Managing FCR at an agent level

Do we really need to be proactive with it? I only ask because I think there are inherent dangers in measuring at such a level. Agents could tick boxes, complete fields which suggest their FCR rate is higher than it really is to preserve a job or enhance a bonus.

I think there is a danger with the FCR rate if the only reason is to check quality of support. Surely we need to put in training and knowledge systems so this is a given. Only then can we move forward and use FCR to identify broken processes across the entire business enterprise - after all the contact centre is the vocie of the company and can find themselves defending other departments and their policies. Surely our energies would be better looking at FCR as a whole and making things chase within the business as a whole?

We gave more coaching resource to our teams and more ongoing training, deskside coaching and general up-skilling. This empowered staff and encouraged sound decision making resulting in more FCR. We wanted to focus on FCR but not to the detriment of customer service. In my opinion the FCR is almost a by product of the other benefits you get.

Jo has asked how agent's FCR performance can be measured reliably and that is a key question. To do this the agents have to believe in the reliability and consistency of the measure they are being asked to support. More specifically agents will want proof that they have caused repeat calls themselves. I know three 2000+ seat contact centres that have used IVR based samples to do this and in each case the method has been pulled within 2 months because the personal FCR measure "flip flops" and the agent with the best FCR score one week can be bottom of the league table the next. This method cannot be used in conjuction with a bonus scheme.

Any sample based approach IVR, agent tagging, CLI, speech analytics, call recording simply does not work!

In November 2008 I sat down having completed FCR training with 130 major UK and Irish enterprises and found that no-one was measuring FCR across all calls (let alone contacts), no-one could provide agents with a reliable personal FCR metric and most importantly of all, the main methods call recording, agent tagging, caller line ID, IVR surveys, post call telephone surveys and speech analytics were all failing and sub-optimal.

Where the organisation just wanted to find the main repeat reasons, then coach all staff on the same problems some of these methods work to a degree. But how does an agent feel when they are being coached to avoid a repeat cause that they do not generate - big demotivator! No, what is needed is the means to coach and reward good FCR behaviours individually, by coaching specific agents on their own repeat call causes - surely that is the goal!

FCR CAN be measured at agent level using algorithmics. Although not widely known in the UK & Ireland algorithmics covers all calls, does not rely on call recording and has even replaced speech analytics in some sites. Algorithmics are totally market proven in North America with large end contact centres with more than 2000 seats such as Chase, Wells Fargo and Sprint Telecom using it for a couple of years now. It is used almost universally by large health insurers and has a strong foothold in utilities. Algorithmics can provide the whole contact centre operation with an accurate and consistent FCR measure. This can also be broken down to the agent level and used to incentivise performance based on a reliable and consistent metric and repeat reasons for each agent.

The issue of real time measurement is an interesting one because some repeats happen within 30 minutes of the first call, but some take perhaps 45 days. Algorithmics can mathmatically compute FCR rates using complex mathmatical hypothesis based on known repeats, empirical repository of repeat algorithms and feeds of information from existing systems within the enterprise.

Algorithmics do take a bit of an effort to understand for sure. But supermarkets use them in their supply chains, the Typhoon Eurofighter uses them to defy gravity and stay in the air, your CRM system already uses them, as does the London Underground, insurance companies for underwriting and Google recruit 1/3 of the USA's PHD's in algorithm! Complex mathmatics have already been packaged to help solve FCR. My UK and North America research clearly shows that algorithmics does the job better than any other method by providing a metric across all calls, providing agent level feedback and directing coaching far more efficiently.

In 2008 many large contact centres thought they could calculate their own algorithmic formula and failed. Many others are still ploughing through the same process doomed to fail too. The maths is complex one of the providers alone employs 50 mathmaticians alone.

At the moment algorithmics are only really viable for 400 + seat contact centre, but those adopting them provide a strong covenant for their use on the basis of significant cost saving and CSAT improvement.

Finally a quick question for readers clinging to the belief that they can achieve this with call recording. If a contact centre claims a FCR rate of 95% and believes it has 100 repeat reason codes (assuming they come in in equal quantities), how many recorded calls would the organisation have to listen to to guarantee finding one example of a specific repeat reason?

The answer is 2000 calls!