Ecommerce and Entrepreneurship Blog | About | Contact | Store

How to Use Metrics Based Customer Connection Rates

Posted on June 16, 2009 by josh

We were recently able to upgrade our phone system. We had spent time, energy and effort to build a better system under the assumption that it would improve a number of important metrics. More importantly, the new phone system made it easier for us to see those metrics. Consequently, I was able to get better reporting on phone utilization, especially as it related to customer service. One of the first things that struck me was our maximum load. We had gone from 8 lines available to 23 lines available. With 8 lines, extra calls would roll over in to voice mail, creating havoc later in the day as we scrambled to try to call people back. Many times, people would call two, sometimes three or more, times to try to reach us. Also, with eight lines, it made it very difficult for us to make outgoing calls during peak periods. With 23 lines, surely we would handle the load. Not so fast! Over the first few weeks I checked our maximum load on all 23 lines and, on several occasions, we had exceeded our maximum load during peak calling periods! We nearly tripled our call capacity and still had overflow. Thankfully, the overflow is not frequent enough, at this point, to justify another 23 lines.

Alrighty then, what about our connect rate? Given all the customers that try to reach us, to what percentage do we actually connect? I was more struck by this metric than our load. Let's just say it was less than half of what we would want it to be. It turns out, we were literally turning away hundreds of calls a week, probably mostly due to long hold time. Even more striking, our average hold time for a connected customer was approaching 30 minutes! 30 minutes?! Dreadful! Sales customers simply will not hold for 30 minutes to buy something. What made this even scarier was the fact that our mix of callers (known by the queue that they select when calling) was more than two-thirds sales calls. We were turning away hundreds of sales a week. This means we were turning away customers who wanted to buy something (and were probably at the end of the buying cycle) in order to provide support to customers we had already acquired.

OK. Now we know. What do we do? We could hire 10 more reps. Nope. Don't have the budget. OK. We don't have an unlimited budget. What else can we do? We could slow down marketing and sabotage our own SEO to get fewer sales calls. Nope. That's dumb. So what do we do then?! How can I get more work out of existing resources? How can we add hours to the day to get more done?... wait... Eureka! We needed more hours in the day! It was ultra-clear that our call and chat volume was much higher earlier in the day than late in the day. Once the clock gets to about 3:45 pm, things tend to slow down. With reps struggling during our rush periods to perform administrative and follow up tasks, it would be near impossible to be more efficient during regular 8:00 am to 5:00 pm business hours. So, why don't we have some reps come in at 7:00 am, instead of 8:00 am, to finish any lingering issues from the previous day and get a head start on the new day before we start taking calls and chats? We could even perform some proactive functions that prevent customers from calling for support reasons in the first place! Why didn't we think of this sooner?! Oh yeah, we didn't think of it sooner because we didn't have great data to stare at.

The effect of having half of our customer service team come in one hour before business hours has been exponential in its positivity. We still have load issues, two or three times a week, during peak times we exceed our capacity. However, our connect rate has doubled! Our hold time has gone from nearly 30 minutes on average to less than 10 minutes! Interestingly, sales are up, but we have reduced our volume of calls and chats by about 10%. Fewer calls and fewer chats, coupled with increased sales, is a strong indicator to me that we are servicing customers more efficiently. We still have a great deal of growth opportunity, but one simple scheduling decision has drastically improved our ability to service customers and, as a positive consequence, also greatly improved morale in the customer service department. Go team!


blog comments powered by Disqus