How to Use Kustomer Data to Help Forecast Headcount

4 min read
Share

How to Use Kustomer Data to Help Forecast Headcount TW

As COVID-19 cases began to spike in February and March of 2020, the economy slowed. Many companies were faced with the difficult decision to layoff or furlough a percentage of their workforce to stay afloat. As we move into the summer months, there have been modest gains in economic activity and employment growth. Reuters reports that approximately 25% of private-sector jobs have since been recovered out of those lost in March and April. Still, recovery has been slow as many contemplate future waves of the virus.

Considering the uneven terrain of our current economy, workforce management has become even more critical to maintaining profitability. It also promotes the health of your customer service team. If you’re running a skeleton crew and looking for ways to justify an increase in headcount for your team, read on.

How Kustomer Data Can Help

There are a handful of important metrics within the Kustomer platform that can help you understand whether your team is over- or under-staffed: inbound messages, average handle time, and agent capacity. For the purposes of this exercise, we’ll focus primarily on a single channel: chat.

Here is the major question to consider: what does the data tell us about staffing needs and restrictions? Additionally, how many agents do we need to staff so that all chat customers are served immediately?

Let’s say that you’re an up-and-coming retailer in the Atlanta area. You currently have a 10-person team that handles all of the incoming chat conversations on your website. Each of these agents is trained to handle five chat conversations at a time. Collectively, their average handle time is five minutes.

Every agent works an eight-hour shift. They take multiple breaks throughout the day that add up to approximately one hour; they work for approximately seven hours per day. Thus, every agent is capable of performing approximately 420 minutes per shift (seven hours is equal to 420 minutes). Sixty minutes divided by an average handle time of five minutes means that each agent could theoretically complete 12 conversations per hour (if not multitasking). If we multiply that number by agent capacity (five, in this case), we can speculate that an agent can handle 60 conversations per hour.

If an agent can resolve 60 conversations per hour, and each of those conversations has a collective average handle time of five minutes, then an agent is capable of performing 300 minutes of work in an hour (in the eyes of our reporting). Finally, when thinking through the amount of work an agent can handle in a shift, that number is 2,100 minutes of chat work (300 minutes multiplied by seven hours).

As the lead of this team, you begin by pulling the average inbound messages per hour within the Conversations tab of your Standard Reports. Break up the data by day of the week. You notice that Mondays, on average, see a typical volume of 6,000 inbound chat messages. Again, if we multiply the total number of messages by our average handle time (five), this represents 30,000 minutes of chat work that needs to be completed on each Monday. If we divide those 30,000 minutes of chat work by the 2,100 minutes that an agent is capable of completing each shift, we can guess that we need approximately 14 agents working on Mondays to serve all of the chat customers as they arrive.

You can replicate this process across all days of the week, or certain seasonal spikes, and even apply this method to other channels. With further calculation, you could provide an hourly view of necessary coverage for inbound chats as well.

One final disclaimer: the important thing to remember here is that we are using past performance to forecast the future. Thus, it will not always be a perfect predictor of future staffing needs. It’s important to regularly monitor the ebb and flow of inbound messages to ensure that your team is adequately staffed.
 

Ready to discover how AI + Data + CRM equals customer magic?

View PricingRequest Demo