a site search best practice. Visual results give users a preview of the results to help them narrow down the search. “Sure,” you might be thinking to yourself, “that makes sense for an e-commerce store, but what about other sites?”

B2B sites can also display site search results more visually as shown below — Zapier is using logos in the autocomplete suggestions to help visitors discover results more visually.

Handling search typos and misspellings

An invisible and often overlooked UI element to site search is how to handle typos and misspellings. Somewhere between 10-25% of searches can contain a misspelling! Including typo tolerance of some flavor is a best practice for site search. 

Here are two examples (below) in how the LA Times and The Guardian handle the search for “nutrition” that was misspelled. By displaying corrected results, The Guardian offers a much better user experience versus LA Times “no-results” experience. This is an obvious typo, but less obvious misspellings happen which can lead users to think the site doesn’t have an answer to their search. 

‍

Best practices for site search optimization

How do you know if your site search is improving conversion rate, lowering support costs, and/or improving user experience? Search metrics are a great start for search optimization. Site search data such as low click-through rates (CTR) could tell you that your site is delivering poor or unhelpful results.

If you can connect your search engine to site search data such as site conversions, sales, support tickets, etc., you can more easily correlate search results and search activity to business goals. 

Nowadays you can design site search with advanced functionality such as:

• Result boosting, either through rules or via the search algorithm itself, to improve visibility of assets, e.g., you may want documentation to be more visible in search results than other pages
• Personalization based on user history, gender, location, or other metadata 
• Machine learning to automatically improve results based on metrics from past searches and results
• Synonyms. You may offer “sneakers” on your site but your users could be searching for “running shoes.” 
• Search operators that allow users to type advanced query parameters. For example, you could give visitors a way to input Gmail-like filters in the search query.
  

These are just a few of site search tools available from different providers which you can use to customize your search experience. For more ideas on improving search relevancy, check out our article on 14 Ways to Improve Site Search.

Site search tracking and Google Analytics

Many site search providers include built-in analytics. You can also use Google Analytics to monitor search queries and build site search reports. We offer more full-featured metrics around search trends that can be connected to business data to actually improve site search results based on website conversion. Whatever search reporting solution(s) you select, it should help you optimize user search experience.  

How is site search related to SEO?

The topic of site search and search engine optimization (SEO) often go hand-in-hand. Optimizing your site for Google search can also help your on-site search engine. Techniques such as adding alt tags to images, adding anchor links on pages, or cross-linking can help your site search engine learn what’s important.

Try our free Site Search Health Report to see how well your site is optimized for search.

The bottom line

Ready to up-level your site relevancy and user experience with brilliant site search? We invite you to try Sajari. With instant indexing, search design builder, and a search reports builder, Sajari can help you get site search up and running in minutes!