Sturdy Statistics Case Studies: Sales Director

Takeaway:  Our model cuts through arduous and time-consuming manual review, instead enabling managers to discover insights quickly and efficiently. Schedule a demo to unlock your unstructured data today.

Another of our users is a sales director for an online appliance retailer. He had metrics on each of his sales agents, and he spent several hours per day reviewing sales calls in order to understand what his top performers did well and how his underperformers needed to improve. Such work was onerous, and despite all the time he spent listening to calls, he was still concerned about “unknown unknowns” — he knew there were more valuable insights in his data, and that he was missing them. He needed a more efficient system to review his calls.

Sturdy Statistics ingested three months of calls, along with metadata associated with each call: the agent name, the call length, and whether or not the call resulted in a conversion. Our technology automatically converted each call into a structured set of topics, enabling us to show the difference in content between any two groups of calls. Thus, we could show things like:

• how calls to agent A differ from calls to agent B,
• how calls which resulted in a purchase differ from calls which didn’t,
• the differences between short and long calls which converted,
• etc.

We used this data to build a custom dashboard for the sales director. The dashboard allows him to monitor the strengths and weaknesses of each agent. Among other things, he learned that one agent excelled at technical conversations: when this topic appeared on his calls, it substantially increased the probability of a sale; for most other agents, this topic slightly decreased the probability of a sale. He used the dashboard to find examples of conversations with this topic, and to coach his other agents how to handle this type of question.

Another interesting result from the dashboard confirmed something the sales director already knew: calls discussing simple facts such as product dimensions or features almost never result in a sale. Callers with such questions are not ready to purchase; they are still shopping and are simply looking for more information. Moreover, such calls were new: product specifications had formerly been readily available on the webpage, but a recent change to the website had buried this information. Thus, this website change drove a number of calls asking for simple specs. The sales director knew from his agents that this had been a problem, and that information-gathering calls were consuming a lot of his agents’ time. This led to longer hold times, and likely, lost revenue. However, he had been unable to convince the rest of the company about this problem. Our results enabled him to quantify the cost of these information-gathering calls, in both hours and lost revenue, and convinced his company to place product information more prominently on the webpage.

Armed with our data and our dashboard, the sales director became both more efficient and more comprehensive in his analysis. Rather than blindly reviewing random samples of calls, he now uses our technology to actively monitor every conversation. He can quantitatively analyze all of that data to see which topics, from each of his agents, abet or impede sales. And if he finds something that requires coaching or review, he can jump immediately to that portion of the transcript. Thus, he has more time to spend coaching his agents, with more actionable feedback and more pertinent examples. He also has quantitative data to motivate changes to the sales script, or even to motivate other changes to the company, such as the website.