For most contact centers, nearly one-third of inbound calls are repeat callers who weren’t satisfied the first time they called. More often than not, the antiquated switches that contact centers use don’t do a great job of reporting on the true first-call resolution (FCR) rate in a given center, so the problem may be even worse than it appears. Why is this?
Think of the iceberg analogy than Lean practitioners often use to explain cost of poor quality and the hidden costs that fall below the “water line.” What rises above that line is easy to quantify – it is the lowest-hanging fruit, so to speak, even though it is highest on the berg. Below the line, especially in contact centers, are problems like repeat callbacks, which is the opposite of FCR.
In pre-Six Sigma days, FCR at a call center was a blue-sky, nice-to-have metric, but nothing that managers held agents accountable for achieving. Instead, they opted for metrics focusing solely on time, like AHT (average handle time = talk time + post-call wrap up and any hold time during the call) or ASA (average speed of answer = how fast you pick up the phone).
What these metrics actually do is incite bad behavior. For example, an agent can keep both AHT and ASA low by just picking up a call and immediately hanging up on the customer, something many of us have experienced when we have phoned a call center. The agent AHT is an average of the daily AHT for a given time period, so naturally, if they hang up, they will have an AHT of a few seconds. Mix that into the normal AHT of an agent and, suddenly, their monthly stats look the lowest of any other agent.
See what I mean by inciting bad behavior? We like to call this “agent badness,” though personally, I think it is more of a sign that the agents are not getting adequate coaching (but that is for another article). AHT is influenced by multiple, equally likely root causes, so pinpointing improvements will only frustrate your Black Belts. I do not suggest contact center AHT projects for Green Belts ever, unless they are working on addressing a component of a larger Black Belt project.
By shifting your objective to improving FCR, however, you will have a much easier time proving and sustaining your results, namely reduced cost of goods sold and improved customer satisfaction ratings. A focus on FCR will also reduce customer churn and improve agent morale.
First, it helps to understand the inputs that affect FCR performance:
- Empowerment – Agents frequently don’t have the authority to resolve an issue, even when the solution is obvious. This results in call escalations to a higher tier, with increased hold time and abandonment. It also means the next callback will be an escalation call that ties up supervisor time.
- Coaching, not training – The agent may not have sufficient coaching time or ability to effectively deal with the customer call. Training has been proven to not move the needle at all whereas coaching is very successful at driving higher FCR rates.
- Information access – Agents need to find information more easily to provide answers or actions for customers. When the information is unavailable or difficult to access, sometimes agents will guess at answers or fail to provide them, both of which can lead to a callback.
- Back-end systems – These systems might not be up to the task. If the agent makes an address change, but it doesn’t propagate through the systems, then the customer will call back.
- Customer behavior – You need clues into customer perceptions and why the repeat calls are happening in the first place. This can help, for instance, when you discover customers are calling back trying to get a different result if their account is being suspended.
Here’s how to get started with Six Sigma DMAIC (Define, Measure, Analyze, Improve, Control) project to improve FCR:
Step 1. Define: Conduct an information-seeking drill-down with subject matter experts (SMEs) who are proven resources in the center. This can help determine the objective and scope of your project. Those SMEs should also be a part of your project team.
Step 2. Measure: Track calls for a 30-day window in an Excel spreadsheet, if you do not have a more sophisticated system.
Step 3. Analyze: Flag repeat calls by analyzing agent logs and any interactive voice response (IVR) system data. (If you have an IVR or any other self-service tools, they should be included here). Identify whether an issue was resolved on the first call and, if not, the actual reasons for repeated calls. Look at all customer interface points because if the customer first contacted you via self-service systems, that counts as a call.
Step 4a. Improve: Make certain you provide a measurement system of record to measure FCR from the first day forward. You need useful information to conduct further root cause analyses and provide data for your control charts. What’s one ideal outcome? Could it be one where you continue to use the technology now in place, but make it work better for you? Or would it be better to leverage new technology?
Step 4b. Improve: Perform a measurement systems analysis (MSA) once that system is in place on IVR transactions and track where calls go once they are handled (defined as “the call is answered”). Look at calls and how the IVR is handling them. Many customers merely opt to zero out. That choice, however, isn’t giving you interpretable data. Callers choose to zero out for numerous reasons, including their desire to speak with a live agent.
One method of measurement can be a standard question that agents can ask the customer at the close of a call: “Is there anything else I can do?” or “Is this what you wanted?”
These questions allow a crude estimate of FCR. The best way to measure FCR as part of Improve is to add a question to your post-call survey (if you have one): “What your call successfully resolved the first time you called?”
If the answer is no, follow-up that question with a series of options, asking the caller why wasn’t it resolved during that first call: “a) Agent lacked authority; b) Agent lacked knowledge; c) etc.”
Step 5: Control – Build control charts around the “done in one” concept (a.k.a., FCR) and publish, like we do, to an intranet to which the agents have access. Make sure you multidimensionalize the intranet to allow agents to drill though center metrics to their individual agent FCR score.