According to my studies it is less than 20%, as users need on the average six (yes six!) attempts to solve a single problem. Do you know your numbers?

Let me first explain what I mean by ‘market share’ in support. When users have an IT problem, they have several options to try to solve it. Instead of contacting the service desk, users may try to solve it by own means, possibly using self help tools. If that does not work, they might ask help from a colleague or a super user or even a family member. In this context, the market share is the service desk’s share of all support requests made by the user community.

A year ago, I did a quick survey about it, in which I asked the respondents what the shares of different sources of support in their organization were. From the results it was obvious that most people guessed the numbers: only five percent were able to give exact numbers. The survey’s key finding is below:

In the graph above, the different colors represent the role of the respondent. While all respondents are IT people, they had different roles concerning the support. Some had the role of end user, some represented the service desk which provided the support, and some were in charge of administration of the support. The bars show the responses each role provided to the questions at the bottom line: “How large is the proportion of solutions provided to users by self, colleague, etc.?”

As can be seen, the role plays a big part in the results. Clearly the service desk people overestimate their market share.

I thought the market share would be important in measuring the success of the service desk; much more important than the customer satisfaction, which can be manipulated too easily. I created a measurement tool for finding out the actual market share. The tool has been used in some companies and the results have been interesting, as well as a bit surprising. The results so far are based on a small sample of companies, but they are based on the responses of several hundred of users.

We can see that the market share of the Service Desk is somewhat smaller than the perceived estimates in my previous survey (37-57%, depending on the role). Overall the results are close to what the users estimated in the survey which is good as the results come from different respondents.

The second survey  also measured the number of attempts to solve a single problem. The results are shown in the graph below.

 A really surprising result, at least for me, was the fact that it often took many attempts to solve one problem. The average number of attempts was 6 and the median 3,5. Less than 20 % of users managed to solve their problem with just one attempt. That is a pretty dismal picture.

One explanation for the large number of attempts is that the users try several channels of support. This was also measured in the survey. The survey offered six possible channels for solving a problem, as is shown in the next graph. The channels were [1] self, [2] automation, [3] service desk, [4] colleague, [5] super user, and [6] other person.

 The results showed that 10% of users use all six and 50% use at least three support channels to solve a single problem. A typical case might be trying first self, then asking a colleague and a super user and finally contacting the Service Desk. Only half of the problems ever end up at the service desk.

Each attempt to solve a problem takes about ten minutes on average, so a lot of time is wasted. The tool measures the time lost in trying to solve the problem and also the time lost while waiting for a solution. The average for the total lost work hours was 1,5 hours per problem.

All my surveys have been done in Finland, so it is difficult to say how well they represent the ROW outside Finland. They may not be so different. I managed HDI Nordic for several years and we did a lot of surveys and compared our results to the US surveys. In my experience, a help desk was a help desk in any country we investigated. If you want more information, please contact me: aale.roos@pohjoisviitta.fi

By the way, doing surveys is a bit of hobby for me. I have created a website about a survey method, called the Three Question Survey Method. The idea is that no survey should have more than three questions. Why? It forces you to think more about the questions if you only have three. It is actually surprising how much understanding you can gain by three well planned questions.