You're Measuring the Effectiveness of Your Help Desk Knowledge Base Wrong


Over the past few years at AnswerDash, I’ve talked to many companies about their website self-service, and I’m bemused at how few of them can answer the following questions:

  • What percentage of your website visitors utilize your knowledge base? Why and when do they go there?

  • What percentage of your visitors find the answers they need? How easily?

  • What percentage of your visitors resume their original task and accomplish their goal?

These are the questions that any website self-service technology should be judged by. They essentially boil down to: How much is the solution used? Does it solve people’s problems? Does it cause people to achieve their original goals?

For knowledge bases, the answers are usually: Hardly at all. No. And, no way.

These questions matter most when trying to understand whether your knowledge base—what I’ve unapologetically called a “help island”—is worth a damn. But amazingly, few companies seem to know the answers to these questions because knowledge bases themselves don’t report the answers to these questions. Instead, they report stuff like:

  • Which articles are viewed how many times

  • How much time is spent on each article

  • How far readers scroll through each article

  • Whether readers mark articles as helpful

I call these types of metrics “introverted” because they only shed light on the use of the knowledge base itself, which tells you very little about how the use of the knowledge base relates to the use of your website—what you actually want to know.

A real-life conversation

I was in a meeting with a customer recently. The conversation went like this:

“We feel our knowledge base is performing very well,” they said.

“How do you know?” I asked.

“Because when someone views an article in our knowledge base,” they replied, “only 4% of the time do they end up sending a support question via email.”

“What percentage of your traffic ever hits the knowledge base in the first place?” I asked.

“Hmm. We don’t know.”

“How do you know that the 96% who never send email actually get the answer they need? Maybe they just give up and leave?”

“Hmm. We hadn't thought of that.”

O.K., so maybe only 4% of your visitors send email after viewing a knowledge base article, but if only .001% of your visitors view articles in the first place, who cares? If 10% of your visitors have questions but only .001% ever attempt to find answers, who cares? If only 1% of those who attempt to find answers ultimately achieve their goals on your site, who cares?

The conversation above highlights a huge gap between the introverted metrics today’s knowledge bases provide and the information companies need to make informed decisions about the effectiveness of their website self-service and their website itself.

The assumption of usage


Introverted metrics all fall prey to the “assumption of usage.” They tell you that given a visitor reached your knowledge base looking for help (that’s the assumption)… which articles did she view? how much time did she spend on each article? how far did she scroll? how many articles did she mark as helpful? And so forth.

These metrics all assume the visitor reached the knowledge base in the first place. As I’ve argued, this is an unrealistic assumption.

In my experience, I have never seen a site for which more than 1% of its visitors went to the knowledge base. Think of all those tech writers producing great content for fewer than 1% of their site’s visitors. That’s a shame.

Obviously the 1% statistic does not mean that fewer than 1% of a site’s visitors have questions occurring between their ears. No site is so well designed, so free from confusion or frustration, so perfectly in harmony with user cognition and behavior that 1% is the sum total of visitors who ever have any questions. Droves of website visitors, above 50% on many sites, have paralyzing questions (ask any Comcast subscriber who has tried to solve a problem online), but knowledge bases still get almost no traffic. That’s because of a few reasons:

  • People realize knowledge bases take effort to use. People don’t want to dig.

  • People realize knowledge bases often fail to yield helpful content.

  • People realize knowledge bases are going to take them away from their current task.

Fortunately, all of these problems are solvable with self-service that is easy to use, in context, and full of relevant content. AnswerDash is just such a solution.

Contextual self-service that gets used

Knowledge bases don’t get used because of the “help distance” between the user and the content. Too many clicks, too many page navigations, too much typing, too much time, and too much effort lies between the user and the help they need.

Contextual self-service solves this problem. What is contextual self-service?

Contextual self-service puts the knowledge people need not off on a “help island” but in the same context as where users’ questions arise. A user’s point-of-action, their moment of need, is where they are trying to get something done. This is the best place to deliver answers when questions arise.


Live chat is an example of contextual help. But it is not self-service. It is assisted service. That means users are subject to an agent’s pace and protocols. It takes 14 minutes 10 seconds for the average live chat session to resolve. The power lies with the agent. And live chat can be expensive for companies to staff.

AnswerDash is an example of contextual self-service that puts point-and-click answers right at the user’s point-of-action and enables control to stay with the user. Small, consumable “Twitter-sized” answers give users just the information they need with links to full knowledge base articles for users who want to know more. AnswerDash puts instant answers within reach of a couple clicks and usually no typing. The reasons people don’t like to use knowledge bases evaporate with AnswerDash. And the results prove it: Instead of 1% of a site’s traffic, AnswerDash is used by 5-20% of a site’s traffic. More people are helped more often and with less effort for both the company and the user.

Answering the questions that matter

Because AnswerDash is not a help island but is in context, on the site itself, it can answer questions that matter to site owners. These aren’t introverted questions. They are “extroverted questions,” going out beyond the self-service solution itself, relevant to the site as a whole.

These are the types of questions that AnswerDash can answer for a site:

  • What percentage of visitors to your site have questions that need answering?

  • What are the top questions arising on your site and where exactly do they arise?

  • What percentage of your visitors find answers to their questions?

  • What percentage of your visitors successfully convert to a sale after viewing answers?

  • What percentage of your visitors never find an answer they need and abandon?

  • How can you improve your site to prevent questions from arising in the first place?

It’s pretty clear that the answers to these extroverted questions are far more valuable to site owners than knowing how many minutes people spend on a knowledge base article.

When it comes to website self-service solutions, we must remember the goals we had in adopting them in the first place. We want to help our visitors achieve their goals on our websites. Anything that complicates this picture deserves to be questioned… and given a new answer.