By Sarah Crossman and Adam Kempler
Our Drupal Client Meetups provide techniques and tips on how to optimize your Drupal CMS. GovWebworks team members Adam Kempler (Lead Drupal Developer) and Sarah Crossman (Director of User Experience) discuss how to move beyond general best practices to create more tailored site search experiences based on user needs and content type.
An effective site search is near the top of the list when we ask clients, stakeholders, and audiences what they want to master with their Drupal CMS. While the out-of-the box site search solution that Drupal provides is a good starting point, users often have needs that benefit from a more customized search experience. In this Drupal Client Meetup, we’ll discuss how to make standard site search strategies more effective.
So, what do we mean by an “effective site search”? It’s about getting the most relevant results at the top of the search results page, and making sure those results have enough context so users know they’ve found what they’re looking for. This is an ongoing goal that involves continual testing, analytics, and refinement.
Effective search is also iterative, which means you don’t just implement it once and forget about it. We’ll look at some tactics for improving search, starting with a plan to identify what site content should be searchable to drive your search strategy as well as architecting for search and the UX of search. These are intertwined—you can’t have good architecture without good UX, and vice versa.
At a high level, effective search is about not just getting relevant results, but presenting them in a clear, informative way. Whether you’re starting a new project or working with an existing one, the following strategies can be applied at any stage of the website lifecycle.
Maintaining an effective site search
Let’s ground ourselves in the importance of site search. Site search is crucial because it’s a common habit for website users. Studies suggest that up to 90% of users typically use site search, but only about 50% actually find success with it. That’s a significant gap, and it shows there’s a big unmet need. Site search is one of the multiple pathways we aim to optimize for site visitors. Some users will come to your site with the intent to use site search, either because it seems the most direct way or because they have a particular need in mind. Some users may prefer to navigate through menus or the homepage, especially if they’re not sure what they’re looking for. Others may have a more targeted approach.
Site search isn’t just another item on the Jira backlog. It’s a product that requires thoughtful planning and continuous improvement.
The following eight topics are the keys to maintaining an effective site search:
1. Test plan
One of the key tactics for improving search is developing a test plan. As we talk about the test plan, we want to be sure we’re identifying what should be searchable on the site. This could be a detailed document or something more high-level, like a spreadsheet, but the important thing is that it’s documented. This search plan will drive the testing strategy and help us reduce noise by identifying what should and shouldn’t be indexed.
For example, if the content type is “Forms,” we’ll identify how it will be searchable. Is it a section-specific search (searching just within forms) or a sitewide search (searching all content)? These decisions matter because they affect how content is indexed and what users will see. As decisions are being made, it’s important to continually revisit the personas of the target audience. This ensures we’re making informed decisions about search based on the people who will be using it.
By orienting the test plan around the target audience, it helps make the testing process more real, and ensures we’re focusing on what these audiences are looking for and need. This way, we can apply personas to test the right things, target the right improvements, and prioritize the correct use cases for optimizing the site. Just like in UX, where we layer best practices with research to understand your unique users, testing for search allows us to understand how to better meet the users’ needs.
Remember that a test plan requires continuous refinement. For example, Google has been around for decades, and they still continuously work on improving their search algorithms. It’s not just science; it’s an art that requires collaboration across teams, including developers, UX, domain specialists, and client stakeholders.
2. Content types
Next we want to identify the types of content involved in our search content. Drupal uses content types such as Legal Form, Form Instructions, News, Events, Conservation Area, and Road Project. Additionally, we want to identify the types of media that should be searchable, such as the files attached to a Legal Form. Understanding this process helps us think about all the related types of content that might contribute to or be exposed when searching for a form.
For example, it’s not just the Legal Form content type that will appear in search results. The title of the file attached to the form, instructions on how to fill out the form, or other related content could all appear when users search for a form. By understanding these content types, we can define which fields should be searchable, such as titles, descriptions, and file names. We also need to identify which taxonomy fields should be searchable, not just as filters but in full text.
On the Legal Forms page, we have taxonomies like “legal situation” and “legal action.” We need to document how these taxonomies contribute to search results and ensure they’re indexed appropriately. For example, if a form is related to divorce and tagged as such, it should surface when a user searches for “divorce.” This process is also essential for legacy content, migrating an existing site over, or refining current content. It ensures we create documentation for approaching search effectively.
For another example, consider searching for legal rules. A rule in a court system website may be made up of multiple content types, such as the rule itself, sections, and changes. Understanding the relationship between these content types is key to effectively indexing them. For instance, when a user searches for “divorce,” do we need to index the rule content, section content, or both? These choices affect how the search results are displayed and need to be documented.
The process of planning and identifying content to be indexed sparks critical discussion about how we can make search results more effective. This drives the next step, which is testing.
3. Testing
Testing is crucial because, while we can build a functional search system, it’s important to ensure it provides high-quality results. How do we know the search system is effective? We need to document the expected results for each type of content and audience in a Test Plan which should be run regularly. We need to test the search results regularly, just like we would with a search plan.
For example, let’s say we’ve created a search plan for “projects,” like road projects. We would define test cases for specific search terms, such as “Chester Bridge,” and document where these terms should appear in the results (e.g. the top result when searching on the Projects page, and within the top 5 results when doing a full-site search). These test cases need to be repeatable and structured to ensure consistency.
Furthermore, when testing content types, we should also consider the variations in how users search. For instance, someone searching for “Chester” may still expect to find the Chester Bridge project. By planning for these variations in search terms, we ensure users can still find what they’re looking for, even if they search with partial information.
Search results can change over time as new content is added and old content becomes outdated. Testing should be conducted regularly to identify this “degradation of search quality”. For example, after a few months, a search result for “Chester Bridge” may not appear at the top of the results if the project is completed. This highlights the importance of continuously testing search results to ensure that outdated or obsolete content is no longer prioritized.
Testing is something that can be overlooked. It’s easy to assume that search is a technology that works on its own, but in reality, we need to validate it with real content and user stories. This allows us to identify areas where search could be improved, even if you’re not yet ready to make those changes. The testing process can also help build a roadmap for future enhancements.
4. Metadata
Metadata is another key element of effective search. Metadata provides additional context for content. It could be visible to users or used behind the scenes to influence search results. For example, if users are searching for a project like Chester Bridge, metadata like the route or county could be displayed to provide more context. Filters or checkboxes can also help narrow down results based on metadata, improving the user experience by reducing noise in search results.
Some metadata, such as taxonomies, are used administratively behind the scenes to track and manage content, such as identifying stale content. Even if not displayed directly to users, this metadata can affect how search results are generated, ensuring that relevant content is prioritized.
Documenting your metadata requirements in the Search Plan is critical for ensuring that search results are effective and meet users’ needs. You want your content to be prioritized in search results, so it’s important to consider factors like the last updated date. Often, older content is less relevant than newer content, though this can vary based on the type of content. For example, a press release from 10 years ago might be less relevant than one created in the past year. For most content on a website that is one or two years old, it wouldn’t make a difference. Think of all the information on the judicial court. That information doesn’t really change unless the laws change.
This metadata is applied differently depending on the type of content, and the key is ensuring that the metadata is carefully thought out, standardized, and adds value. For instance, when tagging content with counties, it’s important to avoid inconsistencies in how you tag the same county. If you have Delaware County listed multiple ways (e.g., “Delaware County”, “Delaware Cty”, “DE County”) it ncan cause issues in filtering or searching, leading to multiple options for the same county or title. It’s important to create a standardized taxonomy to avoid confusion and ensure accuracy in your search results.
5. Facets
Facets, or filters, are tools used to narrow down search results. These user facing checkboxes on a website help users drill down into search results based on different categories, such as content type or location. For example, if you’re searching for news and select a specific district, only news related to that district will be shown. This eliminates the noise and ensures that users get relevant results. Facets can either be manually selected or automatically applied based on user interaction with the search results.
6. Weighting
Another important concept is weighting results, which involves prioritizing certain metadata or types of content to appear higher in search results. Weights can influence which results show up at the top of a page (not just if they show up at all, like a facet).
For example, you might weight titles higher than body content so that search results with matching keywords in the title appear first. Similarly, newer content could be weighted higher than older content, or featured content might be given a boost.
However, weighting content can be challenging, especially when dealing with a variety of content types and taxonomies. The goal is to find a balance and prioritize high-quality content, but it takes time, iteration, and ongoing adjustments to get it right.
7. Indexing
Indexing files is another aspect to consider. For government websites with large amounts of media, like PDFs and Word documents, it’s important to determine whether or not to index those files, as well as the text within those files. Full-text indexing can make files searchable, but it can also create noise if the content isn’t particularly valuable or relevant. For instance, the text of PDF versions of legal forms may not contain valuable content that would help surface it during search. Instead, the title and metadata such as tags would likely be more valuable.
8. Analytics
Analytics play a crucial role in improving search effectiveness. By tracking what users are searching for, you can identify content gaps and areas for improvement. Tools like Google Analytics or Google Looker can help you understand user intent, allowing you to refine your search results. For example, if users are frequently searching for terms like “camping” or “food stamps,” you can create test cases to ensure that the most relevant results appear at the top.
Many of you are probably already capturing analytics on user behavior, but capturing search-specific data is just as important. By tracking search terms and user interactions with search inputs, you can refine your approach and inform the planning and testing stages.
Additionally, you can track searches across different sections of the website to understand how users interact with specific content. This is particularly important for government websites, where users may have very different needs and are navigating a diverse set of content. Understanding how search is used across different sections allows you to fine-tune search functionality and make it more user-friendly.
In summary
As detailed in this Drupal Meetup, Drupal’s default search is a good starting point, but customization is often necessary to meet user needs with effective site search. Crafting effective search results involves considering metadata, weighting, and filtering to prioritize high-quality content. It’s a complex process, but with careful planning and ongoing refinement, you can significantly improve the findability of content on your site.
We reviewed key strategies for improving search, including:
- Developing a structured test plan
- Defining searchable content types
- Implementing metadata for context
- Refining search results through facets, weighting, and indexing
- Regular testing and analytics to ensure search quality doesn’t degrade over time
By continuously optimizing search architecture and UX, your agency can enhance website content findability, making it easier for users to efficiently locate relevant information on your Drupal website.
Learn more
Want to learn more? Contact GovWebworks to schedule a free consultation on how to improve your Drupal site search
- Fill out the Drupal Meetup survey to request a meetup topic
- Read Tending Your Digital Garden for more tips from Sarah Crossman on maintaining accessibility in your Drupal content
- Read Adam Kempler’s blog on Conversational Search Assistants
