At Teleopti’s most recent Americas User Forum, Carlos Munoz, Pre-Sales & WFM Consultant at Teleopti, chaired a panel discussion with customers around measuring and managing KPIS in their customer service operations. In this blog Carlos pulls together some of the key topics and answers that came up.
In a panel discussion I moderated not long ago on Key Performance Indicators (KPIs), we had three panelists participate, with the audience contributing their questions and experiences. What was surprising was some of the differences shared among the three panelists when it came to which KPIs were being measured, how, why and their importance.
Just to give a bit of background on the three panelists and their businesses:
- A senior director for project management (PM) whose company builds websites for small businesses. He directs a WFM and analytics team for operations, now with 13 centers and over 1000 agents in many places, as well as 15 brands under their belt.
- Another panelist helps with WFM for a contact center (CC) representing 200 auto mechanics, the agents helping customers negotiate mechanical issues.
- A WFM specialist working for a credit union, whose members are in control. Their CC has grown from 5 to 80 agents over the past 15 years.
Introductions – Core Measurement Focus
The PM Director of the website business started out by saying their main KPI is customer satisfaction – from which everything is driven. They try to keep it simple. Their most important metrics are for their front line, including average speed of answer (ASA). If a call is answered quickly, it reaffirms the fact that they have a good product and process in place, which hopefully delight customers!
Standard agent metrics are measured – e.g. schedule adherence – but not average handle time (AHT). The latter is used as a coaching tool, but not as a measurement of agent performance since agents are encouraged to work with their clients as much and as long as possible.
The credit union WFM specialist reported that in her business, the focus was on service level, adherence and ACW (After Call Work). What drives their KPIs the most is member satisfaction. That said, however great it is to have personal interaction with members, it should not be at the expense of other callers having to wait. They don’t focus on AHT either, but she wished they did so agents could learn to have more call control because sometimes they have delays of up to 20-30 minutes. Some agents have an AHT of ten minutes, which horrifies her: If this is the AVERAGE handle time, what then is the reality? This should not be the case and she presented the idea of supervisors using AHT as a coaching tool to her boss as a tool e.g. to identify if something is missing, if agents are too chatty, etc. She considers it a helpful tool to look at and gauge who needs help or identify if agents are, for example, spending too much time looking for information.
As for the auto mechanics contact center (CC), similarly, handle time isn’t used but instead schedule adherence. Customer savings are looked at, as well as the negotiations that take place with those dealerships. Another key KPI measured is ASA (Average Speed of Answer).
Do you employ any specific methods or tools to both gather data and manage bad habits?
The manager for the auto-mechanics CC confirmed there was a balance between the two but confessed they sometimes get in trouble because although some of the agents showed results, they also showed bad habits. On paper, they may look good but their behavior is definitely not something they were coaching or trying to teach. It forces those supervising the coaching to be more dynamic so that the right things can be done. If customers are treated the right way but the handle time goes up, then agents are not doing what is expected of them. Company culture or not, supervisors and managers have to deal with it.
Tying in with this the manager and one of her groups felt the metrics being employed weren’t working for them. They wanted to get rid of them and instead, have something simple and easy in their place. Teleopti was the best way to go with that. Tiers, with gold, silver and bronze prizes were set up, with agents able to see what they had attained. This was implemented last week and is still being tweaked. “We would like to get more of our CC on board with this, but it takes time getting people bought in and to simplify,” said the manager for the auto mechanics CC.
The credit union WFM specialist related that many things were looked at regarding targets in order to determine if they were met or not, and that this was also based on research on other credit unions. Most was done through consulting – not through looking at industry or white papers, engagement or benchmarking. She expressed there was a need to go back and look at those again to make sure it fit their company and culture.
One thing they do use is gamification, which the credit union agents love! Agents check their stars, if they are gold, bronze, etc. Adherence goals are not set but are watched to assess the highs and lows to find a happy medium to strive for.
Reporting tools are used to determine where your business has been and shape some of the benchmarks. While the hard metrics are more commonly used, how are the softer KPIs – customer experience and engagement – handled and what targets are defined for these?
Metrics for customer satisfaction were never used before by the web site organization, nor were other companies looked at for best practices. Eighty percent was arbitarily chosen. Agents were pushed and paid for their best performance, and not told it would become an incentive metric. “That’s how benchmarks are set many times. Once an official benchmark is set, we may realize something else is more reasonable, so we work it into agent incentives and actual performance metrics, rolling it all the way up to the supervisor and managerial levels,” said the PM Senior Director.
The web company also has a huge data warehouse. Trends are analyzed and the numbers crunched: If, for example, call Y answered by Bob or a particular team is identified as giving a much better experience, calls are directed there for a while to figure out the reason and how it can be used in training. The frontline and supervisors also come in with suggestions. This feeds directly into empowerment. Sometimes, however, it’s a demographic, which can’t be trained. Still the data gives clues and often leads to an easy solution.
An audience member piped up, saying much is said about coaching and using metrics, but not about agents’ behaviors associated with those metrics. “How well do we go in and understand what those metrics are telling us? It should lead us to identify the agent’s behavior on calls that are leading to those metrics. Counterintuitively, I feel handle time is an important metric to focus on and look to the behaviors that drive it.”
He continued by saying: “For example, one rep had a good track record but calls took too long. I partner with reps to listen to calls to identify behaviors and then drive behavior performance. In line with empowerment, we use a grow-coaching model to get feedback from the reps about how they can change within their own behavior-set to impact the metric. That’s how we loop metrics into performance management and behavioral change.”
The WFM credit union specialist shared a similar experience; that besides call quality and adherence, a team is utilized to go side by side to listen and dissect calls with reps.
Measuring deviation and discussing where a center went off the rails: Do you do post mortems or is it just pass or fail?
For the auto-mechanics group, it is never pass-fail because: “That would imply if you can’t hit the hard numbers, you’re in trouble. In reality, it’s all shades of gray.” “Performance analysis is done in a few key areas to figure out how exactly agents are doing, see where they need to improve and help them dial down into what they need to apply to fix it. Feedback is given in coaching at least once a week, with a formal recap at the end of the month.”
By contrast, at the credit union, it’s strictly pass-fail: either you meet it or you don’t. It affects the morale negatively, all the way down to the agents. However, deviation is being considered in order to avoid the risk of managing to a number instead of a behavior.
For the web creation company, it depends on the level that the data is displayed to: agents or supervisors. Data is shown in real time as much as possible, depending on the metric, but it is also shown every day for the previous day. A post mortem is held every day for the workforce where every key metric is reviewed. Given the forecasting is done correctly, site leaders are expected to explain why a certain metric is missed. It took a while to get everyone in that tempo but everyone is ultimately held accountable for their own metrics.
Are there specific drivers or data that actually transform the business?
The credit union WFM specialist said that having looked at the math and when members were frustrated, she reached out to the CEO, they used that data to implement a Voice ID, which decreases the time the system needs to identify members when calling in by voice and answering a few questions. The next time they call in, it automatically picks up the information and they’re ready to go to the agent. They also changed the way agents initially enrolled them to decrease that time, by using their driver’s license number. With both these things, handle time was decreased, positively affecting their business.
The website company manager said he tells his team, supervisors and managers that they shouldn’t be afraid to blow things up if they’re no longer effective. They had a long-standing business structure in the operations organization, which they completely annihilated because the data showed that doing so would give better results.
The data shows things like: “We’ll eliminate these lower-complexity calls or eliminate these people on the phone so we can shift them somewhere else,” and not: “If we do this, we can eliminate 10 people,” he said. “We make sure we let the data lead us to solutions. We sometimes drive people nuts because we move at such a rapid speed but as long as the data supports what we’re trying to do, we’ll take any march forward that we can.”
The mechanics CC manager, in turn, shared how they upgraded their phone switch a few years ago, allowing them to implement e-mail contact. The inbound CC had an e-mail team running a pilot to see how they could get to where they wanted to go. The analytics, performance and efficiencies clearly showed what they would gain. Ultimately, they were able to treat every e-mail and call exactly the same and made it a true, blended CC.
What do you do that helps reduce the level of effort in report creation so you can focus more on analysis?
The website PM director provides his teams with the data and tools that are crucial to get what they need quickly, so as not to spend a lot of time finding these figures, crunching data together or figuring out the code behind the scenes. In this way, he makes it easier for them to focus on analysis.
The WFM expert for the credit union said she pushes leaders, senior management and agents to use Teleopti reports; either encouraging them to go and look at it themselves or she trains them to pull information. She pushes it as a positive: that they don’t have to wait for her to have the time, that they can go look at it anytime as it’s there for them and is quite informative.
The mechanics CC manager is grateful Teleopti syncs up well with Shoretel and their 3rd party report application, called bio-metrics. “I have plenty of ways of getting what I want and provide answers to people who ask. It also helps me educate them.”
Other ideas of how to measure and manage better, faster, cleaner?
A member of the audience discussed talk time: “The implication is if you’re measuring, you’re rushing. But talk time also exists on the short side and it can reveal that an agent is not helpful. Very short talk times means they’re hanging up on people or transferring them straight to another department. So, you have to look at talk time in a call center, not just to hold the cost down but to look for problem people.”
What about abandon rates? What is an abandon in your world?
The mechanics CC manager answered: “We consider a short call abandoned after 5 seconds. I can’t justify why 5 seconds. It’s just there.”
The credit union WFM specialist responded: “After 20 seconds we consider it an abandoned call. We try to stay under 5% abandon.”
The website PM director said “For us, it’s 30 seconds just because of how customers come in. We do get a lot of wrong numbers sometimes that bypass our phone menu – the IVR. We’ve just done a lot of historicals that reveal 30 seconds seem to be right, but we have ways to measure from moment of when they got on the phone with us to not.”
Conclusion from me, as Moderator
All the items analyzed and how to go about it show that: it’s clearly not one size fits all. For instance, schedule adherence proved to be an often examined KPI, yet Average Handle Time was a point of contemplation and debate in the panel. Overall, decisions have to be made based on that contact center’s situation, customer needs and agent behaviors. I appreciated all the good ideas that were shared in the panel and thank the three panelists, and audience, for their interaction and honesty.