The following is a Q&A with Mapbox’s Rafa Gutierrez about his experience working with Swiftype Site Search.

About Mapbox

If you’re not already familiar with Mapbox, they are a location data platform for mobile and web applications. They provide building blocks for developers to add location features like maps, search, and navigation into any experience they create. Mapbox’s strong customer portfolio includes The Weather Channel, Tableau, General Electric, and National Geographic. Together, Mapbox’s customers are shaping the way 300 million people explore the world.

Mapbox + Swiftype

Mapbox has been a Swiftype customer for a number of years now, and they use our search technology in a few different places on their website. Mapbox’s core use case for Swiftype is in their help center where they leverage our search technology to give their customers easy access to the documentation that’s relevant to their needs.

Recently, we spoke with Rafa Gutierrez, a Support Engineer at Mapbox, about his experience with Swiftype. During our conversation, Rafa touched on his use of Swiftype’s customizable web crawler, powerful search analytics and synonyms feature. Here’s our full conversation.

Question and Answer with Mapbox’s Rafa Gutierrez

Q: How does Swiftype compare to your previous search solution? What benefits have you experienced since implementing Swiftype?

A: Before we started using Swiftype, we had to build in specific page filtering for each API documentation page. This was cumbersome and we knew it wouldn’t scale for additional SDKs and other frameworks that we’d eventually have come online. We also needed a way for the search to cover the entirety of our site while still having some selective control over what it crawled. With Swiftype, we could just add a code snippet to all the areas of our site we needed covered by search.

We use Swiftype throughout the site but its main customization has been for our help and support pages. Swiftype crawls our pages selectively for content and pulls various categories and tags for filtering. We’re able to blacklist portions of our site to keep things relevant. We also have the ability to tailor content to the specific user and weigh out specific terms to differentiate content on similar searches.

Q: Tell us about the types of users that visit your website regularly.  What are they looking for? How are you creating experiences that are tailored to their unique needs?

A: We’re a developer platform and our customers range from government organizations to autonomous vehicle manufacturers to on-demand transportation companies. Whether it’s finding a coffee spot on Foursquare or building navigation systems for autonomous cars – Mapbox does the geo stuff so developers can focus on building their product.

As we look to meet the needs of our wide array of users, we want ensure a streamlined and tailored experience that gives users fast access to relevant documentation so they can make the most of their use of Mapbox technology. Swiftype helps us deliver relevant content so our customers can find what they’re looking for faster.

Swiftype helps us deliver relevant content so our customers can find what they’re looking for faster.

Rafa Gutierrez
Support Engineer

Q: In what ways have you utilized the data from the Swiftype Search Analytics dashboard? What is your search data telling you about how users engage with your help center and product?

A: We dive into our metrics dashboard regularly to look for oddities or patterns that give us insight to how customers are finding the info they need. When customers search for terms that we’re not listing in our glossary or used for keywords in our categories, we can find these and add them to synonyms. For instance, we found a number of customers that would search for the phrase “add marker to map”. There are a few ways to do this with our tools so we restructured our docs to surface ones relevant to getting started with creating markers on a map. We also added synonyms to Swiftype to capture words that often get conflated with markers on a map like “pins” or “points”.

Q: How are you using the combination of search and quick glance features to drive down customer service requests?

A: We use Jekyll for a static HTML site generator so we can load up all relevant keywords, categories, and excerpts as metadata for our all of our help docs. By doing this, our search can pull this info into the quick glance as customers type in their search terms. At glance, they can see which category their term falls into and read brief excerpts before they click through. This was really valuable for us in that we could apply design treatments to results and watch search results get clicked through more than 50% of the time, reduced our searches with no clickthroughs by 16%, and nearly eliminated results with multiple clicks.

We also used the search in our contact form to provide an extra layer of assistance to ensure our customers didn’t overlook a doc before contacting us. We’re continuing to iterate on this approach for our next release.

You can say 2017 was an eventful year for Swiftype. From starting the year with a product launch of Swiftype Enterprise Search, to helping countless customers migrate from their discontinued Google Site Search solutions, to joining forces with Elastic, the makers of popular open source products such as Elasticsearch, Kibana, Beats, and Logstash. To celebrate the new year, here’s a look back at the top blog posts of 2017.

1 — Swiftype Joins Forces with Elastic

“We are thrilled today to announce that Swiftype is joining forces with Elastic, the makers of Elasticsearch. This new alliance will accelerate the development of new features and tools that will make it easier for you to manage and deploy world-class search functionality to your website or organization.”

We’ve built some great search products at Swiftype over the past 5 years, but honestly, we are just getting started. Joining forces with Elastic enables us to partner with the best search experts in the world, so we can continue to innovate and deliver an increasingly diverse suite of powerful search solutions.

2 – Google Site Search is Sunsetting: What to Expect and How to Move On

“By now you’ve probably heard that Google plans to sunset Google Site Search (GSS) and direct existing customers to use their free replacement, the Google Custom Search Engine (CSE). Unfortunately for these customers, the CSE forces websites to show ads alongside their search results, so the transition isn’t as simple as it may seem. As the dominant alternative to GSS for the past 4 years, we at Swiftype are very familiar with the migration process, and we’re happy to share some of our advice to those weighing their options.”

In early 2017, Google announced it was sunsetting their widely used Site Search product. Since Swiftype Site Search is similar to GSS in its ease of use and out-of-the-box relevancy, many former Google customers turned to us for their website search needs.

3 — Customers Say It Best

“Each month, nearly 900,000 buyers utilize G2 Crowd for unbiased user reviews to assess which products are best for their businesses. Swiftype’s reviews on G2 Crowd both reaffirm the value of our product to our customers, and allow us to improve our products to make sure each and every user is satisfied with their site search.”

When potential Swiftype customers are evaluating our solution, we really encourage them to do a free trial and spend some time using our product. We believe in our product and consistent positive product reviews from our customers help to reaffirm that belief.

4 — Introducing AI-Powered Enterprise Search

“Today, we’re excited to announce a new product that addresses this problem head-on: Swiftype Enterprise Search. By integrating artificial intelligence with our industry leading search technology, Swiftype Enterprise Search is a powerful new way for employees to seamlessly access the data they need, from any device, across the many applications their jobs require.”

Earlier this year, we launched Swiftype Enterprise Search  — a modern, AI-powered search solution that enables companies to unify content from all of their apps and data stores. Our enterprise search solution is available on web, desktop and mobile and integrates with your existing workflows so you can find the files you need without breaking your concentration.

5 — 10 Examples of Well-Designed Search UIs

“Are you implementing search and looking for some inspiration? In general, intuitive search UI designs are easy to access, match the look and feel of the overall website, and offer autocomplete results and filters when appropriate. To give you a better idea, here are 10 examples of well-designed search UIs.”

At Swiftype, we’ve helped thousands of businesses to index their content, customize their search engines and then implement search interfaces on their website, either with our JavaScript snippet or our search and autocomplete jQuery libraries. Over the years, we’ve learned what makes for a good search experience and can confidently tell you that these 10 companies have nailed their search UIs — helping them to get the full value of their search technology and enabling their customers to easily find what they need.

6 — How Asana, Twilio, and SurveyMonkey optimized their customer support centers [Infographic]

“Asana, Twilio, and SurveyMonkey have all invested heavily in creating high-quality support content. To ensure this content is easily accessible, these companies leverage advanced search technology that provides relevant results to their customers and actionable analytics for them.”

One big use case for our search technology is in customer support centers. Companies like those listed above utilize Swiftype Search as a flexible portal for their customers to access their support documentation. Since search is a natural way for people to access a wide array of support content, our customers have been very successful in reducing their manual support loads while simultaneously helping their customers to resolve their issues more quickly.

7  — How site search makes higher education websites more accessible

“When a website visitor elects to use the search bar on your website, they have a specific goal in mind, so it’s important to provide them with relevant search results. At Swiftype, we specialize in building relevant site search technology and making it incredibly easy to set up on your website. We’re fortunate to be able to work with some great universities, and we wanted to highlight their Swiftype success stories.”

University websites generally have lots of pages designed to cater to a diverse constituency of prospective students, current students, faculty, staff and parents. Powerful search technology like Swiftype helps higher education institutions to improve the accessibility of their websites so that their site visitors can find the content that’s relevant to them.

8 — Great Developers Ship, They Don’t Configure Search

“Engineers and developers want to spend more time developing products and websites, not configuring search. It’s pretty obvious, and understandable. Developing allows them to be creative, solve problems, and build new things. Search, albeit a critical feature for site visitors, is part of a site’s foundation. It should already be there. And it should work, and work well.”

As you can imagine, we love to build search products at Swiftype and can tell you that they’re difficult to get right. When you’re working on a product or website and need search functionality, you may not have the time or resources to allocate towards building search. And that’s where we come in with our robust search as a service solution and well-documented developer APIs.

9 — Welcome to the Swiftype office in San Francisco [HQ Photo Tour]

“Welcome to Swiftype Headquarters in San Francisco, California. Founders Matt Riley and Quin Hoxie have created a comfortable and fun office environment with an open floor plan, scattered lounges, and various collaborative areas. Take a look at our office photos below to experience a day in the life of a Swiftype team member.”

Take a tour of the Swiftype/Elastic San Francisco Office and get a feel for what it’s like to work at a high-growth startup.

10 — Swiftype API overview: Customize your search engine

“The Swiftype API gives you full control over the schema of your search engine and the content that you index. With the Swiftype API endpoints, you are able to index data, execute search queries, and access rich analytics.”

One of our main focuses for the Swiftype Site Search product is to make it extremely easy to get your data indexed so you can quickly install a working search bar. Some of our customers prefer to interact with Swiftype through our API or client libraries which enables them to build search to their exact specifications, from the data ingestion process to user interface design.

Test drive Swiftype Site Search

At Swiftype, we provide powerful search as a service complete with customizable search UIs. We make it easy for you to get a search engine up and running by providing a web crawler and API for ingesting data, and an intuitive admin dashboard for managing your search experience. Ready to get started with Swiftype? Sign up for your 14-day free trial here.

Retaining customers is hard

Getting customers to try your product is really hard. Retaining them as happy customers year after year might be even more difficult.

“Our customers are loyal to us right up until the second somebody offers them a better service” – Jeff Bezos

When a customer uses your product, whether it’s an HR platform like Gusto, an internal messaging platform like Slack or a marketplace like Uber, they are going to have certain expectations about how it should look, work, and feel. When a customer lands on your product’s home page for the first time, the clock starts ticking for you to make your mark and retain them as a long-term customer.

Great customer experience leads to increased retention

A great framework for thinking about customer retention is customer experience. What do people think about when they hear your company’s name? Would they recommend your product to their friends? Optimizing the customer experience that your company provides will help you to improve retention and boost revenue. When thinking about providing a better customer experience, there are many different customer touch points you can work to improve. Here are 5 ways you can improve customer experience at your company.

1 – Simple customer onboarding

Drive new users to your core features

A new customer has just signed up for your product. Your goal is to quickly help them to realize the core value of your product so they will incorporate it into their workflow or daily life. According to CloudLock, there are 150,000+ unique apps used by their 10 million users at over 750 companies. With so many apps on the market, you only have a short window of time to prove your product’s value to your new customer.

In order to help your customer understand the key benefits of your app, you should have a simple onboarding flow. You should be very strategic with your onboarding flow and not walk through too many of your features as this may overwhelm a new user. Your onboarding flow should highlight your key features and encourage users to take actions that you know lead to retention. You can use a tool like WalkMe or Appcues to easily add in-app onboarding flows to your product.

Furthermore, you can send new customers emails or other notifications to nudge them towards taking specific actions. For example, a project management app might send you an email with instructions for creating your first project and inviting teammates to collaborate.

Don’t rely on your documentation to guide new users

It’s really important to remember that most of your customers are not going to read the documentation before they try your product. They’re just going to start using it and expect to learn how it works. By providing in-app onboarding and sending informational emails, you can add structure to your new user experience and boost your retention rate.

2 – Have a well-documented help center with advanced search technology

Customers prefer self-service

According to the HBR, 81% of all customers attempt to take care of matters themselves before reaching out to a live representative. Customers want to be able to help themselves, but all too often, companies do not enable customers to find what they need on their own.

Companies like Asana, Twilio, SurveyMonkey and Lyft have invested in creating high-quality customer knowledge bases and then making them accessible with advanced search technology. At Swiftype, we work with these companies and others to provide search technology that integrates with their help center platforms and enables effective self-service customer support. 

Resolve support tickets before they happen

Some companies that use Swiftype have been able to drive down the number of support tickets they are receiving by making their support ticket description field searchable. By suggesting relevant help and documentation to their customers as they file support tickets, companies are able to resolve support tickets before they happen and reduce their overall support load.

Data-driven customer support

Furthermore, using search in your help center enables you to recognize trends in search data. These trends, such as frequently asked questions or questions that return no results, will help you to decide what new support content to create and can also inform your product roadmap.

For more on using search in your help center see:

• “Want to improve your customer support experience? Think search.”

• “How Asana, Twilio, SurveyMonkey and Lyft optimized their customer support centers“

3 – Have a scaleable system for answering customer questions

What about questions that customers can’t find the answer for in your help center? Whether they issue a support ticket, come in through live chat or send you a tweet, you want to provide consistent and timely answers to their questions.

Give your support team the tools they need to succeed

In order to scale your customer support operations, you need a help desk platform like Help Scout or Zendesk and then a reliable platform for aggregating customer questions from disparate channels like Salesforce Service Cloud or Front. This will help you get the most out of your customer support team by quickly assigning new questions to team members and making sure they don’t duplicate their work.

In general, customers aren’t used to quickly getting helpful answers to their questions. If your company can do this, it will be a huge plus to you customer experience efforts and will positively influence customer perception of your company.

4 – Iterate based on customer feedback (listen to your customers)

“Listen to what your users tell you, improve your product, and then listen again.” – Sam Altman

This is a pretty straightforward tip but one that gets easily overlooked or even dismissed. As you continue iterating on your product, you’ll want to take customer feedback into consideration. Every company knows they should do this; the hard part about listening to customer feedback is deciding how much it should actually influence your product roadmap.

Customer feedback and your product roadmap

When it comes to building new features based on customer feedback, take a look at this post by YC Partner Geoff Ralston: http://blog.geoffralston.com/startup-priorities.

His post focuses on helping resource-constrained startups figure out what to build next, but I think it applies to software companies more widely as it’s very common to be constrained when it comes to developer resources.

Balancing feedback with product vision

Although customer feedback is a valuable tool for shaping product vision, it’s also important to draw some inspiration from other sources. Steve Jobs is well known for saying, “people don’t know what they want until you show it to them” (Source). This makes sense considering he conceived of the iPod which most Apple customers at that time probably would not have thought of, let alone have considered technically possible. So, it’s important to listen to customers but don’t let their feedback steer you away from testing out some big innovations.

As for collecting customer feedback, you should have a system in place for doing so. More on this in tip #5.

5 – Measure customer happiness (so you can keep improving)

Simply put, it’s important to measure customer happiness so you can work to improve it. We all know that it can be difficult to get customers to answer surveys, so you should aim to measure customer happiness in a simple and concise manner.

Easily measure customer happiness and loyalty with Net Promoter Score

One popular way to measure customer happiness is the Net Promoter Score (NPS). NPS is a popular measurement because it’s a single question and research has shown that strong NPS ratings correlates with revenue growth (Source).

How NPS works

If you’re unfamiliar with NPS, it is calculated based on the responses to a single question: How likely is it that you would recommend our company/product/service to a friend or colleague? The scoring for this answer is based on a 0 to 10 scale.

Customers who respond with a 9 or 10 are called Promoters and you can expect them to be loyal customers. Customers who respond with a score of 0 to 6 are considered Detractors, and those who respond with a 7 or 8 are labeled as Passives.

NPS is calculated by subtracting the percentage of customers who are Detractors from the percentage of customers who are Promoters. So, if 40% of respondents were Promoters and 15% were Detractors, your company’s NPS is 25.

Other NPS notes:

• Because many companies measure their NPS score, you can benchmark your score against your industry to get a feel for how you’re doing.
• As part of your NPS survey, it’s a good idea to ask a follow up question like, “Can you share why you rated us a 6?”
• Tools for measuring NPS: SurveyMonkey, Promoter.io, Wootric

Getting started with Swiftype

After implementing Swiftype, SurveyMonkey noticed that over 90% of people that visit their help center find the information they’re looking for without contacting a support specialist.

Swiftype Customer Support Search enables companies to quickly help their customers find what they’re looking for. After implementing Swiftype, SurveyMonkey noticed that over 90% of people that visit their help center find the information they’re looking for without contacting a support specialist.

Interested in implementing Swiftype for your help center. You can sign up for a free trial here or learn more on our website here.

Do you use Zendesk for customer service and support tickets? Check out Swiftype for Zendesk.

When it comes to solving issues, customers prefer self-service

Studies have shown that a majority of a company’s customers prefer to solve problems themselves rather than reaching out to a representative. For this reason, customer support search is incredibly important to help those individuals find the information they’re looking for.

Better customer support through search

Asana, Twilio, and SurveyMonkey have all invested heavily in creating high-quality support content. To ensure this content is easily accessible, these companies leverage advanced search technology that provides relevant results to their customers and actionable analytics for them.

Asana, Twilio, and SurveyMonkey have been successful with focusing their help centers around search. To highlight some of the key strategies they’ve used, we created this Customer Support Search infographic.

Get Started with Swiftype

Want to try Swiftype Customer Support Search for yourself? You can learn more on our website or sign up for a free trial.

Congrats on your Swiftype implementation! You’ve already started delivering a powerful site search experience for your visitors. Trust me, they’ll thank you for it.

We’ve been listening to our community and we’re here to answer 4 of the most commonly asked questions from Swiftype customers.

1. Why do my results all look the same?

When Swiftype’s crawler, lovingly called “Swiftbot”, crawls a domain, there’s potential for repetitive template elements to be indexed with the page body. Most notably the navigation/header, sidebars, and/or footer content. With all the great, meaningful content encased in template noise, the quality of the customer’s search experience can be negatively affected.

For example, Apple.com has these elements on their website:

Since these elements appear on all pages, when their site it indexed, in the Swiftype dashboard, it looks like this:

You can easily clean this up by using our Content Inclusion/Exclusion tag recognition. By adding Swiftype specific meta tags (data-swiftype-index=’true’) to the HTML container(s) that holds the primary page content, it’s possible to instruct Swiftbot to index only those sections of the page body.

The best practice is to set the main content container to true. If you want to further refine what’s indexed from that section, you can add additional tags with a value of false to containers nested within.

<body>
    <nav>Blah Blah blah</nav>
    <div id="main_content" data-swiftype-index='true'>
         <p>All of my sweet, sweet 5/7 content is going to go in here.</p>
             <div id="ad_widget" data-swiftype-index='false'>This bit really isn't as important which is why I'm going to add a 'false' exclusion parameter.</div>
         <p>This bit will be indexed though, because it’s still within the ‘main_content’ div that’s set to 'true'. Everything outside of the ‘main_content’ div container will be ignored, yo.</p>
             </div>
    <footer>Copyright Attempting to Sound Official© 2016</footer>
          </body>
        

2. How do I prevent Swiftype from indexing certain pages of my site?

For crawler based engines there are three approaches you can take to determine what pages are indexed from your domains: URL path rules, a customized robots.txt file, and robots meta tags.

Path Rules:

From the Manage > Domains section of the Swiftype customer dashboard, there’s an option for each domain that will allow you to ‘Manage Rules’ for that domain. From here, you can define specific paths to include (Whitelist) or exclude (Blacklist) when crawling your site.

Examples of common cases are the exclusion of /category/ paths for ecommerce sites, so the focus is exclusively on crawling product pages. For other CMS based sites, you’d likely see paths to login or administrative pages excluded, as well as dynamically generated content, like pages based on tags or categories. More examples and tips on using this feature can be found here.

Robots.txt Files:

A robots.txt file is a plaintext document that you can upload to the root directory of your website’s domain. With the robots.txt file, you can define URL path exclusion rules for all or only specific web crawlers to follow. Many websites will commonly have a robots.txt file already in place, and it’s presence is one of the first things Swiftbot will look for when starting a crawl process.

Check out our Robots.txt documentation to learn how you can leverage this with Swiftype.

Robots meta tags:

If you need to exclude content in a more precise way (page by page or page template basis), we recommend and fully support robots meta tags. We adhere to the robots tag standard that’s a companion to the aforementioned robots.txt file.

This means that we’ll pass over any page we attempt to crawl that contains a meta tag:

<meta name="robots" content="noindex">

You can also configure these tags so they only apply to Swiftype’s web crawler:

<meta name="st:robots" content="noindex">

Similar to the robots.txt file, meta tags are configured and managed outside of the Swiftype dashboard via their webhost/CMS.

You guessed it. We’ve got documentation of robots meta tags support here.

3. “Why are pages missing from my search engine?”

Here are 3 reasons why Swiftbot, our crawler, may not be able to locate and index pages on your site:

A. We’re unable to find the content because it’s not linked to from other pages.

When spidering a domain, the Crawler will examine all links within a page to discover URLs that are part of the domain submitted to the engine and also adhere to any configured path rules (see question 2 above). If content exists within a site that is not linked to from another known page, within the site’s navigation menu, or listed on a domain’s sitemap, chances are Swiftbot will not be able to locate it.

One of the best ways to ensure our Crawler is able to index all desired content is to include a current sitemap. Sitemaps are files that are typically stored at the root of a URL domain and contain a list of links to pages on their site(s) that are available for crawling. Our documentation for sitemap support and installation notes can be found here.

B. Improperly configured canonical URL elements/tags.

A canonical link element is often used like a meta tag, in order to prevent duplicate content indexing issues by pointing web crawlers to the preferred (or canonical) URL version of a web page. A scenario that can occur is that a customer’s site content is configured with static canonical tags that all point to the root domain URL.

Due to this misconfiguration, even if the crawler finds links to all the content on the site, it is being given instructions that all pages on the site are another version of the home page. With that directive in place, only the home page will be indexed. If the customer re-configures or removes those elements, a recrawl will be able to index their content successfully.

For best practices on canonical elements, you can refer to Google’s documentation here.

C. The content is being excluded by one of the methods noted in question 2.

Just as with misconfigured canonical link elements, misconfiguration or conflicts in a customer’s path or robots rules can cause pages to be skipped over.

4. My site is password protected / behind a firewall / hosted on our company’s intranet.

It is possible for Swiftbot to crawl secured content, but you’ll first need to make minor configuration changes to your web or intranet site’s host server.

All Swiftype accounts have an account specific User-Agent ID string. By whitelisting this identification string with your server, you can allow, or disallow, crawlers access to your site’s content.

Swiftype has a unique security feature where we encode our crawler’s User-Agent with a secure key that is uniquely tied to your Swiftype account. This approach enables you to limit access solely to Swiftype’s crawler, and is an extra level of security many customers enjoy.

If you’re interested in using the Swiftbot web crawler to access your secured content, please contact our support team and we’ll be happy to supply you with your account specific User-Agent ID string.

Hopefully these answers to commonly asked questions will point you in the right direction. If you ever have questions, suggestions, or feedback, you can always email [email protected] to reach our team. We’re happy to help!

We live in a world where the solution to nearly any problem can be solved instantly. When you need the answer to a question, there’s Google, or Siri. If your refrigerator is empty at dinnertime, you have Postmates or Munchery to the rescue. So when you need help for a product you purchased or service you subscribe to, waiting is not an option. As a Customer Success Manager at Swiftype, I help our clients build upon the success of their own customer service departments through the business of instant answers.

Many of Swiftype’s customers use our search software for their Knowledge Base. Knowledge Bases can vary from a FAQ or Help section provided for end consumers as a first stop in the customer support chain, to internal documentation for developers using a product. No matter what the final application looks like, help centers are becoming an integrated part of the product experience today.

Customers are searching now more than ever
Thanks to Google, we’ve all been trained to search. Customers and employees alike expect a self-help channel to service themselves prior to getting on the phone for support. In fact a Forrester survey shows that for the first time in the survey’s history, customers of all ages are using the FAQ pages on a company’s website more often than speaking with an agent on the phone.

This change in user behavior has generated a significant need for our customers to have not only a robust Knowledge Base that they themselves can navigate and find answers within quickly, but the ability for their end customers to quickly search those databases and find answers to their questions instantly.

The need for instant answers is not confined to one industry or solution. Speed and accuracy are critical in a search solution within knowledge bases across the board.

Providing Instant Answers Three Ways
Customer Support teams are using Swiftype as a means to aggregate content in one location. A customer of ours has content that sits in multiple subdomains of their site for developers, designers, their sales team, within their blog, training center and help center. As each individual portal grew, it became increasingly difficult to quickly find the information you were looking for. They’ve successfully leveraged Swiftype to search across all content types, creating a one-stop locale for all internal users when they are unsure of where to locate content.

Another way Swiftype has been implemented by customers internally is to better support call center teams. We are the internal search engine for a global enterprise organization with many call center teams who need to be able to answer questions for their end customers quickly and accurately. If decreasing call time and increasing customer satisfaction are critical metrics for your team, consider ways to make your employees more successful with a customizable search solution.

Lastly, we’ve even been leveraged for instant answers before your customer is even actively searching. Swiftype results are shown to users in a right-hand drawer that slides in when users are in their dashboard for our customer’s product, suggesting answers that are related to the page they are on.

Maintaining a best-in class customer support system is critical to businesses. Knowledge bases are a key component to providing an excellent experience. Gone are the days where your customer is willing to wait 10 minutes on hold with a customer service representative, and another 20 minutes to explain their issue, and hopefully receive a solution. The faster you can provide an answer to your customer, the happier and more satisfied they leave.  And I’m in the business of making happy customers.

Adding powerful search to your company knowledge base or help center is an essential step toward helping users resolve issues without filing tickets, but the process of implementation can often be a daunting first step. To make this process easier, we added a new resource to our website for customers looking to improve search on their help center or knowledge base.

Last week, we added a new resource to our website for customers looking to improve search on their help center or knowledge base. Written with a non-technical audience in mind, 15 Steps to Help Center Search Implementation introduces readers to the most important questions that companies should familiarize themselves with before getting started, including:

• What are the pros and cons of building search internally?
• How long will implementing search take?
• What team members are required to implement new search?
• How can I evaluate the performance of a new search experience?
• What ongoing work is required after implementing search?

With a step-by-step checklist that clearly lays out each phase of implementation, this resource will be an invaluable guide. Implementation is often an intimidating prospect with any new software, but with this guide in hand, your team will start with a much clearer sense of what needs to be accomplished and what team members will be involved.

To access the checklist, follow the download link below. To receive a personal demo of what Swiftype can do for your knowledge base or help center, contact us today.

While personal customer support remains an essential ingredient for attracting, retaining, and creating loyal customers, this level of individualized attention is expensive to support difficult to scale as your company grows. In response to this challenge, creating a comprehensive user-facing knowledge base that enables customers to resolve support issues without contacting your team can save companies time and money while also improving the overall user experience.

Once in place, any steps that a customer support team can take to optimize this knowledge base and create a more efficient user experience can substantially decrease inbound support volume. In this pursuit, creating a centralized search experience that allows support center visitors to quickly and easily find the content they are looking for is vitally important.

Still, while the importance of search across knowledge bases is fairly self-explanatory, customer support teams often overlook the valuable insights available from user search behavior—insights which include:

• What are the most pressing issues my customers are facing?
• What issues are users unable to resolve on their own?
• What new support content needs to be created?
• How relevant are the results for users who search across my knowledge base?

To give support teams a clearer sense of precisely what information they should be looking for from their search analytics and to provide actionable recommendations about how they might improve their on site search, Swiftype has created the Knowledge Base Guide to Site Search Analytics.

To access this guide and learn how customer support teams can leverage knowledge base search analytics, follow the download link below.

When creating search engines, the first and arguably most important step is indexing website information in a structured format that is optimized for a specific search algorithm. The specific information you index and the structure by which you organize this information (also known as the schema) dictates how your search engine will determine relevance, what your users can search by, and what information you can display in search results.

How does indexing work?
While there are numerous ways to customize and control the information you index in your Swiftype search engine (for example, via our API or one of our platform integrations) we aim to make this process as simple as possible for non-technical users by automatically indexing website information with Swiftbot—our high performance web crawler designed to index information from a specific URL.

Swiftbot allows non-technical users to get up and running with a working search engine in minutes by simply entering their website URL and letting Swiftbot index their website for them. A major component of Swiftbot’s technology is the logic that our engineering team has built in to parse website HTML and index it in a structured format that works with Swiftype’s advanced search algorithm and information retrieval method. (To learn more about the technical challenge of building a search engine, read our white paper on the subject, written for a non-technical audience).

Building an intelligent web crawler
Because almost every website is built and structured in a different way, teaching Swiftbot how to effectively read, sort, and organize information from a website’s HTML base is an ongoing challenge. While we do allow site owners to completely customize the default information Swiftbot indexes from your website with custom <meta> tags, not all users have the technical resources or knowledge to do this on their own, so Swiftbot is also built to make many of these indexing decisions on its own.

With every website structured differently, how do we teach Swiftbot to intelligently index this information?

Still, with websites differing so dramatically from one another, indexing the right information in the right format from each page is no easy task. In particular, identifying the most important image from a web page and associating that image with a search result is a multifaceted problem, since there are many images on every page and these images often have different filename structures and/or occupy different locations on a page.

Adding images to search results pages and autocomplete menus can create a much more engaging search experience.

Nevertheless, indexing images allows site owners to create much more engaging search experience, adding thumbnails of varying sizes to their autocomplete and search results that let users see a preview of the page content before selecting a result. So, in a recent update to Swiftbot, we’ve built in conditional logic that automatically indexes images from your website pages (provided there are no Swiftype specific image tags already in place).

How does Swiftbot decide which image is “best”?
To teach Swiftbot how to index the “best” image from web pages, we had to build in logic that would overcome a series of challenges that result from the varying nature of website pages.

1. As a starting point, we decided to leverage existing open graph <meta> tags (such as Facebook and Twitter <meta> tags) that many site owners use to prepare their content for sharing on social media platforms and other content distribution networks. By teaching Swiftbot to obey these <meta> tags if no Swiftype specific <meta> tags exist, we created hierarchical indexing logic that more intelligently sources images from existing website metadata.
2. Secondly, we know that many websites have a large number of images that repeat across many, if not every page on their website (for example: a company logo, images in the header, footer, and sidebar, author headshots, ads, etc.). To ensure these images are not considered the “best” image for a specific document, we built in logic that identifies and rules out these repeating elements as candidates. Similarly, we do not want to index advertisements, so we run any images on the page against an ad server blacklist to ensure these remain out of consideration.
3. Thirdly, we compared data in the alt attribute of each <img> with the url and <title> of that page, assigning a relevance score to those images based on how closely the alt description matched this page information.
4. Lastly, Swiftbot looks for common CSS classes and id’s to locate the main content area of each page—another step that helps rule out extraneous information such as the header, footer, and sidebar.

Taking all these pieces of information together, Swiftbot assigns the images on the page a relevance score and indexes the image it judges to be the “best” image for that document. As this new indexing process gains wider use and we gather feedback from customers, we will continually work to improve our image extraction technology over time.

Adding these images to search
Once these images are indexed from your website and in your search engine, the question becomes: how do I display these image thumbnails in my search results and autocomplete dropdown? While there are many ways to style your autocomplete and search results (including using Swiftype’s web components or jQuery library) the best choice for users with very little technical experience is the Result Designer, which allows users to style their search results entirely from the Swiftype dashboard without writing any additional code. To learn more about the Result Designer, watch our dedicated webinar explaining this tool and offering best practices advice from the Swiftype customer success team.

As you’re getting started with Swiftype, you may be wondering how to index thumbnails from your website and serve them to users in your search results. The answer to this question lies in using Swiftype’s custom <meta> tags, which allow site owners to pass detailed web page information directly to Swiftbot, our web crawler, as it moves across your site. As Swiftbot encounters these custom Swiftype <meta> tags, it indexes their content and incorporates that information in your search engine index schema.

To index thumbnails from your website, all you need to do is add a Swiftype image <meta> tag to the <head> section of your website template that indicates where images are located on your various page types. For illustration purposes, the Swiftype image <meta> tag is formatted like this:

Swiftype recommends placing these <meta> tags at the template level of your website to ensure that image files are dynamically populated within the tags, rather than being added manually for every page on site.

NOTE: the value of the “content” attribute must be HTML encoded. For more information see this guide.

Alternatively, you can wrap images with a body-embedded Swiftype image <meta> tag to avoid changing your website <head>. For example, Swiftbot will index example.jpg into the image field from the HTML below:

<body>

Hello world

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Sed ut risus sed ante dignissim pharetra aliquet a orci. Maecenas varius.

In in augue molestie, bibendum velit vel, luctus erat. Curabitur cursus, tellus at feugiat lacinia, tellus est suscipit lectus, non commodo diam elit sit amet justo.

http://fullurl.com/example.jpg </body>

It is important to note that in both the <head> and <body> embedded <meta> tags, you need to specify the data-type attribute as enum. For images, this will always be the case. For any other custom meta tags you choose to define, each attribute must be a valid, Swiftype-supported field type, which you may read about here.

Once you index thumbnails from your website, you can easily customize your search results and autocomplete to feature thumbnails in a range of shapes and sizes with the Swiftype Result Designer.

To learn more about using custom Swiftype <meta> tags to refine your search engine index, check out our tutorial. As always, if you need help or have any questions, feel free to reach out to us at [email protected].