Using the metrics that matter to improve your knowledge base

Support teams know the most about customer issues and the best way to solve them. That’s why a knowledge base is a crucial part of your customer experience strategy. It organizes frequently asked questions, product details, policies, and more, and empowers customers and agents with that information. Integrated seamlessly with your ticketing system is a smart help center. It’s built to help you and your team continuously improve your content, keep it up to date, and serve it to customers before you can say “FAQ”! A great knowledge base means your customers are happy because they get help faster without having to contact you. This leads to a faster and more efficient service operation for you and your agents.

However, this leads to the question "How do I know if my knowledge base is doing its job?"

That question has different answers depending on your business and your customers. In this article, you'll learn some useful tips to help you measure the effectiveness of your knowledge base. You'll find that Zendesk includes a bunch of tools to help you analyze your knowledge base, but you'll also learn about other tools you can additionally use to get even more information.

Important: Zendesk and third-party products are constantly evolving and we’ll keep this article up-to-date as things change. Make sure to hit the Follow button at the top of this page to be the first to know when we make updates.

This article contains the following sections:

• Is anyone looking at your articles?
• What are your readers searching for?
• Calculating your self-service score
• Understanding article votes
• Using Answer Bot to analyze article performance
• More tools and resources to help you monitor your knowledge base
• Join the conversation!

Is anyone looking at your articles?

A pageview is a count of how many times your page has been viewed in a web browser. Page views are a useful indicator that people are using your knowledge base. You can use them to show which articles your users are most interested in, help find out if your pages are optimized for search engines, and to help you understand user behaviors.

Advantages of monitoring pageviews

• Shows that people are actually using your knowledge base.
• Gives an indication about which of your articles are the most popular.
• Collecting pageviews over time can give you a good trend as to how your articles are performing over time.

Disadvantages of monitoring pageviews

• A pageview doesn't indicate whether the reader found what they needed in your article.

Always monitor page views in conjunction with other monitoring methods to ensure you have a complete picture of your team's success and areas where you could improve.

Examining page views with Explore

Explore currently tracks all pageviews to your articles including repeat visits. If you want to track only unique visitors, you'll need to use Google Analytics.

In this section, you'll learn how to create a quick count of all of your pageviews using Explore.

To create a pageview report using Explore

1. In Explore, click the reports ( ) icon.
2. In the Reports library, click New report.
3. On the Select a dataset page, click Guide > Guide - Knowledge Base, then click Start report. The report builder opens.
4. In the Metrics panel, click Add.
5. From the list of metrics, choose Article views > Views, then click Apply. Explore displays the total number of article views in your knowledge base.
6. In the Rows panel, click Add.
7. From the list of attributes, choose Article > Article translation title, then click Apply. Explore displays a table showing the number of page views for each article in your knowledge base. Click the Views column header to sort the table into order by number of views.

   

Tip: Usually, you'll want to look at page views over a period, like "Last month" rather than over the whole life of the article. To do this, add a filter for Time - Article views recorded. For help with Explore filters, see Filtering a report.

Examining page views with Google Analytics

Google Analytics tracks both total page views, and unique page views only. It also gives you an indication of how long the visitor spent on your page. This can help you understand the level of engagement visitors have with your content. Additionally, Google Analytics can help you to understand some information from Zendesk Gather (though this capability is on the roadmap for Explore).

Google Analytics has a lot of features and could fill a series of articles by itself. In fact, that's what we've done. Take a look at our Google Analytics and Help Center series of articles to take a deep dive into what you can do.

What are your readers searching for?

By understanding the terms that your customers search for on your site, you can ensure that relevant content is returned when they search.

If you're using Guide Professional or Enterprise, you use the search analytics dashboard to review the search words that customers have entered in the help center search field. For each search term that's been used, you can see the number of searches for that term, number and type of search results returned (if any), click-through, and the next action taken.

To open the search analytics dashboard

• In Support, click the Reporting icon () in the sidebar, then click the Search tab.

  

For more information, see Analyzing help center search results with Explore.

How to improve search results

End users generally search your knowledge base in one of two ways:

• From the search bar in your help center.
• From an internet search engine like Google or Bing.

The methods for optimizing content for search are similar both for your help center search and search from a search engine like Google or Bing. You can improve end-user search results by updating content in the following ways:

• Add labels to content. (Guide Professional or Enterprise) You can add labels to content so that the appropriate content appears in search results. For example, if you see that end-users are searching for "e-mail" but you use "email" without a dash in your content, add a label with the similar term to the content.

  You can test how well this works in your own help center. To get started and find information about optimizing your search results, see About Help Center end user search.

• Update content titles. You can rewrite titles to more closely match end-user searches. For example, if your article is titled "Deleting an email account" and end-users are searching for "removing a user from email," consider updating the title.
• Break content into smaller articles. You can divide large articles into smaller articles to help customers find what they are looking for. For example, instead of "Managing email" consider smaller articles about "Setting up email", "Adding email accounts", and "Deleting email accounts".
• Update content body text. You can add common end-user search terms to the body of appropriate articles so that the article appears higher in search results.
• Remove old content so that it does not clutter search results and confuse your customers.
• Ensure content is available. If people are searching for an article in your knowledge base that doesn't exist, or if you see a high volume of tickets for a topic that’s not covered, write an article about it!

Calculating your self-service score

One useful indicator of how well your knowledge base is doing is your self-service score. This measures the number of unique visitors to your knowledge base against the total number of users who've submitted support tickets.

Currently, Explore can't calculate your self-service score directly as you can’t combine data from tickets and pageviews in the same report. In this example, you'll calculate your self-service score over the last month.

To calculate your self-service score

1. In Explore, create a report using the Guide: Knowledge Base dataset. Add the SUM(Views) metric and a filter containing the attribute Time - Article views recorded. Ensure that the date range on your filter is configured for the past month.
2. Next, create a report using the Support: Tickets dataset. Add the COUNT(Tickets) metric and a filter containing the attribute Ticket created - date. Ensure that the date range on your filter matches the date range in your knowledge base report.
3. Divide the number of pageviews by the number of tickets created over the same time period to obtain your self-service score.

Continue to monitor your score over time to track improvements, or areas in which you need to improve. For example, after you created a number of new articles, did your self-service score improve?

Understanding article votes

At the bottom of each of your articles, you'll find two voting buttons that let you vote an article up or down. A number representing the difference between positive and negative votes is displayed.

If an article has more negative votes than positive votes, the number of votes is shown as a negative value. Here are some things you can consider when an article has negative votes:

• Read the article. Is it up to date? Check if any procedures still work.
• If you have commenting turned on, read the comments to see if there are any clues to what your readers don't like.
• If you wrote the article, consider asking another team member or even a customer to peer review the article.
• Try to discover what people don't like about the article. Maybe organize some customer meetings. Sometimes, you'll find that the article is fine, but customers have an issue with the product itself. In those cases, you can pass along that feedback to the relevant teams at your company.

Reporting article votes with Explore

In this simple Explore example, you'll learn how to report the top ten articles by the most votes.

Note: Explore only reports the current article votes. You can't report on the history of votes over time.

To create a report by votes

1. In Explore, click the reports ( ) icon.
2. In the Reports library, click New report.
3. On the Select a dataset page, click Guide > Guide - Knowledge Base, then click Start report. The report builder opens.
4. In the Metrics panel, click Add.
5. From the list of metrics, choose Article views > Article votes, then click Apply. Explore displays the total number of votes across all of your articles.
6. In the Rows column, click Add.
7. From the list of attributes, choose Article > Article title, then click Apply. Explore displays a list of all of your articles together with the vote for each one.
8. Now, you'll restrict the results to only the ten most voted for articles. In the result manipulation menu (), click Top/bottom.
9. In the Top/bottom panel, select Top, and set the value to 10.

   

10. When you are finished, click Apply. Explore displays a table of your ten highest voted articles.

    

Using Answer Bot to analyze article performance

Answer Bot uses machine learning to answer customer questions with content from your knowledge base. By using Answer Bot, you can help customers to get answers to their questions without an agent ever becoming involved.

A great read to get started using Answer Bot is Quickstart guide: Answer Bot. You’ll find lots of information about how to get things up and running and how to optimize your content to make Answer Bot work better.

When you’re ready, Explore offers a whole host of ways to report on how well you are doing with Answer Bot.

See Analyzing your Answer Bot activity for more help.

More tools and resources to help you monitor your knowledge base

In this section, you'll learn about other useful tools that will help you measure the effectiveness of your knowledge base and the articles it contains. Some of these are third-party tools so contact each company if you want more details:

• Zendesk Answer Bot, an add-on for Guide, uses machine learning to scan the text of incoming support requests and then automatically responds to tickets with a list of relevant knowledge base articles that may help your customers resolve their issues without having to interact with an agent.
• The Zendesk Knowledge Capture app enables agents to easily share and direct customers to knowledge base content, to help customers solve their support issues themselves.
• If you’re using a Suite or Guide Enterprise plan, Content Cues can help you discover opportunities and tasks that will improve your knowledge base health by identifying tickets that could make good articles and also show you under-performing articles that might need further review.
• Pendo gives you powerful usage analytics about your product and web pages. You can use Pendo to help understand the user's journey through your project which can help you to understand where you might need to write further usage assistance. It can also help you understand where your knowledge base visitors came from. Pendo has an app that works with Zendesk to give you information right inside your Zendesk tickets.
• Crazy Egg: takes article analytics a step further by allowing you to see how readers interact with your page. For example, you can generate heat maps that show where readers spend the most time on your page, or click maps that show you where they click. A similar product to Crazy Egg is Hotjar.
• We have a bunch of recipes to help you get started monitoring and improving both your knowledge base and your support operation in general. To get started, see Explore recipes reference.
• A knowledge base is a great way to organize both internal and external-facing content. We've got some great best practices and tips to help you create and structure your knowledge base to make things easy to find, and to keep things up to date. See Best practices for creating an internal knowledge base.
• Want to go further with Google Analytics? Our series of articles, Google Analytics and Help Center takes you on a deep-dive into how use this analytics tool with your knowledge base.

Join the conversation!

What metrics do you use to help measure the success of your knowledge base? How have you managed to improve them? Let us know in the comments below!