To test the effectiveness of the approach to measuring your actual use of support analysts, we compiled a test year by pulling out all the numbers for 2002 and plugging them in to review the results. Though the figures supported the conclusions we had already drawn from our existing experience and team monitoring, the new system is more objective and repeatable. It’s also more transparent so the analysts can understand and buy into the way in which they were being monitored and rewarded.

Another lesson we learned from the test exercise was that it let us fine-tune the percentage targets. When you implement the system yourself, you’ll want to set levels based on your test quarters and on your own business – you may find that the percentages are too difficult to achieve or, as we did, that the targets were a little on the low end.

Calculating the bonus
We recently completed our first year based on this method for awarding bonuses and we made sure to include a team target (as you can see from the spreadsheet in last issue’s article). Our goal for team targets was to discourage individuals from holding on to an excessive number of cases, in the hope of closing them to boost their individual numbers. This would clearly go against the best interests of the customer and the team targets ensure that it’s also goes against the best interests of the team and of each individual.

We implemented this by setting a team target as a requirement to activate the individual bonuses. The team target is determined by dividing the total “Actual cases closed” by the total “Available cases” (i.e. if one analyst answers more cases, then somebody else will have answered fewer but the overall team score will still remain the same, so there is no incentive for individuals to “hoard” cases).

Once the team goals have been met, the bonus level each individual receives is determined by their Average utilization score (i.e. ensuring that their availability to take cases is still factored in to their reward).

The table above shows how the analysts are heavily encouraged not to drop below 100%. However, they are also rewarded in smaller increments for over-achievement.

Some initial experiences

It’s important to validate very high scores and, even more importantly, very low scores. One argument I often receive is that two weeks’ training or vacation severely limits analyst’s ability to meet the individual target. This is only true if a very large number of cases come in (i.e. more than normal) while that analyst is unavailable. That said, in a previous quarter, Analyst 2 was out of the office for six weeks out of 12 and still managed to meet her 100% (Average) target, this was due to the effort she put in while she was in the office.

Once analysts are trained and really understand the measurements, they can work towards achieving them. One major change we saw in the team was that no cases were turned away (from call administrators) or passed along without a genuine reason. Analysts took ownership of cases and worked through solving them. It also encourages proper tracking of time on and off the desk and of cases taken.

It’s important that team leaders monitor case ownership to ensure that analysts only hold onto cases when it benefits the customer – not just to benefit their statistics. To make this work, the team leaders are responsible for moving cases between analysts, so the customer gets the appropriate resource and knowledge, and analysts don’t lose out by giving up a case.

We also monitor self-logged cases. If an analyst is asked more than one question, two cases are created. This works in practice but it can also boost analysts’ numbers.

You also need to track the time not open for taking on new cases. This is the time used on other tasks like building a test environment, working on the knowledge base, etc. We track this and deduct anything that would prevent the analysts from receiving and working on a new case from the “Hours worked” column.

Evening out the load

How do we manage to get an even spread and keep it that way? By evaluating and measuring our call administrators and team leaders on case distribution and any variance. This makes it their responsibility to ensure that the analysts meet their targets and, by doing so, leaders meet their own.

Before developing this approach, I envied anyone lucky enough to have business process that could be measured through a basic standard utilization method -- not any more! This method has given us hard numbers to support what we thought we knew about individual performance but having objective and quantifiable information makes a world of difference.

About the author

Matt Hinchliffe is Director of Support, Americas for CODA Financials, Inc. - the US-based division of global Financial Intelligence solutions provider, the CODA Group (www.coda.com). His background covers a full range of experience in software and support roles: from support analyst to trainer and manager, including a “ground-up” reorganization of CODA’s Support Centre in Golden, Colorado that has produced significant gains in efficiency and effective customer service delivery.

When implementing any new measurement you should ensure you are aware of the impact on your organization, customers and employees. The SSPA, author, or CODA may not be held liable should the application of this, or any other measurement negatively impact your business.