The other side of outliers
Outliers – any analyst worth his or her salt knows that even a few of these abnormal data points can skew the results of the most meticulous analysis, rendering any conclusions suspect. Therefore it is standard practice to comb through and cleanse any data set of values which are far enough from the other observed values, mitigating any damage they can do before the analytical heavy lifting is done. While it is often the case that outliers do detract from the quality of any analytic work, dismissing them outright belies their hidden value. Outliers, when paired with solid qualitative data, can lead us to knowledge we would not have otherwise reached.
We realized the potential benefit of outliers during a recent engagement with a pharmaceutical distribution company. A portion of the project had us measuring how the client’s sales force was spending their time. In prepping our data for analysis, we noticed a few sales managers had reported spending an inordinate amount of their time sourcing and qualifying leads. These managers were spending over double the time on these activities than their benchmark role would suggest. Rather than presume these responses to be invalid and toss them out, we took a closer look at who was reporting these extreme values. Sure enough, the respondents showing as outliers shared a primary customer segment. Interview evidence corroborated the fact that lead generation in this segment was weak, and that managers were supporting their reps with lead generation and qualification in order to enable them to maximize their engaged selling time.
In the end, we recommended that our client strengthen this segment’s lead generation function, allowing their higher-priced sales managers to concentrate their time on coaching and supporting their reps, a more valuable use of their time. Alone, the anecdotal evidence we heard in interviews would not have been enough to support this recommendation. When paired with the time profiles of our outlier managers, however, we had enough to convince our client that this was the right move.
About the Author
Mike Pap Rocki is a business analyst at AGI’s San Francisco office.