ICMI is part of the Informa Tech Division of Informa PLC

This site is operated by a business or businesses owned by Informa PLC and all copyright resides with them. Informa PLC's registered office is 5 Howick Place, London SW1P 1WG. Registered in England and Wales. Number 8860726.

Advertisement

Get a Clearer Picture of Hurdles to FCR

Whenever I ask, “What is the FCR in your contact centers?” I often hear, “Oh, it's doing well. It's up from 71% to 73.5%.” But when I probe with “How do you measure FCR?”, the answers are often vague or evasive, including “If the customer doesn’t get back to us on any issue within seven days, we consider that the interaction was resolved.”

Does this make sense? Of course not! FCR has to be measured much more precisely, ideally from the customer’s point of view – not immediately after the contact in an IVR or web survey, since they often don’t know yet if the situation has been fixed – but by asking them later, timing the questions based on the intent or reason. More on this later.

Then, when I ask, “What is the FCR in your chat bot or app?”, I often encounter shrugs or scratched heads followed by, “Well, it's about 25%.” Here they might present stronger evidence measuring FCR in these unassisted channels by tracking the percentage of customers who do not need to speak to an agent right after the chat bot or within an hour or two after being on the app. But this misses those customers who simply give up and don’t bother to get connected, and those who bring a different intent or reason to the agent.

Let's look at it from the customer's point of view: If a customer uses your chatbot or app in order to get something fixed or to find an answer and only gets a 25% success rate, doesn't that mean that it's a 75% failure? And if they wind up speaking to one of your frontline agents and they report a 73.5% first contact resolution, doesn't that represent a 26.5% failure rate? Plus, if they started online and then tried the contact center, that’s already a second contact so measuring the assisted FCR by itself is meaningless!

Try this on for size: If your FCR is less than 100%, then you are failing your customers and your frontline staff.

But don't despair! There's a lot that you can do to move the needle and increase FCR as close to 100% as possible. Here are seven steps that you can use to shrink the failure rates:

Break down the customer contacts for unassisted and assisted channels to 25-50 MECE (mutually exclusive and collectively exhaustive) intents or reasons, as my coauthor and I shared in our most recent book The Frictionless Organization. This ensures that you will have the right level of granularity and the appropriate assignment of ownership across the organization. Stay away from averages and look for the ranges and “long tails”3.

Use AI and Big Data to produce detailed FCR by intent or reason instead of surveying customers or applying broad brush metrics like “no call back within seven days”. Start using 100% minus FCR as the “failure rate” and ask yourself what percentage of failure is acceptable for your website uptime or delivery performance or software bug4.

Start with the upstream unassisted channels, chat bots and apps or portals, and IVR systems if you’re using them for automation. Create an FCR-equivalent metric (some folks call it “containment rate”) for each unassisted channel. What you'll discover with the unassisted channels is that a big reason for the low FCR is that you're trying to provide too much to the customers and too many options for them to consider, especially considering that most customers today are using their mobile phones, which has a very small window to be able to discern these different options. Instead of offering every possible alternative and option in the unassisted channels, it’s important to focus on those that have the highest cost and the greatest customer pain as we also described in our most recent book using the four-box “Value-Irritant Matrix”.

Analyze the FCR in your contact centers for each intent or customer reason. You will discover that some of the intents might have FCR in the high 90s, while others are significantly lower, more like 25 to 40%. It is most likely that these low FCR intents also have the highest customer churn and lowest NPS or worst Customer Effort Scores (CES)5. Be sure to check! Examine most closely these low FCR intents using the classic Ishikawa fishbone tool; you can break down the reasons for low FCR based on process, knowledge content, agent training, customer education, and other buckets. Then it's important to split each of these buckets into component parts and begin to build a weighted average for each root cause in order to determine where to focus energies.

Determine which of your agents is producing the best FCR results and which agents are doing poorly, meaning that they are creating repeat contacts. Related, figure out which agents are able to resolve previous agents’ inability to resolve the issue. We define repeat contacts as “snowballs” rel="noopener noreferrer" that need to be “melted”. This will reveal the effectiveness of agent training, coaching, and motivation that you can use to learn from the best performing agents and apply across the entire frontline workforce. It also will begin to reveal the quality and the usefulness of the knowledge articles that are available for agents to help the customer through the resolution process.

Start using AI-based tools that are generically called “Agent Assist”. These tools tee up for agents the best-fit response or offer or solution based on previously successful resolutions, thereby shortening the agent's search time and the overall handle time and producing significantly higher customer satisfaction rates.

Report FCR on a weekly basis overall and on the lowest unassisted and assisted FCR by intent and by agent. Share widely, especially with your digital tool dev teams and frontline agent trainers and coaches. Associate FCR with the other CX pain metrics, again by intent, including churn, NPS, and CES.

See if these tips help you quantify FCR at a more granular level and better address the sticking points.