As I’m done bashing MBOs (for now), I might as well revisit my long-standing concern with the ubiquitous red/yellow/green traffic light metaphor. While the metaphor is intended as a simple summary of performance (green = good, red = not good), for most business situations three levels of performance are not enough to truly judge results. A simple – but compelling – example of this disconnect came from a recent exercise in helping a sales team design their MBO targets.
This team designed a dashboard that showed each region’s and individual contributor’s achievement versus sales quota (ie. the target). They recognized performance should be compared against quarterly targets, not just yearly ones. However, because quotas varied widely by individual and region, sales management didn’t want to display actual or target values but rather an “at a glance progress towards goals.”
When I arrived on the scene, the team was struggling with how to systematically convert these values into colors without relying on manual intervention. Using my terminology, they needed a grading system.
The team had used a simple device that leveraged the fact that the business plan was based on an average sales rep achieving 85% of their quota: 100%+ achievement was green, 85%+ was yellow and below 85% was red. However, sales organizations typically have an incentive or accelerator that is invoked for achievement at a threshold (such as 120%) above quota. This grading system failed to distinguish between these overachievers and those who met their quota.
After a brief discussion, we concluded that this scheme also masked an issue with low performers. In this organization, some fraction of the sales reps achieve less than 50% of their target for several quarters in a row and are put on a sales improvement plan. Failure to fix their performance can lead to termination. Grading these low performers the same as those who achieved 84% of their quota (nearly the expected performance) was both unfair and inaccurate.
Given the predominance of the traffic light metaphor, the team continued to struggle to come up with a three color grading system until I reminded them that a traffic light is only designed to cover two outcomes: proceed or stop. The third color – yellow – signifies a transition from green to red but doesn’t create a new potential outcome. More on this here.
In our example, we wanted at least three separate outcomes (performance improvement plan, track earned commissions, engage accelerators) and possibly a fourth (notify achievement of quota). Clearly, we needed more than three states and more than three colors. We explored a variety of metaphors (Harvey balls, emoticons, thermometers) but the organization wanted something close to the classic red/yellow/green traffic light. In the end, we decided to keep the colors but get rid of the stacked lights, as in this design:
The corner cases are clearly handled: over-achievement is dark green while improvement plans are dark red. Furthermore, those “close to” the expected average are graded yellow, even though they haven’t reached it yet. This scheme is simple enough to explain to an individual contributor but provides enough granularity to judge performance.
What do you think?
To me; Traffic Light is a way to indicate where there is a problem and where to pay more attention. For example; if I see a “red” light on a KPI, I would want to drill-down to get the full picture. I like the extra granularity your solution provides. However, 12-15 of these on a single dashboard might get overwhelming. Traffic light can point you quickly to the area where performance is lacking. A 2nd or 3rd level of granularity should also be provided. How much time you spend drilling down to the contributing factors and root causes, should be left up to the user
Thanks Mo. I worry about 12-15 KPIs on a dashboard and instead recommend 5-8. I don’t think humans can understand the trade-offs between that many KPIs. Also, as I mention in my next post, I don’t recommend looking at red KPIs but rather those that are trending down.
Adding a trend indicator to that graphic looks like it would be pretty easy and not too cluttering – just put a bit of a “motion blur” effect to one side of the needle or the other to indicate the direction of movement. The wider the blur, the more movement of the trend over the unit of time.