BI is taking a central spot in the heart of most progressive companies these days.
BI is a powerful tool (and a hot buzzword). It is, in some ways, the perfect companion for customer service, where you have a good chunk of data to process. Still, it is also a double-edged sword because, unlike Sales data, for example, in many service teams, the variation in a specific data set can be quite significant…
for example, let’s look at this example chart of Time to First Response per agent for a Tier/Level 1 Support team:
We can see that Steve is the fastest responder in the team, Matt is somewhere in the middle, and Jenny takes the longest to respond. So, by this chart, if the manager who reads it looks only on this parameter isolated, Jenny should be on a Performance Plan, Steve should get employee of the month, and Matt is mediocre. Only, there are other factors that are important to managers except for the speed of response or the number of tickets handled. Good customer service is also about efficiency and quality. While speed is a critical factor for customer satisfaction, there are other very significant factors. These other factors are where we start to build great BI dashboards and reports, and they are our Key Performance Indicators, also known as KPIs.
Start by defining clear individual and team KPI’s.
We looked at Time to First Response, it could be a KPI on its own, but it could be too generic to indicate performance without combining it with other factors.
The most important function of BI is the ability to take different data sets from other systems and combine them into one clear picture that provides insights you might not have gotten to just by looking at an isolated data set.
The basics of reporting and BI reporting are the same. You need to understand what is important to you. In our example, let’s say I look at the following KPI’s as indicative of the performance of my Tier/Level 1 team members:
- Time to First Response (number)
- One Touch Resolution (Percent)
- Number of touches – agent side (number)
- Time to Resolution (number)
- Escalated to Tier/Level 2 (percentage)
- Customer Satisfaction Score/NPS Score (number/percent)
Note that I didn’t define the desired values for the KPI’s just yet. Before I do so, I will go through the below stages.
Make sure everyone you benchmark should be benchmarked together
Comparing KPIs between team members will give you very insightful data, but this is where it gets tricky. Let’s say you have a team with different types of people (like every team) with varying levels of experience and tenure, they should be doing the same function you hired them for, BUT they don’t really.
For example, if we continue from the example above: Jenny is more tech-oriented, and she’s also the most experienced of the bunch. She sees a complex ticket in the queue, and she jumps on it because she likes to investigate the more difficult tickets.
On the other hand, Matt has been on the team for three years now. He might respond a little slower than Jenny, but he will not nitpick tickets. His one-touch resolution is high, the number of touches is low and quantity-wise, the number of tickets he can handle is relatively high.
Steve is new to the team, fresh from training, he’s enthusiastic and takes tickets the fastest, but his responses still need work and lack experience, so his number of touches is significantly higher.
So how do you come up with the right benchmark? Divide and conquer.
Micro-group your data
To get good benchmarks, you need first to make sure you’re splitting the team right. My philosophy is that if you have a group of people with different skills and abilities but they are not doing the same thing, find the middle ground. If there is not a sufficient amount of middle ground, you need to make some changes to your structure because it does not reflect the role and the people who are doing it.
If you have too many variations in performance and a small sample pool (a small team), check out your medians. They might be more indicative for you in the beginning to indicate performance.
If you have a lot of people, then it’s time to split them into groups. Before you do so, here are some questions and data you might want to look at to help you define them:
- How long does it take your employees to ramp-up?
- How does the improvement graph look for new employees over time?
- How many tickets are escalated to the higher support level/tier by each in these stages, do you see a pattern?
- What level of experience each of them came with into the position?
- Were some of them promoted from a different team? Does their data look different than the other team members?
- What’s the correlation between TTFR (Time to First Response) to OTR (One Touch Resolution)?
- Etc. (you get the direction)
Once you’ve examined and researched your data, you probably reached some insights and found some things in common. Now it’s time to make additional grouping to your BI system (or Excel sheet, whatever rocks your boat). To see if they are indicative over time as benchmarks, picking the correct time frame to measure is essential to ensure you’ve got good performance indicators that give you visibility and improve the team’s performance.
Choose a time frame which will show you true actionable data
KPI’s and BI can be excellent tools for Micro-managers. Let’s get that out of the way. I’m assuming that if you’re reading this, you are not one of these managers who would look at the daily results of each of their agents and ask why they had a long response time yesterday. This is not an article on micromanagement and its effect on employees. Please don’t make your people hate data by utilizing it as a tool to micromanage them.
To ensure strong and indicative KPIs, we need to take a representative time frame. I’d recommend two weeks to a month. Once we have enough data looking at the last six months’ performance split by month would be a great way to learn the current state of an employee or team, but the minimal time sample we define is the most important.
