When you create a help center for your customers, you’re providing them with a self-service channel to solve their own problems instead of opening tickets. This helps you scale your customer support organization because self-service can result in fewer requests for agents to handle, known as ticket deflection.
This article introduces the tools and metrics that work together to measure the effectiveness of your self-service channel. To get the most out of this article, you should already be providing your customers with self-service content using Guide so that you have some user activity to measure.
- Analyzing knowledge base engagement metrics
- Analyzing search engagement metrics
- Monitoring help center traffic and activity with Google Analytics
- Calculating your self-service score
- Analyzing Knowledge activity for knowledge base content and solved tickets
- Analyzing automated ticket resolution via autoreplies with articles
- A summary of your self-service channel reporting options
Analyzing knowledge base engagement metrics
Analyzing knowledge base activity begins in the Knowledge Base dashboard in Explore. In this dashboard, administrators can measure essential engagement metrics for the knowledge base.
For more information, see Analyzing knowledge base activity with Explore.
Analyzing search engagement metrics
If your customers can’t find the information they’re looking for in your help center, your self-service channel will be of little help to them. The Search dashboard in Explore is where you’ll find metrics that help you track what your customers are searching for and what actions they take after searching.
For more information, see Analyzing help center search results with Explore.
Monitoring help center traffic and activity with Google Analytics
Just as with any other website, you can monitor and analyze the traffic and activity in your help center using Google Analytics. Google Analytics provides industry-standard metrics for website traffic, user activity, and user engagement. When used with the Knowledge Base dashboard in Explore, which gives you a snapshot of essential activity data, Google Analytics helps you to dig much deeper into important user activity and engagement metrics.
Although these metrics can’t tell you how many tickets have been deflected by using your help center, they do provide you with a comprehensive understanding of the use and effectiveness of the content in your help center.
To get started, set up Google Analytics for Guide. Then you can track your help center activity in your Google Analytics account.
- Page views: This is the number of page views in your help center. You can track views in both Google Analytics and the Knowledge Base dashboard in Explore.
- Unique page views: This is the number of unique visitors to your help center. Each visit to your help center counts as a session, and each session (usually) results in multiple page views. Tracking the number of users visiting your help center gives you some perspective about its use compared to the total number of views in a specified period. A monthly views total of 10,000 compared to 1000 unique users within that same period tells you that those users are viewing, on average, 10 pages per session. This helps you understand how many of your customers use your self-service content.
- % New Sessions: Understanding how many new versus returning users visit your help center helps you focus on the content that addresses the needs of those users. For example, rolling out a new product may result in a spike of new users, which you can address by providing the information needed to use the new product.
- Average session duration: The average duration of a user session in your help center tells you how much time they spend in your help center and, if you drill down deeper, how much time they spend reading specific articles and FAQs. Ideally, they spend enough time to read through the information you provided them. If they don’t, that tells you something as well—that perhaps your content isn't engaging or isn't the information they need.
- Pages per session: This is the average number of pages viewed during a session on your help center. Once again, this tells you how much of your self-service content is being used.
- Bounce rate: This is the percentage of single-page sessions in your help center. A bounce means that the customer left your help center after viewing the first page they landed on. A user may have visited the help center unintentionally, or didn’t like what they saw when they got there.
With Google Analytics, you can also analyze what users are searching for and what actions they take after those searches.
For more information, see the following articles:
- Google Analytics and your help center - Part 1: Asking the right questions
- Google Analytics and your help center - Part 2: Measuring the effectiveness of search
- Google Analytics and your help center - Part 3: Tracking customers' actions
- Google Analytics and your help center - Part 4: Fine-tuning your help center
- Google Analytics and your help center - Part 5: Capturing help center user data
Calculating your self-service score
To begin more directly quantifying the effectiveness of your help center as a self-service channel, and its impact on ticket deflection, you may want to determine what your self-service score is. This metric, also known as the self-service ratio, is a manual calculation you can make using this formula:
Self-service score = Total user sessions of your help center(s) / Total users in tickets
- Set up a Google Analytics account and connect it to Guide as described in Enabling Google Analytics for your help center.
- When you have several months of user activity available in Google Analytics, take a 30-day snapshot (for example) of the number of visitor sessions in your help center.
- Divide that number by the total number of users who have submitted tickets in that same time period. See Explore recipe: Finding how many users submit tickets each month.
When making this calculation, you may also want to define what you consider to be active use of your help center in an attempt to self-serve. In 6 steps for measuring self-service success, Erin Cochran of RJMetrics says, “We defined ‘content interaction’ as someone who did more than just visit the help center landing page or navigate straight to a new ticket form. This allowed us to get a better idea of how many visitors were actually trying to self-serve before submitting a ticket.” Erin shares other useful tips for evaluating self-service in her article—give it a look.
Analyzing Knowledge activity for knowledge base content and solved tickets
Knowledge in the context panel enables agents to easily share and direct customers to knowledge base content to help customers solve their support issues themselves.
There’s manual intervention needed here because agents add the links to knowledge base content into their replies to customers, but you can then track whether the linked content helped the user solve their own ticket. Tickets aren’t deflected in this case, but their resolution may be the result of the use of your self-service channel.
For more information, see Analyzing your Knowledge or Knowledge Capture app activity.
Analyzing automated ticket resolution via autoreplies with articles
The autoreplies with articles feature 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.
Like Knowledge in the context panel, you can view analytics for autoreplies with articles activity in Explore. Most importantly, you can see how many tickets were solved using your knowledge base articles.
This includes overall performance (how many times the links resolve tickets), and also the performance of individual articles (which articles are the best and worst at helping customers solve their problems).
For more information, see Analyzing article recommendations.
A summary of your self-service channel reporting options
Reporting tool | Location of reports |
---|---|
Knowledge Base dashboard | |
Search dashboard | Search dashboard in Explore |
Google Analytics | Google Analytics dashboard |
Knowledge Capture dashboard | Knowledge Capture dashboard in Explore |
Article Recommendations dashboard | Article Recommendations dashboard in Explore |
19 comments
Kate J.
Just wanted to clarify, based on this information, that in using the knowledge capture tool to link articles to customer tickets we can essentially track issues that could have been self-serviced but weren't, correct? Along for content improvements, of course.
0
Jim
Hi Kate!
Technically yes, that would be correct. But if you're looking for a report, in order for you to track those; then you may use the documentation below as reference.
Analyzing your Knowledge Capture activity
Explore recipe: Analyzing Knowledge Capture activity
1
Mary Jane Reese
when I go to the Search tab, I see the search stats for 'no clicks' and 'no results'. when I run the CSV from the top of that page, the data doesn't match. Several of the responses don't show on the CSV but they are on the list on in the reported data on that page.
0
Christophe Tiraboschi
Hi Mary Jane Reese,
Since you posted, we made available a search dashboard and a search dataset. You can find more information about this in the following article
We recommend using it.
However, if the issue you encountered is about results before October 15, 2021 (that are not in Explore), please let me know so I can open a ticket on your behalf to investigate further.
0
Lance Gu
The picture disappeared shortly after uploading the guide, it's very weird
0
Geerten Goossens
hi Christophe. About the search dashboard. On the helpcenter (of funda.nl) it says I have 0.0% clickthroughrate. Google analytics tells a different story. Any clue why this is happening?
The mobile SDK is showing figures on the CTR.
0
Christophe Tiraboschi
I am going to open a ticket with you so we can investigate this further.
You will receive an email shortly.
0
Jessica Cohen
According to this article, the page views and unique page views can be tracked in Google Analytics or in the report, but our figures do not match up. There's a difference by the thousands! Is anyone else having this problem?
Which did you find most reliable? We're trying to measure the impact of our Help Center and these measurements are key to our understanding.
1
Wade Fitzner
Im with Jessica C on this. I am seeing a variation in numbers between Zendesk and Google.
Also, we use a non-Zendesk chatbot, and curious about how we are accurately able to track clicks? If they are clicking article links through a chatbot, how can we tell how the customer is getting to each article?
1
Scot Sroka
I agree with Wade and Jessica, our Zendesk Article Views and Google Page Views number is off by thousands. Would be interested in hearing more about why this is? It definitely can change the outcome of the Self Service Score.
1
Dane
Normally the discrepancy is caused by how Explore and Google calculate views. In Explore, HTTP user directly accessing the API stats view endpoint of those articles are also being counted. Our Explore system counting method does associate this with views, however, Google Analytics won't as it wasn't actually a real visit on the page.
If you want to check the real page visit, Google Analytics is more reliable. This is currently being discussed on how to avoid this type of behaviour in the future, but we do not have a near-term resolution for now.
0
Jessica Peck
Is there an option in Explore to add a self service measurement? I don't see anything like that as available reports but could we build that calculation?
1
Graham Ferguson
I agree with Jessica Peck - is there a way to build this metric in Explore?
0
Dane
When it comes to self-service, the first thing that comes to mind is Answer Bot resolution. If this is the case, please refer to Explore recipe: Answer Bot attempts through to resolutions. However, if this cannot be applied on your use case, please provide more context on your use case.
0
Jessica Ison
Hi Jennifer Rowe, Dane -- can we calculate self-service metrics using Zendesk reporting alone? Since Google Analytics is phasing out, we want to ensure we're calculating these metrics using a sustainable future-proof process. Thanks!
0
Jessica Ison
Dane just following up -- any suggestions?
0
Dane
It seems that you have already submitted a ticket for this inquiry. Please continue to work with the Advocate assigned to the ticket so that you can be assisted more efficiently.
0
Mirek Kokes
Hi Dane would you be able to provide me a step by step instructions how to create a formula in Zendesk explore which would calculate self-service ratio and gave me the e.g. 40:1, 4:1 result?
I was able to create 2 separate queries, but I am stuck and I can't make them work together because they are from 2 different datasets.
Thanks for your help.
0
Jonathan Suchin
Multiple google analytics articles linked here don't seem to be connected to an article anymore; these did not load
0