Skip navigation
All Places > Industry News and Best Practices > Blog > 2017 > May

Providing key metrics and clear numbers is primordial in any industry, and it becomes particularly challenging in the field of call centers. This is why managers have developed a number of techniques to quantify results and improve efficiency over the years. One of these methods is Call Center Service Levels.


The most basic definition of a service level is: a measurable number of services provided to a customer within a given time period.


In the context of call centers, this is often employed to measure the percentage of incoming calls that agents answer live during a set amount of time.


Calculating call center service levels

Unfortunately, calculating call center service levels is a highly contentious issue. This is because the rate can easily be manipulated depending on the formula employed to calculate it. In the following examples, we will look at 5 different ways to calculate service levels and see how they offer different results.


All the formulas are based on the same data. For this example, we will limit the time threshold to 30 seconds.


During these 30 seconds:

  • ● 1000 calls were answered
  • ● 60 calls were abandoned
  • ● 860 calls were answered within 20 seconds
  • ● 40 calls were abandoned after 20 seconds
  • ● 10 calls were abandoned within 5 seconds




Formula #1

The simplest formula for calculating call center service levels is the following:


number of calls answered within threshold / total calls answered * 100%

In our example, this is ((860)/1000))*100% = 86%


The service level rate of 86%. While this looks good, we should be aware that it does not represent the abandoned calls.

Formula #2

This formula is designed to take all calls into consideration. However, it also uses abandoned calls during the time threshold as a positive,


Total calls answered within threshold +calls abandoned within threshold/ total calls answered + total calls abandoned * 100%


Here, the result is (860+20)/(1000+60)*100% = 83%


Formula #3

This version tends to impact the results negatively, as it treats all abandoned called as a negative.


Total calls answered within threshold / Total calls answered + Total calls abandoned*100%

With our numbers, this gives a result of (860)/(1000+60)) *100%= 81%


Formula #4

The fourth formula ignores calls abandoned before the threshold, while abandoned calls after the threshold impact the result negatively.


Total calls answered within threshold / Total calls answered + Total calls abandoned after threshold*100%


Our data gives us the result of (860)/(1000+40)*100% = 83%


Formula #5

Finally, the last method uses a threshold that accounts for short calls, counting abandoned called before the threshold as a positive.


Total calls answered within threshold +calls abandoned within a shorter amount of time than the threshold / total calls answered + total calls abandoned * 100%


(860+10)/(1000+60)*100%= 82%


Finding the right call center service levels

Since different Center Service Levels offer varying results depending on the data selected, you need to ensure all your parameters are well defined. This can easily be done by following this 10-point checklist:


  1. 1. Decide how to classify abandoned calls: are they counted, missed opportunities or ignored.


  1. 2. Select the appropriate formula: as we’ve seen in our example above, different formulas offer different results.


  1. 3. Choose a time interval: different time thresholds will significantly impact your results.


  1. 4. Decide how often you measure: ideally, you will want continuous monitoring. Since this is not always logistically feasible, you should ensure that you measure at intervals that accurately reflect your day-to-day operations.


  1. Use the right tools: using a dedicated call center software could drastically improve the accuracy of your service levels. By collecting and processing huge amounts of data, the right software will help you make the right adjustments based on your measurements.



Final Thoughts

As we’ve seen, Call Center Service Levels can be confusing, which is why it is important to ensure they are measured through the right method. As is often the case with call center data, this is an area where advancements in software technology, such as speech and customer engagement analytics, can drastically help managers get a clearer picture of their performance.


A final point to note is that if you are working with clients who require Call Center Service Levels, your formulas should be made transparent with everyone in order to be understood. This point is usually defined during a contract between the call center and the client, and as a reasonable solution, most of them will include a clause to accept 10% variance between the results.

In the US market alone, there are hundreds of customer service consultants offering thousands of customer service improvement strategies which begs the question: does anyone need yet another customer service improvement plan? I think, decidedly, yes, for the simple reason that most customer service remains lackluster and inconsistent—while executives routinely believe their customer service is better than it really is. (For more information on this, just ask, we’re happy to share.)


So why does customer service tend to be largely reactive, inefficient, and overly transactional?


From having evaluated tens of thousands of customer service interactions, I find that when customer service disappoints it’s almost always because it has been managed in an overly general, cookie-cutter way. The result is that customers are treated more similarly than they really are, as though they have the same needs, expectations, and perceptions. But of course, that’s not true. Each customer is unique, making their inquiries at least a little bit different. So when companies treat everyone the same, rarely are customers fully engaged or completely satisfied.


Antidote! What I outline here is a plan that actually improves customer service. I know this plan works because we’ve been using it for more than a decade to improve customer service for clients in a wide array of industries. And the reason it works is that the entire plan hinges on a single proven concept—one that’s paid huge dividends for our clients: specificity. That’s specific ways to add value, relative to specific scenarios, measured by specific scoring rules, summed with specific metrics and last but not least, coached with specific model answers to build necessary customer service skills.


If your immediate reaction is, “…but that’s not scalable!”, I assure you it is. There’s a well-crafted process behind this plan, so it’s actually more scalable than the usual approaches to customer service that are less formally conceived.


Step 1: Decide What Specifics You Will Add

First, you need to decide what specific, extra value you can add to each customer service interaction. This “specific extra” becomes a way to involve your associates—and it’s a powerful way to create a lasting, positive impression in customers’ minds. Examples of “specific extras” include brief, meaningful educational content; or a policy that is clearly and frequently articulated like Zappos has with its easy-to-return shoes.


Adding value through “specific extras’ is about consistently doing a little bit more, on top of addressing the question at hand or solving the problem.

Where to start? Gather your customer service improvement team and brainstorm. Then see how each of your good ideas can actually play out in real interactions. Sometimes those great ideas are clumsy when put into execution. So adding a specific extra is both imaginative and iterative, and requires a little bit of trial and error to land on what’s right for your brand and goals.


Step 2: Take a Complete (and Specific!) Inventory

In order to improve your customer service, you need a clear and specific picture of who contacts you and why. Don’t assume customers who ask the same question need the same answer. And don’t assume that your customer service report or software analytics are picking up on unique scenarios, because at present, software is not sophisticated enough to tease out this level of differential nuances.


The solution is to observe a statistically-valid number of your customer service interactions (emails, chats, face-to-face, etc.) and classify them by touchpoint, inquiry type, customer state of mind and customer objective.


Once you’ve figured out the possible combinations of touchpoints and customer characteristics, you’ll have your list of specific customer scenarios.


Step 3: Define Specific Criteria

You can’t manage what you don’t measure. So for each unique customer scenario, develop specific scoring rules. When figuring out what to measure for each customer scenario, start with the four dimensions common to all customer service interactions:

  • Timing: Was the customer’s time valued?
  • Information: Were the customer’s questions answered clearly, accurately, and proactively?
  • Connection: How engaged was the associate? Was the interaction tailored to the customer’s situation?
  • Differentiation: Did the associate demonstrate that your company is special in some way?

To make your scoring rules usable, break the four dimensions down into specific elements (usually there are between 8 and 20 elements) and weight these elements depending on what’s most relevant to the specific scenario.


For example, when a caller asks a retailer where their package is, connection and information will be most important. But when a caller asks about products they have not yet bought, providing persuasive information and differentiating your brand will matter most.


There is no doubt that developing scoring rules that measure each element is extremely time-consuming. But to be accurate (and truly useful), scoring rules must be specific and include explanations about how to apply each rule.


Step 4: Track Specific Metrics

Measure often and keep track of progress using specific metrics based on the elements you’ve defined in your scoring rules. Metrics that are specific show you exactly where and how you need to improve. Metrics that lack specificity (read: net promoter scores and C-SAT scores) don’t give you exacting details about where your customer service is going wrong.


Manage and share these metrics with a dashboard that enables you to coordinate improvement efforts across teams. Dashboards are also a great way to engage associates with the customer service improvement process.



Step 5: Provide Specific Examples

Finally, provide specific examples that show associates exactly what you are looking for in how they handle each specific scenario. If you can’t show your associates model answers, you’re missing a vital tool, because while it’s possible that associates could build out these models, they probably don’t have the time.


And without clear models, while some of your associates may make great choices, the fact is, some could unknowingly tarnish your brand.


Specific examples may sound like you want rote answers to customers’ questions, but you don’t. To prevent that hollow, robotic quality that creeps into customer service, coach associates on the structure behind each model answer, giving them the customer service skills they need to improvise off those structures and develop their own unique responses.

Superior customer service is specific, and specificity is the key to customer service improvement. It’s about specific ways to add value, understanding specific customers and their specific situations, measuring with specific scoring rules, tracking specific metrics, and providing specific examples that give associates the skills they need to deliver the highest levels of customer service. When you follow these five steps and embrace this concept of specificity, you will be well on your way to improving your customer service.


Perhaps the best way to think about this ‘specificity concept’ is as a mindset that actively focuses on awareness of variation and difference. This is a decidedly different way of thinking and perceiving that social scientist Dr. Ellen Langer describes as mindfulness, in contrast with the usual mindless ways most of us tend to our experiences. Let me know how it goes!



Managing a call center is no easy feat, which is why it is imperative to deploy tools that can measure progress, results and performance. The first step to charting your progress is defining your objectives. The next step is deciding what key performance indicators, or KPIs, you will use to verify how effectively your call center is achieving its objectives?


While you can find excellent resource listing such as the best KPI-measuring tools, and in-depth overviews of the best call centers KPIs, this blog will focus on how to choose meaningful metrics for your company.


Start with A Clear Strategy in Place


The best way to use the right KPIs is to start with the objectives you want to achieve. Measuring data is only valuable if you are aiming towards a clear goal, which is why your strategy map is a primordial first step before identifying your indicators.


Moreover, these objectives need to align with the larger goals and targets of the general operation. Since most corporate objectives will eventually be financial, it is important that your call center goals focus on the KPIs that support these goals, such as enhancing customer experience and satisfaction, improving agent performance and streamlining processes.

Defining KPIs That Make Sense for Your Organization


Since there is no one-size-fits-all for KPIs, they need to be focused on specific outcomes. For instance, if your only worry is to increase customer satisfaction, it makes little sense to spend time, money and resources on measuring revenue per call, a KPI that would be essential for telesales or performance marketing contact centers.


Similarly, if you are working in debt collection, a metric like “Promise to Pay” will be a much more important KPI than Average Handle Time.


Ensuring KPIs Are Shared and Understood by All Employees


This is an important point that is often overlooked: if you are trying to find meaningful KPIs for your call center, they should be easy to understand by your employees as well. KPIs should remain an integral part of their decision-making when interaction with customers, especially when you are trying to increase performance.


Keeping your workforce in the dark about which data you collect could make them suspicious of your actions and cause them to disengage. Likewise, giving them too many KPIs to focus on could confuse them, which is why it is important to ensure they are assimilated and understood seamlessly by everyone. You might also consider a phased approach, introducing one set of KPIs at a time, giving agents the opportunity to show success before introducing the next set.  

Some Examples of Call Center KPIs


One good exercise to see if you can derive meaning from your KPIs is to have a look at the compiled list of the most important call center metrics according to managers:


  1. Quality Scores: providing an overall score for the caller experience. Quality scores can be monitored at the agent and group levels. When you track 100% of customer interactions, supervisors and agents can get an accurate and unbiased view on how well they are doing.


  1. First Call Resolution: measures the percent of conversations that get resolved on the first contact. Simply divide the number of calls that are resolved on the first contact by the total number of first contacts. You can get even more sophisticated by segmenting data by call reason, call type, etc. For example, you might want to segment calls for tech support, billing issues, account information updates, marketing, or legal. Ultimately, the objectives for measuring FCR are to enhance the customer experience and satisfaction, reduce costs and improve agent performance.


  1. Customer Satisfaction: looks at the agent behaviors and processes that most affect the customer experience. Beyond looking for words and phrases that indicate dissatisfaction, speech analytics also looks at sentiment by monitoring, silence, and loudness and micro-tremors in the voice.


  1. Average Handling Time (AHT): the total amount of time it takes to handle a call, measuring different areas such as talk time, on-hold time and wrap-up time. Keep in mind that the optimal handle time depends on the complexity of the customer’s issue. To get an accurate assessment of agent performance it is important to average handle times over several calls.


  1. Transfer Rates: expressed as a percentage of the number of calls that are transferred to someone else to resolve. Transfers occur for a variety of reasons including the request of a caller, the fault of the agent or incorrect routing from the IVR. Insights from speech analytics can identify the root cause so that issues can be quickly addressed, thereby reducing transfer rates.


  1. Net Promoter Score (NPS): a metric used to predict customer loyalty by looking at how many of them would recommend your service to others. NPS can indicate that there is a problem. But NPS alone can’t tell you what is causing the problem. Is it an agent, a process or a product? By listening to what all your customers are saying when they interact with your brand you can gain insights that can help you improve NPS. In addition to monitoring what is said, you can also monitor customer sentiment. 


These are just a few of the main KPIs you might consider for your organization, and there are of course many more. As previously stated, finding the right ones for you depends on your goals and strategy.




What’s Next?


Once you have established what you are going to measure, the next step is determining the tools you will need to effectively measure the KPIs you have set. Speech Analytics can help you monitor all interactions across channels – phone, email, chat and social, and gain a 360-degree view of customer likes and dislikes, preferences and concerns. It is one of the most effective tools for monitoring call center and agent performance and helping you to deliver a great customer experience at every touchpoint throughout the entire journey.

Final Thoughts


As the list above shows, some KPIs may overlap while others might not be essential to your organization. It is important to remember that, like with all key decisions, planning and adjusting your KPIs is just as valuable as deploying them effectively.


Finally, don’t forget that KPIs are only as useful as the meaning you can derive from them, whether it is in terms of how easily you can share them with your staff, or how much they can help you adjust course to improve performance.


What KPIs do you measure and which ones do you find to be most effective in achieving your contact center and business objectives?