Reply
Fri 11 Jul, 2014 06:45 am
Let me start by saying I am a super novice with statistics. I had some classes in college but that's been years ago. Here's my problem: I am an operations manager in a call center environment. Our client is currently very concerned with our outbound call percentage. This is computed as the total outbound minutes/(total inbound minutes+total outbound minutes). They are comparing us to the enterprise which comprises of a total of four sites, including ourselves. We are an extremely small group with only 10 agents taking calls. On a weekly basis, we take around 262 calls and the enterprise takes around 3,815. Based on those numbers, I'm assuming the enterprise is around 150 total agents but the true number is unknown to me.
My concern is that they're pushing my group to be at enterprise level on a weekly basis. The enterprise generally runs around 30-32%. We usually are close but are generally a bit higher than enterprise but do drop below on occasion. This is an escalation group which deals with technical calls that are unable to be handled by the tier one agent. This results in necessary follow-up with customers, sometimes multiple times. I don't want to sacrifice customer service by not allowing these agents to make the outbound calls and with our small slice of the total volume, I feel like it's statistically unreasonable to expect us to be at the enterprise level. I feel we'll be way more volatile and need help understanding how to present this to the client.
I was thinking potentially a control chart may be the way to go but , again, I have no clue. I would assume there's a way to calculate a range that as long as we fall into from the enterprise, we should consider it normal. Would this be one standard deviation from the mean? Any insight is appreciated.
Thanks,
Tom
@TSteveH,
The problem you have is that over time, you should regress to the mean, meaning that if the enterprise average is 30-32%, your small group might run 28-34% over a small interval, but as you average over longer periods you should be doing 30-32% as well. If your sample size is 262 and the enterprise is 3815, you should be pretty close to each other. If your sample size was 20, you might have a better argument. My advice to you would be to take the concern seriously and look at your operation to see why your response times are longer. Does your team have the right tools? Do they understand the requirements? Failure to meet the same standard as other sites over the long term could mean job losses at your site so you should probably jump on it rather than try to explain it away.
Stumble It!
Tweet This
Bookmark on Delicious
Share on Facebook
Share on MySpace