As I mentioned earlier, I will be speaking at the MIT Enterprise Forum at MIT tomorrow.

I’ve got a few thoughts jotted down, mostly on the key issues I see (and have personally seen before I defected to consulting) startups face when they finally get to thinking about support. Usually it is let’s build it and sell it (sometimes in reverse order!) and if support does come into play, it usually involves engineers doing support until too much of their time is ‘wasted’ with break-fix.

It was a pleasure building out a global support model at Akamai providing complex, award-winning support when we had over 15,000 servers to support and almost a million hits a second on the network. The whole philosophy was to build a support model to scale in an environment where a support call for break-fix was too late. By the time we figured out where the needle (your IP address) was in the haystack (among the million that *second*), it was too late. That was just to start the troubleshooting process…

Some of the philosophies are documented in my book, and more of it will be in the form of war stories we will talk about in my upcoming workshop, ‘Be the Voice of the Customer‘ on June 5th just outside Boston.

In my April newsletter, I wrote about how companies are beginning to relalize that the ‘tiered’ support model is counter-productive to customers and to staff. Given the number of responses to it, I’ve decided to reproduce it as a blog for a more widespread discussion…

————— (From the April 2008 edition of my newsletter, The Verghis View)

Ah, is there any part of the service experience that is more reviled by savvy customers* or by upwardly mobile support staff than the dreaded ‘tiered’ model of support? It made sense back when our customers didn’t know much about technology, and we knew best. When we first built these models, we put friendly but clueless support people on the front line to protect expensive resources (who often were the reason the problems existed in the first place) from ’simple’ questions.

From a customer’s point of it, the tiered model is horrible. They’re always forced to start at the bottom answering basic questions. There is no acknowledgement of their expertise or the context of their query. For example, all calls are treated pretty much the same, whether they are calling because they can’t print a document to read on the plane, or if they’re having trouble printing a proposal for a multi-million dollar project that is minutes away from the submission deadline. Every caller has to jump through the same hoops until someone determines the context and urgency of your issue.

From the staff’s point of view, the tiered model is treated as a necessary rite of passage. Traditionally, rookies had to spend a certain amount of time dealing with repetitive, low-level questions, then either defect to another group at the first opportunity, or if you survive, move up to Tier 2, which offers more interesting, less routine work.

One of the emerging best practices in support involves getting rid of tiered support for live calls completely. In this emerging model, callers who need assistance are connected with a highly skilled generalist, someone who understands both your business and the context of your call. He or she is paid just as well as the experts they may escalate to, because their skills are as just highly valued by customers. These professionals need to know a lot about the customer’s context and bring the right level of resources to resolve that particular issue.

I call this type of support ‘Savvy Support.’

To accommodate this change to a ‘just in time’ model, support tools must deliver more than just traditional CRM touches. To connect with people who have talents and interests not encompassed by their job description or title, these tools must also layer in context and social networking aspects as well as nuanced linguistics capability. This ’savvy support’ approach will force us to break away from the ‘if this person has a title they must be good’ mentality and embrace the ‘everyone has talent, most of it is not apparent in titles and traditional organizational charts’ mode of finding out who has the right skills that can be applied at the right time to the right issue.

This Savvy Support model works both for the customer and the support team. The customer gets a knowledgeable advocate who can help them resolve issues on the first call (if they even need to speak to a person). Support staffers win because they love to solve problems, rather than be enforcers of procedures and rules. This model is exceptionally customer friendly because the important metrics are things like customer impacted minutes, not internal efficiency metrics like time to answer.
However, a couple of prerequisites are required for this approach to work.

• The first prerequisite is leadership, because conventional support wisdom must be turned on its head. Leaders will need to work with HR and other departments to implement changes in compensation, reward and metrics.
• The second prerequisite is a support team that relishes solving new problems, rather than becoming complacent. It will no longer be enough to be good at handling routine (or known) problems. In the Savvy Support model, recurring issues are ruthlessly eliminated and fixed so they don’t reoccur.
• The third prerequisite is a solid organizational culture of sharing knowledge, such as Knowledge Centered Support.

Are you ready to learn how to get out of your comfort zone and smash your tiered support model? Call me at 617.395.6613 or toll-free at 800.494.9142. Another way is to attend my upcoming June 5th workshop, “Be the Voice of the Customer”

I was invited to post on the MIT CIO Corner, so I put my last newsletter article titled ‘Perception is reality, but what about the other way around?’ online. You can read the post here.

Have any of you considered (or adopted) prediction markets for service and support? (Prediction markets are markets where participants trade in contracts whose payoff depends on unknown future events. For example, how low will the US dollar go against the Euro?)

In the service and support world, an example could be where you create a market to see how internal staff, partners and *gasp* customers predict whether your product or service offering will be bug free and delivered on time. If developed properly, I can envision this support prediction marketplace with helping service and support leaders get more relevant in critical conversations about the importance of the customer in an organization (beyond lip service).

One of the best known examples of prediction markets is the one run by the business school at the University of Iowa — Iowa Electronic Markets. You can put real money in and participate in academic research.

For those of you interested in this prediction markets, there is a great article in the Journal of Economic Perspectives that explains this concept nicely.

So, are you doing anything like this in your organization related to support? If you could, what would the key questions you want answered (and predicted) to be?

As part of a white paper I’m writing, I have been doing some research on ‘mashups’. (Web applications that combine data from multiple sources in one user interface.)

The idea behind mashups is simple. As anyone in IT can tell you, there are far more demands for their services than there is time available to get to them. No wonder people find corporate IT not moving fast enough for their needs, particularly if their needs are simple and highly situational. Enter mashups. The idea is that you can create simple ‘widgets’ where you combine content and services from within the corporation and out on the web to create new, simple applications. This can be created by the end user, not IT.

Two interesting mashup makers are available from IBM and Microsoft, called QEDwiki and PopFly respectively. If you don’t know what mashups are, check out the IBM video on YouTube, and the 30 second overview from Microsoft. Will give you a good idea of what is possible…

Are you using mashups to help your customers? Let me know how!