Using the Right Analytics to Improve Contact Center Performance

The contact center's role as an enabler of customer service has always been driven by metrics. But are we now seeing fundamental changes in how performance is measured?

We still ask ourselves how long our customers have to wait for a response, or for an answer to their problem. We still ask ourselves how productive are our agents, and whether we're getting enough value for the cost.

So, are we OK sticking to the same reports and data we always have?

In a word, no.

The what and the how of performance measurements are evolving in lockstep with how your customers now expect to leverage easy communications for rapid solutions. Before getting to the kind of analytics you should be thinking about, consider the recent past and why these legacy metrics won't keep you competitive today.

The World You Know

Conventional call center reporting has historically focused on operational metrics, such as number of calls, abandoned calls, talk time, service level (percentage of calls answered within an acceptable time period), and so on. Legacy facilities have been capital-intensive, and certain core metrics were needed to demonstrate good ROI to management. These data points were fairly easy to measure. We know how long a customer is listening to hold music; we know when an agent picks up and hangs up the phone. In any given week, we could tell our managers that volumes are up, but talk time is down, so our service level is getting better and better. Life was good, or so it seemed.

Then, as the Internet grew, call centers evolved into contact centers, and multichannel became the norm. With a choice of communication channels, talk time became interacting time; the time it took to compose an email reply or complete a text chat.

Acceptable service levels were relative to the channel: It's OK for a customer to wait 24 hours for an email reply, but Web chats demand response in 30 seconds or less. These same concepts were applied to what was reasonable for each channel.

The World You Thought You Knew

Before we make the jump from multichannel to omnichannel, let's take a step back. On its surface, talk time (or interacting time) seems like a valuable metric; after all, if I minimize the amount of time each agent spends on each interaction, they can handle more. Right?

More interactions, lower cost translates into more customers, higher revenue. True?

But, consider what if agents, who are incented to get off the phone quickly don't actually solve the problem? What if customers have to call back multiple times to address the same issue?

What if agents artificially inflate their number of interactions to appease management, at the expense of quality and customer satisfaction? Talk time may look great, but issue resolution timeand customer satisfaction level could suffer dramatically.

To address this concern, more sophisticated contact centers began to look at combinations of multiple metrics. Let's examine the impact of higher talk times on shorter issue resolution time, you might say. Or let's take a look at the number of interactions per issue. As such, that leads us to metrics that more closely indicate true customer satisfaction, like first-contact-resolution rate, or our customer survey rating or Net Promoter Score (NPS).

Some organizations are examining the value of speech analytics to monitor and analyze spoken words that bring structure to interactions and reveal the information buried in the customer contact center. In a recent report, MarketsandMarkets detailed its potential: "[T]he technology is used to extract important business intelligence material, which is used to relate it to strategy, product, process, operational issues, and contact center agent performance. All this information helps to give an insight regarding what customers or clients actually think about their company, which in turn will help the enterprise to react quickly in favor of their working structure."

Collectively, we are getting closer to identifying real interaction value, but it's not an easy road.

How does one measure first-contact resolution for example? A post-interaction survey ("Did we resolve your issue?") might work, but it could also be biased. Unhappy customers, for example, might complete the survey to make a point.

Agent wrap-up data is also viable, but its value is derived from the agent's honesty. And one must ask whether the agent really knows if the issue was fully resolved. Supervisor evaluations of recorded interactions are an option, but could be very time-consuming and expensive.

The World You Need to Know

As contact centers boldly move into the world of omnichannel customer communications, truly insightful metrics might require even more intricate calculation. For example, if you text message an appointment reminder to a customer, he clicks through to your Web site to change it, and then he requires a live call to discuss options, are these three different interactions with your business, or just one (which combines three different channels)? Would you call this first-contact resolution?

It is no doubt a complex equation. Here are four suggestions to help define valuable metrics in this brave new world:

  1. Leverage a knowledgeable partner. Choose a vendor and/or integrator with proven experience in omnichannel communications and analytics solutions. A good partner will assist in determining the right metrics, with a focus on customer experience, value, and cost.
  2. Determine the analytics that are most appropriate for your business. Do you require segmentation by customer value, by product, by region, and/or by team? Do your key metrics vary seasonally, after promotions, by time of month or year, etc.? What are your key determinants of cost, revenue, and growth? A good analytics system allows you to control how the data is measured, including dimensional views and trend analysis.
  3. Tie contact center data to customer (CRM) data. The problem of connecting related interactions together is simplified when communications are associated with customer and issue/order data. This enables a consolidated view of customer and issue history in reporting and might even empower agents to see a complete picture of a customer journey in real-time.
  4. Use innovative analytics methods for immediate benefit. Tools are now available to interpret data that previously required extensive manual intervention to analyze. Speech analytics can automatically examine real conversations (during or after a call), to provide information on script adherence by an agent, to report on volume/clarity/tone, and even to determine stress levels for agents and customers. Knowledge management can use search data and history to determine what information is most requested by customers and agents. Survey tools can perform post-interaction or post-issue data gathering to determine issue completion and satisfaction levels, then tie results to immediate action (such as a disgruntled customer campaign).

How you measure well-defined analytics depends on your business, but one age-old principle survives. The best results come from a never-ending loop: measure, improve, measure again, improve some more.

Even as customer communications evolve, the value of great metrics will always be based on how you use them to continuously move the needle and advance progress.

John Cray is vice president of product management at Enghouse Interactive.

Related Articles

Effective self-service functionality can make a company's customer experience stand out, and solid search capabilities play a key role in ensuring that customers quickly find the resources they need. With the introduction of Coveo's new Sitecore native search app, Coveo for Sitecore Pro Cloud Edition (as well as an on-premises version of the same tool), companies can add context to customers' searches and improve the relevance of the content that is returned in a search.

Posted April 29, 2016