A Different Way of Looking at Categorization

Categorization is a critical aspect of many IT service management (ITSM) processes.  Categorization helps us:

• Route work associated with an incident or a request
• Produce effective management reporting that enables further analysis or process improvements
• Translate what a consumer is telling us into something that IT can understand and do something about

There’s a lot of good advice about categorization that is readily available via an internet search.  A couple of great articles can be found right here on the SysAid blog.  Stuart Rance shared his thoughts about categorization in his post titled Improving Categorizing Incidents. Another SysAid blog post, Incident Categorization – Reasons and Consequences, discusses the benefits of a good categorization scheme and the consequences of poor categorization.

But how do you go about defining your categorization scheme? I have some ideas about categorization that I’d like you to consider.

Good Categorization Is About Balance

The most important thing to understand about categorization is that it is all about balance.  If a categorization scheme is too simple, it will inhibit work flow and trend analysis. The categories will simply be too broad or vague to facilitate effective process execution or continual improvement.

The other end of the spectrum is a categorization scheme that is overly complex. While further definition may help with more precise workflow and analysis, this comes at a cost – the up-front time it would take a service desk agent to determine the exact category of an incident or request. While good documentation may mitigate this risk, keep in mind that a primary objective of a service desk is to complete its tasks in a timely manner. Service desk agents don’t necessarily have time to read through a lengthy procedure while the consumer is on the other side of a telephone call or chat session.

Also, categories can and should change over time. As services are introduced or retired from the managed environment, there will be impact to a categorization scheme.  A good categorization approach recognizes that, and should strike a balance between the need for consistency for managing day-to-day activities with the flexibility needed when services do change.

Common Categorization Mistakes

Unfortunately, many organizations don’t take the need for balance into consideration when defining their categorization schemes. They often don’t think about how they want to leverage their categorization schemes in terms of workflow routing, reporting, or identifying improvements. As a result, I’ve seen them make the following mistakes:

• Define categories based on the organizational chart
  This might work, and seem very straight-forward…until the organization chart changes.
• Top level categories are “too technical”
  Generally speaking, the consumer has no knowledge of the underpinning technologies that are used in delivering service, and are unable to articulate an issue in technical terms. As a result, the service desk has to translate (read: guess) from the business-reported symptom to the technical taxonomy.
• Too many top-level entries or a scheme that is “too deep”
  I once worked with a client that had a six-level deep categorization scheme. I’ve seen clients that had dozens of top-level entries in their categorization scheme. If there are too many entries (keep in mind that “too many” for one organization may or may not be too many for another organization), handle times for logging an incident will become needlessly excessive.
• Assuming that the consumer “knows” the categorization scheme
  Yes, it may make absolute sense to the IT organization, but how (or even why) will the consumer know about the categorization scheme? The consumer has no idea how the scheme is defined, how the scheme is intended to be used, or even the cause of their issue so that it can be properly categorized within an IT-centric scheme.
• Using “other” or “miscellaneous” as a category
  Frankly, this is laziness! If these categories are defined, they will (and do) inevitably become a dumping ground, nullifying the reasons for categorization to begin with!

Another Way of Looking at It

Consider this – what is the trigger for categorization?  How an incident is categorized is essentially based on what the consumer has said.

So why not take a consumer-centric approach to categorization?

In my experience, most categorization schemes have been defined following an “inside-out” approach. The scheme is defined based on the IT organization’s processes and services, and is pushed from “inside” IT “out” to the consumer. So why not look at it from the consumer’s perspective? Looking at this from the “outside,” or consumer’s perspective, what should the consumer’s experience be? This is known as “outside-in” thinking.

So rather than trying to categorize incidents by “hardware” or “software” or “network” (all IT-ish ways of looking at an incident), why not categorize based on the consumer’s perspective? This might look something like “can’t print,” “error message,” “not working or slow,” or “need help.” From this top level, the service desk agent can then drill into the issue to identify further symptoms that will assist in determining how to manage the issue. Following this approach essentially defines the roadmap for helping the service desk agent categorize the issue; for example, “not working > intranet > broken link” or “need information > software > excel”.

Advantages of an Outside-In Approach

Here’s a few advantages of taking an outside-in approach to categorization:

• It improves the customer experience – service desk agents can have a conversation (rather than follow a script) with the consumer to manage the incident.
• Strikes the right balance – it allows for the needed flexibility (changes in offered services can be easily added or removed at a sublevel), while having consistency at the top level.
• Removes the ‘tech speak’ from categorization – consumers need not be intimated by a call to the service desk, as the service desk will be enabled to speak and use terms that the consumer understands.