We think it’s fair to say we are experts on site search. We have over 9000 customers large and small, including some of world’s top brands like Under Armour, Lacoste, Medium and Twitch.

Why, then, are we calling this the almost ultimate guide?

For one thing, because site search is a complex topic. Search UX is studied by research firms and institutions. Relevance alone is a book worthy topic. The list goes on.

Secondly, search is built and used by a wide variety of folks, from developers and product technologists to merchandisers and digital teams. Writing a guide for different search users and experts would look very different.

Lastly, search is specific to your industry, use case, and the complexity of your project. We’ve been providing developer documentation search, but also highly rich and personalized experiences for e-commerce giants. Writing a guide for each use case could make for several tomes.

The goal of this post, however, is to give you an essentials toolkit that will empower you to:

1) Understand the most important aspects of site search

2) Be well equipped to assess the tools you are using or plan to use, and

3) Identify areas of opportunity to improve your site search — and its outcomes

With that said, let’s dive in.

What do we mean by site search?

First, let’s settle on the terminology.

Our working definition: Site search is the functionality that enables users to search a website’s or app’s content or product catalogs with speed and relevance.

For more than a decade, Google has been serving blazing fast and super relevant as-you-type site search suggestions (also known as query suggestions). We are so used to this functionality that we don’t even notice it any more.

In a different use case, Amazon revolutionized the way we shop. To provide access to their huge catalog, they built a fast autosuggest dropdown menu and powerful dynamic refinements.

It is this type of search that we are focusing on here: it connects people with products, content and key pieces of structured data. It is fast, reliable, works on multiple platforms, and offers highly relevant results.

At Algolia, we usually call it “search” and have a separate, crawler-based solution  called site search). People also call it on-site search, internal site search, custom internal search, website search, but for the purpose of clarity, we’ll run with the most commonly used and well known term: “site search”.

Elements of great site search

What does “great” mean for site search?

You could provide the most exquisite site search experience for the user, but one that’s expensive and hard to build, and even harder to maintain, optimize, and analyze.

On the other hand, you could apply an out-of-the-box search tool that’s wonderfully easy to implement, use, and modify, but leaves the user experience lackluster, or, worse, irrelevant.

Great site search should be both relevant, rich, and engaging for the end-user, but also easy-to-use, agile and performant for the business.

Here are the key elements of great site search

For the user:

1. Relevance: surfacing only the most relevant results
2. Speed: modern users don’t wait; they bounce
3. Great UX/UI: to make it rich, beautiful and engaging
4. Discovery capabilities: so users can get inspired to read or buy
5. Omnichannel availability: so users can find what they need on any device
6. Personalization: so that users feel known, understood and efficient

For the business:

1. Analytics: so you can maximize business results
2. Merchandising/promotion capabilities: so you can control your offerings
3. Ease of use for developers and non-technical users alike

Site search relevance

A great digital experience is most directly correlated with high search relevance. Whether you are offering products, news, courses, or recipes, the better you are at determining your users’ intent and serving them relevant results, the better the user experience will be. Artificial intelligence plays a key role in modern relevance strategies.

Textual relevance

The name of a product, its brand, keywords in the description — what we call attributes — constitute textual relevance. Here are some elements of textual relevance:

Typo tolerance: reading correctly into the user’s intent even if they misspelled a word or used a stop word (the, and, at, with…).

Synonyms: offering users a synonym result (e.g., a parka when they typed in “jacket”). This includes prefix search (if your user searches for pants, then searches for “p”, “pa”, “pan” should show results for both “pants” and “trousers”).

Plurals: Making sure that plurals are accounted for (searching for “feet warmers” should display foot warmers).

Business relevance

To make for a great user experience, it is important to offer the users content or products they are most likely to act on. You can use your business metrics like seasonality, popularity, or click-through rates to tune relevance, and also use site search  to promote and filter items.

Regular tuning and optimizing

It is important to regularly assess opportunities to tune relevance. Turn to search analytics to review results for:

1) your most popular searches

2) searches with high-click positions (e.g., where the most popular item is in position 6 instead of 1)

3) most popular searches with no results

This way, you can identify searches that are not serving the users the right results,  and then adjust your relevance. It’s important to tune your search regularly, because your users’ needs and search patterns will change over time. It’s also important to do 18when there are bigger changes in your website content or product catalog.

Great UX/UI

In an ideal world, the majority of your customers will find what they want on their first attempt, and it will be in one of the first 3 search results. But if they are forced to hunt further, dissatisfaction will raise its ugly head (and with it, increased bounce and site abandonment rates).

Giving your users the ability to find what they are looking for — and refine it in an intuitive manner — is key to getting the most out of your search. Here is how.

Display results from the first keystroke

Just like in a real conversation, it is important that the end-user feels the UI react to their intent and queries. Don’t wait for the 2nd or 3rd keystroke to return results!

Search-as-you-type experiences encourage users to type twice as many search terms, which, in turn, leads to more relevant search results.

Display and highlight searchable attributes

Let’s say your customer is searching your catalog for comfortable loungewear, and looks for items made with spandex. If you simply show results with clothes containing spandex, but the user doesn’t understand why those results showed up, they may think they encountered a search error, and abandon your site in disappointment. This is why it’s always best practice to display and highlight searchable attributes — in this case, highlighting “spandex” in results.

Use federated search

Federated search lets you surface all of the great content you’ve built to your users. A user’s query can reveal results from any product or content catalog on your digital property. Take a look at how PubNub lets their users discover relevant results from different content sources: eBooks, blog, documentation, solutions.

https://www.algolia.com/doc/guides/solutions/gallery/federated-search/

Personalization

Consider this:

• 74% of customers feel frustrated when website content is not personalized
• 79% of consumers say they are only likely to engage with an offer if it has been personalized to reflect previous interactions the consumer has had with the brand.

Online giants like Amazon and Netflix have made personalized search and discovery experiences a must. But before setting out on a personalization path, it is critical that you understand your users, their needs and signals.

• Define what attributes of your products or content are the most important to personalize on: brand, color, type of product, size, author, a mix of them… Sometimes what matters is your customers’ purchasing history, sometimes it is their brand affinity.
• Decide what user actions are the most important for personalization: is clicking on “add-to-cart” more meaningful than a user opening a product page?
• Use A/B testing to assess the impact of your personalization strategy on your search metrics. Let’s say that you’d like to personalize search results based on both views (e.g., if a user has seen a product page, or a landing page), and clicks (what a user has clicked on). A/B test this strategy to see if it positively impacts your search metric. Then, as the next test, you might test personalizing results based on conversion data, like when a user adds an item to cart or makes a purchase, watches a video or reads an article, etc.
• Iterate! No matter what strategies you implement to improve user experiences, be sure to plan on testing, adjusting and iteration.  It’s almost impossible to get everything right the first time, so iteration is essential to getting actual business results from your search.
  

  See how Decathlon drove 50% higher conversion rates with personalized search

Omnichannel search

Customer journeys start with search. With over 3.8 billion global internet users searching for anything and everything online, great search must meet the user intent not only in an instantaneous manner, but anywhere the users find themselves: be it is on their phone, tablet, or laptop.

Mobile search

Building great omnichannel search starts with mobile. Users are on the go, more time-constrained, and easily distracted by notifications, so their experience on your app must be as polished and frictionless as possible.

However, mobile poses unique challenges, from smaller screens to connectivity challenges. Here are a few suggestions on ways to conquer them – you can find more in this eBook. 

• Make your search bar prominent and well-placed:

• Instead of  1-2 “rich” search results (with several attributes and an image), a great alternative is to show query suggestions: complete queries of what the user might have had in mind.

Make the most of the device capabilities.

Whenever possible, use:

• Wallet/Apple Pay/Google Pay rather than ask users to input a credit card number
• GPS to pre-fill the address
• Biometric sensors rather than asking for a password
• Voice, speech to text, and images for input

All of this reduces friction and provides a better user experience.

See how Lacoste used site search to increase mobile conversion rates by 62%.

Voice search

Until recently, consumers had to be in front of a keyboard or screen to build relationships with brands and make purchase decisions. The starting point was always a search box.

But today, conversational search—users interacting with technology and getting responses in the form of a natural conversation—serves as a conduit for contextually relevant and personalized interactions, including search and discovery.

Voice search affects everyone who has a website or a mobile or voice application.

This opens up a world of new possibilities for how humans interact with brands and businesses. However, successful voice search is fundamentally more challenging compared to traditional text search:

• Voice queries are more complex than keyword-based searches; users use a richer vocabulary through natural language.
• With voice-only interfaces, you only have one result to return to the user, which requires significantly higher degrees of relevance.
• Results must be delivered at what appears to the consumer as speed of thought, mimicking the pace and style at which natural language conversations occur.

Modern search and discovery technologies must meet these challenges to make sure that every flavor of the end-user experience is not only functional but enjoyable.

See how WW (formerly Weight Watchers) uses voice search to enhance the digital experience for their users.

Search, yes, but what about discovery?

Users typically come to your site or other digital properties to either:

1. Search, in an effort to find exactly what they are looking for (e..g., a pair of jeans in a certain brand, color and size), or
2. Browse, with a potential intent to get something, yet not exactly sure what (e.g., a business casual outfit that may include a pair of jeans).

You should allow users to start by either searching or navigating, then further refine their results by leveraging the other strategy in addition, for example offering facets after the user performs a search, or making facets searchable:

In other words, you should be offering them a unified search and discovery experience.

Search analytics

Site search should be measured and used as a lever for growing your business. Defining KPIs and then analyzing your search are critical first steps towards getting value out of your site search. Enter search analytics.

Search analytics should enable you to follow your users’ every search: results they get, no results, “empty” results, etc. It should help you model your users’ behavior by breaking down their searches into metrics like popular searches, click position, number of “no results”, and filter usage.

Search analytics will get you thinking more objectively about how to structure your product data, and how to configure your ranking and relevance.

You will see whether your products are properly described or represented, whether the correct products show up in your search results, and whether, based on the searches, you have too much of one product and not enough of another.

Merchandise your products, promote your content

Your business teams can align your search result strategy with your catalog, revenue targets, promotional campaigns, and marketing strategies. For example, you could:

1. Run seasonal campaigns:
   
2. Hide items from search results:
3. Pin items:

Ease of use for developers and business users

Behind every great user search experience are busy teams: developers who build or implement, then maintain the search technology, and business teams (product managers, merchandisers, e-commerce marketers and omnichannel strategists) who own search data and want insight into it.

• The developers want solutions that are easy to implement and maintain across platforms and devices, while still retaining control to fine-tune, evolve and innovate.
• The business stakeholders want to understand and adjust the user experience without having to understand the tech behind it or needing to rely on IT/developers to make changes.

Great site search means balancing the needs of both.

Want to get the most out of your site search?

Empower your business teams — digital strategists, product managers, merchandisers, marketers and content editors — to manage and optimize your search. For one thing, they are the ones defining growth strategies for your products or content. Secondly, they are most directly responsible for ROI and improving business results. Last but not the least, this will free up precious IT/developer team time that can go towards improving user experiences.

In summary…

While thinking about all these aspects of site search may be overwhelming at first, keep in mind that they can be tackled in phases, in the order that makes sense for your business and bandwidth. Think of it as an easy roadmap to success: each iteration will contribute to happier users and better business results. And the beauty of it is that, with site search, one directly translates to the other.

And if you wonder where your site search sits with all these elements, we have an easy, free self-serve Search Grader.

Grade your own search in 5 minutes!

Please don’t hesitate to give us feedback and suggestions on this post: hey@algolia.com.