Searching the web is something we often take for granted. It’s hard to remember what life was like pre-Google and it’s evident that we are very reliant on how easy it can be to find information online.
Here at Yellow Pencil, municipalities are one of our main client groups. I thought it would be interesting to perform some analysis on how different cities across north America are using search within their website.
I’ll be offering the audit results from both a technology platform angle as well as user experience perspective. I’ll be using three common search terms in this study: “pet license”, “parking ticket”, and “property tax”.
I’ll also be evaluating how the search experience works on mobile; whether a separate mobile site or responsive design is available to make navigating search results easy for users.
There are a number of terms that I’m going to use in this post while discussing search, so before I get started here are some helpful terms to know:
This allows a website visitor to choose what type of content they are searching for by applying multiple filters. A common faceted search allows the user the option to view only HTML content, or only documents from a set of search results.
Google Custom Search
Google Custom Search allows you to create a search engine for your website and display via an embedded widget. This is an entirely free solution, but it’s unable to provide customized search results. Google branding must be displayed alongside results. Also, this search is not physically hosted within your environment, and uses Google’s infrastructure.
Google Site Search
Google Site Search also allows you to create a search engine for your site and display via an embedded widget. It differs from Custom Search in that you are able to customize the display of search results and display of ads. This is a paid service, and Google has tiered pricing available depending on the number of search queries per year.
Google Search Appliance
This is a physical device available from Google that is premised alongside your web servers in a data centre. The search appliance is able to index both publicly accessible pages as well as internal pages that require authentication. The search appliance learns and adapts results based on the patterns of the website’s users. This ensures that relevant pages are returned in search results.
For this post, I looked at the five largest American cities and Canadian municipalities (by population):
While it is unclear what type of technology runs New York City’s (NYC) search, the search itself is powerful and offers several facets to users such as events, 311 content, or PDF documents. NYC’s search sometimes highlights a recommended page based on the search term. For all three sample search terms, the first page returned was the page a user would be looking for, and subsequent pages in the results were also relevant.
In the case of the “pet license” search, the dog license page was the first result. The search was able to understand that “dog” and “pet” have similar meaning.
From a user experience perspective, it’s easy to enter search terms on the site, and the search results are displayed very clearly. A brief description is offered for each page, and the descriptions help decide which page the user actually looking for.
The whole nyc.gov website utilizes responsive design, so it is incredibly easy to navigate on a mobile device.
The city of Los Angeles is using a Google Custom Search for their entire website. When searching a term, a sidebar containing media releases is displayed as well. While this offers a second facet when searching, it isn’t user selectable and is limited to one type of content.
When searching the sample terms, the first page of results contained the term being searched. However, a lot of similar or irrelevant content was also displayed. Page descriptions are not consistent across the site, and in some cases, were not helpful.
While the website itself is responsive, the search results are not. This is due to the way Google Custom Search has been implemented. The user is forced to pan left and right to read the title and description of a search result.
Chicago is using a Google Search Appliance, but there are a few issues with their particular implementation. While they do offer a faceted search for both documents and web pages, for "property tax" and "pet license" searches, the relevant pages appeared 3rd or 4th in the search results. The page descriptions that are displayed in the search results are not the actual meta tag descriptions, or actual page content. It appears as if their search returns the navigation menu as the description of the page. The resolution of this configuration error could vastly improve their search results pages.
Autocompletion of search terms is utilized when entering search terms, but it does not narrow the user down to a more relevant search term. This autocompletion offers little assistance in finding specific content.
The website is not available in a mobile format, or within responsive design, making it difficult to use on a mobile device.
Houston is another city using a Google Custom Search via an iFrame within their website. An iFrame is a way to embed a ‘window’ containing another website. While search results were accurate, page titles and descriptions were inconsistent across the results. In the case of the “parking ticket” search, a specific page was highlighted and recommended above other results. This page suggestion provided the content we were looking for.
The website is not responsive and does not offer a mobile version. As mentioned earlier, this makes it difficult as a whole to use on a mobile device.
When searching phila.org, the user is directed to the Google Custom Search directly on Google’s domain. Faceted search is offered, allowing the user to search for 311, contracts, public safety, or all website content. However, it is not immediately apparent that these facets are available due to the overall visual design of the search results.
The search results were very accurate, and offered relevant descriptions.
Functionally, a phila.gov search provided the relevant information, but there is a lot of room for improvement in the overall user experience. The average user may not be aware of the powerful search options on the site that are available via faceted search, and the results themselves are not mobile friendly.
Toronto is using a Google Search Appliance, and offers a second facet to narrow a search to 311 Knowledgebase content. Related searches are displayed in a sidebar, but aren’t necessarily related to the term being searched. PDF documents are listed among the search results, but are marked as a PDF for those who want to avoid opening a PDF on their mobile device.
The search results have highlighted colours for any page titles or descriptions that contain the search term. While this can be helpful, it can also be distracting when trying to discern the search results. The scanability on the search results page isn’t the greatest. This can be remedied with a better vertical rhythm and more consistent line heights.
Toronto doesn’t offer a mobile version of their website, or responsive design.
Montreal is the only city in our analysis that is using Coveo for search. Coveo was recently listed in the Gartner “Magic Quadrant for Enterprise Search” alongside Google and other technologies. Search results using our example terms were relevant. One issue that was noticed was the inclusion of HTML code within some page titles in the search results. The pages themselves were actually relevant, but may be skipped over by the user due to the strange appearance of HTML code.
Coveo search offers many parameters on the site, but a website visitor may be overwhelmed. One of the parameters when searching is “Source”, and one of the sources is “Oracle Portal”. For most website visitors, it is likely that they wouldn’t know what Oracle Portal is, and how it is relevant to their search.
The search results page doesn’t align with the overall visual theme of Montreal’s website, and does not offer a mobile or responsive version.
Calgary made waves in 2011 when they launched their search-based website. When arriving at calgary.ca, there is limited navigation presented to the user, and it is expected they will use the search field to find what they are looking for. Paul wrote about this when the site first launched, and you can read some of his thoughts here.
Calgary’s search-based website is powered by a Google Search Appliance, and Google has published a case study that reveals some hard data on how Calgary increased engagement and returned more accurate search results.
Autocompletion when entering search terms is snappy and relevant. For certain search results, “best bets” are displayed at the top of the results page. These “best bets” are relevant, and are easy to visually identify in the search results.
Faceted search is available for webpages, documents, and news, allowing the user to further define what they are looking for. The search results are clean, easy to scan, and offered clear page titles and descriptions.
Calgary offers a mobile version of their site, making it very easy to navigate search results. However, the rest of the website is not entirely available in a mobile or responsive format which, in some cases, leads to a jarring experience.
Ottawa is using a Google Search Appliance within their responsive website. Popular searches are displayed alongside the search field, allowing the user to complete these popular searches.
Page descriptions are inconsistent in the search results, and it becomes difficult for the user to discern which page may have the content they are looking for. Faceted search is not offered, but visitors are able to sort results by date or relevance.
Several page descriptions within the search results included breadcrumb navigation as content. An audit and overhaul of page description content on the site would aid in the overall search functionality of the site.
Ottawa utilizes responsive design, and results rendered properly on a mobile device.
Edmonton is using Thunderstone search technology on their website. Popular searches are available in a dropdown menu, allowing visitors to see what others are searching for.
The search results of the three sample terms had appropriate content returned. Results are easy to scan, and page descriptions are relevant and consistent.
It was not clear until digging into the Advanced Search that a visitor is able to search within different sections of the site and types of content. For more specific, less common searches, the user probably doesn’t realize that they are able to limit their results in this way.
Edmonton’s website is not currently mobile friendly, but the city is examining their mobile web strategy at the time of this post.
5 best practices for municipal search (and other search too!)
Provide helpful autocomplete
Autocomplete can guide your users to utilize meaningful search terms that are within your site’s vocabulary. It’s important that this autocomplete doesn’t suggest terms that will produce ineffective search results. In some of the sites reviewed, autocomplete was used but the user is bombarded with a smattering of autocomplete suggestions. This can be confusing, and isn’t helpful in narrowing down the scope of a search.
Ensure your search page is mobile friendly
A surprising amount of web traffic is coming from mobile devices, and even if you are utilizing techniques like responsive design, your search may not respond appropriately. Iframes can be difficult to integrate within a responsive framework, and will require a user to pan around within the frame if they are on a mobile device.
It’s also important to provide a seamless experience between the search result page and the rest of your site. If the user is being pushed to an off-site template that is not mobile friendly, they may abandon their search.
Limit PDF content
This is a valuable practice, not just in relation to search. PDF documents are difficult to navigate, especially on a mobile device. Certain search technologies can also have issues properly indexing PDF content, so keep that in mind when considering PDF’s.
Provide meaningful faceted search
Allow your users to search for content by a content type. Common content types include events, city council minutes, and bylaws. This allows the user to further narrow down their search to the exact type of content they are looking for. Note: certain search technologies may not allow you to configure faceted searches.
Populate your metadata
Populating the metadata of your pages not only helps your in-site search, but your external search engine optimization as well. Using a consistent format for your metadata will allow users to easily decipher page descriptions and scan for relevant results. Your search platform will use this metadata to index your pages and return relevant results to users.
If you would like to have us perform an audit of your site’s search functionality and usability, contact us!