Hi, this is a late night post about a talk I did a little while back at Confoo and Web à Québec. I didn’t realize that my slides were popular online. Since they reached more than 10K views, I thought I would take the time to make a clear recap on my blog. There’s a Slideshare link of the talk right here if you want to see it or download it.
Search experience and performance are heavily influenced by non-visible factors search as logic, product data integration and user experience optimization. Bottom line: it’s hard for a competitor to steal your internal search features easily. They need to invest the time and audit your internal search to replicate things.
Out of many e-commerce websites, few provide a truly great internal search experience. This means that they are leaving money on the table. If we can’t find what you sell on your own website…it makes it hard to buy things from you as customers no?
A Few Things to Keep in Mind When it Comes to Onsite Search
- Gauging the competition’s search experience requires extensive testing and evaluation.
- You internal search efforts can’t be easily copied by competitors.
- Poorly performance search experience can be pretty…doesn’t mean that it’s useful to the customer.
Onsite Search vs Onpage Navigation
Website users were tested and people are pretty evenly divided between two big practices: some prefer on-page navigation to find products while 47% (or a little less than 47%) prefer to filter down to the product page. This data is coming from the Baymard Institute. For those wanting exact references, please check my slides. I have placed all references in my speaker notes.
Humans Have a Hard Time Searching for Things Online
65% of test subjects required 2 or more attempts to complete their search on a site. I have seen in my career internal site searches be so bad on e-commerce websites that customer service specialists preferred to look for products on Google.com or Google.fr. They simply could not rely on their own internal search tools to find their own products on their own website…
3 Big Categories of Search Intent
Humans have 3 big intents when they search for something. Their search can be navigational, informational or transactional.
- Navigational search intent means that the user is looking to reach a specific page in the website.
- Informational search intent means that the user seeks to acquire information that is located somewhere on the website.
- Transactional searches seek to perform a web-mediated activity such as an online order, signing up for a newsletter, etc.
Internal Search Best Practices for Beginners
Here are some tips on how you can go about improving your internal search:
- Make sure you place dummy text in the search box
- Place your search box in spot where the user can easily locate it – the norm is on the top right hand corner of the page
- Make the search box stand out in a different color
- Provide a simple search with a clear option for the user to navigate to an advanced search options pages
- Call it a search bar and not something creative. We need to know it’s a search bar!
Websites with semantic search engines have lower rates of cart abandonment
Wondering why that’s the case? Well, someone searching for “red sneakers size 9” is further along in the path to purchase than someone looking for “sneakers” in that search bar. While your basic internal search engine can handle a broad query such as “sneakers”, I’m not sure it can handle a sophisticated query that would entail figuring out a color (red) a type of shoe (sneakers) and a shoe size (size 9).
Your search engine should be able to do the following things:
- handle common misspellings
- provide a robust autocomplete feature
- handle long tail semantic searches
- highlight keywords in the search results
Internal Search Results Pages Best Practices
Here are 4 simple tips to get your started on internal search results pages. Here’s how to make them better:
- keep the search query visible in the search bar for your user
- provide the user with filterable options
- go beyond generic filters and adapt to your product features
- have clear product titles and descriptions available in the results
You should always fine tune the number and the presentation of search results. You should also fine tune how and what gets returned in the results. The reasoning behind this is that you can’t force users into a tunnel of limited search results. If they don’t find what they want from you, they will seek products elsewhere on the web. So let them check, uncheck, clear, refine and opt their way into a better search experience on your website. This will often lead to higher conversions as customers can truly find the products they seek to purchase.
Benefits of Investing in Internal Search Optimization
There are 4 big benefits that I can see:
- Search Analytics: get data on the way your users search for things on your website.
- Let the search behavior guide your site structure: this will help your information architecture and user experience.
- Let search behavior guide site content: are people looking for something that isn’t there? How about you craft content or provide the product they are looking for?!
- Using constrained search can help you also reflect a strict information architecture in the search experience. For some websites, it is important to ensure that visitors go through the same experience (for example custom ordered products).
How to Capitalize on Internal Search
Here are a few points to get you started:
- Avoid returning low relevance results. Audit your internal search to check what shows up when people search for your most used keywords.
- Map synonyms and misspellings to ensure that you show products when people are searching for them. It doesn’t matter if they can’t spell or use synonyms; give them what they want!
- Map symbols and abbreviations for the same exact reason as number 2.
- Allow users to iterate on the search query.
- Implement faceted search.
- Provide hierarchical breadcrumbs and history-based breadcrumbs. Show the user where the results sit in the site categories so that they can open up their search by category and let them return to their previous search results (aka by locating them in their initial context on top of the category context of your website).
Search Engine Optimization and Marketing Tips
- You should seek to provide unique, user-friendly (aka readable by a human) search results for long tail SEO traffic. Most of the time, people will ask you to no-index your internal search categories or results to make sure that they are not hijacking organic visibility while providing low quality content to users. Here’s the deal though: if you optimize your search result pages, they could also help you attract a lot of qualified traffic. It’s all in how you design those pages and what your SEO strategy looks like!
- Use those search result pages for pay-per-click ads. That’s right! You can also use those pages as landing pages for highly specific ad campaigns. If someone is looking for pink converse shoes and you happen to have a filter combination for that…This page should be used to help users convert.
- How about using Google to show a search box within a search results page?
- You can use those results pages to highjack search queries. When customers are looking for the shipping page, do not show them potential matches to their queries. Simply take them to the search page directly!
- You should seek to provide unique, user-friendly (aka readable by a human) search results for long tail SEO traffic. Most of the time, people will ask you to no-index your internal search categories or results to make sure that they are not hijacking organic visibility while providing low quality content to users. Here’s the deal though: if you optimize your search result pages, they could also help you attract a lot of qualified traffic. It’s all in how you design those pages and what your SEO strategy looks like!
- Use those search result pages for pay-per-click ads. That’s right! You can also use those pages as landing pages for highly specific ad campaigns. If someone is looking for pink converse shoes and you happen to have a filter combination for that…This page should be used to help users convert.
How to handle auto-suggestions
Autocomplete affects how and what a user searches for.
- Audit auto-suggestion from the website’s search logs.
- Machine learning should be based on the success rate of a query.
- Filter out duplicate suggestions.
- Allow users to iterate on auto-suggestions.
Autocomplete design patterns
Aim for 6 out of these 8 things in your internal site site :
- Style auxiliary data differently
- Avoid scrollbars and show 10 items maximum
- Highlight the differences
- Support keyboard navigation
- Treat hover expectations as a non-committal actions
- Show search history with the CSS visited. Persistent search makes the iteration process less frustrating
- Reduce visual noise as much as possible
- Consider labels and instructions to help the user along in their search
Faceted Search For Internal Search
Users can’t always specify their queries. If you offer faceted search, they can simply filter out what they don’t want to focus on finding what they are seeking. If you think of design details and filtering logic to improve your internal search, you are sure to tap into a big portion of customers online!
You should aim to have a dynamic labeling system. There’s nothing sadder than seeing filters in a search that simply shouldn’t be there. What do I mean? Well…sleeve choices on a pants search don’t really make sense. That’s what I mean by have a dynamic labeling system. It should adapt to the search context. Please, provide product specific filters. You should also take it a step further by trying to map filtering types to the users’ purchasing parameters.
Breadcrumbs: Hierarchy and History
Hierarchical breadcrumbs are great for non-linear navigation. They allow the user to jump back a few steps and widen their search to a larger category in the hierarchy of the site.
History-based breadcrumbs give a way to go back to previous search results. This is helping the user remain in their own specific search context.