If you Google "call center data analytics," over 11 million results will promptly appear, many of which are promoting data analytics products for sale (naturally).
And it's a given that we need data analytics in the call center in order to improve customer service.
But what exactly is your current data telling you that you don't already know? Since the early days of Average Handle Time, call center operations have struggled with what we can measure versus what we should measure to improve the customer experience. So let's take a few minutes to review what, at a minimum, your data analytics application should be providing you.
1. Interactive Voice Response analytics
If your organization has an IVR, you should have at your fingertips a wealth of data that can help to streamline and improve the customer experience. Many organizations create an IVR and then just tack on new prompts as needed without any analysis as to how changes affect the quality of customer experience. But why do that when there is so much data available to help improve the IVR every day?
You'll want access to the following data:
- Reason for Call analysis: What reason for call did customers give and in what proportions? What percentage of each Reason for Call bucket was properly routed? Does the Reason for Call logic and routing need to change based on actual caller use?
- Caller Identification analysis: What percentage of callers were accurately identified? For the ones that weren't, why weren't they? What prompt should be changed or added to increase the identification match rate?
- Opt Out Rate: What percentage of callers opted out of the IVR call flow (by hitting zero to speak to an agent)? At which prompt did the largest percentage of callers do so? Can the prompt or call flow be rewritten to decrease opt-outs?
- Misroutes analysis: What percentage of callers ended up in the wrong queue and had to be transferred? What Reason for Call did those callers give? Can the Reason for Call prompt and routing logic need to be altered to decrease misroutes?
- Self-Services rates: What percentage of calls were handled exclusively within the IVR, without speaking to an agent? How were the quality of those calls rated in comparison to those handled by agents? Are there additional types of calls that would be good candidates for self-service?
2. Speech analytics
"Speech analytics" is becoming the hot buzzword of the call center (and stay tuned for an exciting announcement from Spoken in this area). Whether performed on live calls or after the fact for quality and marketing analysis, speech analytics can provide valuable insights into the customer experience. So what exactly can this next generation of speech analytics provide to improve the customer experience?
In the past, speech analytics was limited to keyword search applied to recorded calls after the fact. However, in this exciting next generation of speech analytics, organizations will be able to analyze using the following tools:
- Real time keyword search Apply keyword searches to real-time transcriptions of live calls
- Real time sentiment analysis Apply sentiment analysis to real-time transcriptions of live calls
- Privacy redaction Redact private customer data such as credit card numbers from recorded calls
- Real time issue flagging Use machine learning to create algorithms that can flag live calls that are at risk
3. Agent evaluations
The workhorse of the call center has always been agent evaluations. Never quite a sexy as newer data analytics, agent evaluations often fall into the category of "should" and get dropped down in the priority list, behind newer and shinier quality applications. And in part, I believe that is because many of us believe that these evaluations are NOT going to tell us anything we don't already know.
However, that is exactly why it's essential to insist upon third-party evaluations of agent performance. Why third-party? Because most of us have our own personal biases, and it's nearly impossible not to let them affect agent evaluations. For an accurate view, selecting a third-party vendor to conduct agent evaluations is a best practice. And that vendor should be able to provide:
- Form evaluation An evaluation of the quality form to ensure what is measured will accurately reflect agent success and improve the customer experience
- Attribute analysis Evaluation of each quality attribute, viewable by agent, score or coaching notes
- Calibration Several vendors offer online calibration, which saves a significant amount of supervisor time and energy
- Customized reporting Any evaluation service should provide the ability to customize and schedule reports and alerts
4. Real time and historical reporting
The beating heart of call center analytics is the dashboard showing both real-time and historical performance data. Can your reporting tell you which agent or line of business is in trouble right at this moment? What about which team has improved the most over the last month?
Daily reports and roll ups are a thing of the past; today is all about viewing performance data in context, both real-time and historical. Many vendors offer online reporting and dashboards that curate a variety of data feeds, including our own Spoken Engage advanced reporting:
- Automated data from the Automated Call Distributor (ACD), including agent ID, call length, and Average Handle Time (AHT)
- Silence An analysis of silence time on calls, usually a predictor of low quality
- Talkover An analysis of talkover time in which both parties were speaking at once, usually a predictor of low quality
And many tools offer the option to integrate with other third-party quality metrics, such as:
- CSat scores Customized customer satisfaction scores from past post-call surveys
- First Call Resolution scores Historical FCR scores
Analytics are abundant in today's modern call center. The key is to be able to glean insights from them to determine what you don't already know. When you select your analytics products, feel free to be a skeptic and ask, "What is this going to tell me that I don't already know?"