© 2022 Black Swan Telecom Journal | • | protecting and growing a robust communications business | • a service of |
If we were still in the era of simple landline phones and voice mail, maybe telecom customer care jobs would be in decline.
But in the age of smartphones and a dizzying array of app and network technology, customers expect call center pros to step up and help them solve tough problems.
In fact, the technical expertise of many customer care pros today puts them in the league of “Computer Support Specialists”, a $48,500 a year job the U.S. Department of Labor says will grow an above-average 17% in the next decade.
But technocrats undervalue the role of humans in the customer experience mix. The truth is that full, end-to-end care automation is neither affordable nor wise. The best strategy is a semi-automated one that puts great intelligence at the fingertips of highly trained reps.
And stepping up to deliver that greater — and big-data-based — care intelligence is comScore, a Washington DC based firm that has led the website audience measurement business for two decades.
In this interview, Brian Jurutka, comScore’s Senior VP of Global Telecom Solutions, explains why his firm made a deep investment in telecom-specific analytics and the payoff the firm’s now delivering in customer care.
Dan Baker: Brian, it would be great if you could first give us some background on comScore, whose background is rather uncommon among telecom software firms. |
Brian Jurutka: Dan, comScore is best known for our audience measurement and analytics of websites on the internet. So when a Microsoft or Yahoo says: “Hey, you should advertise on our site because we have 21 million visitors a month,” comScore is often cited as the independent third party authority who verifies those numbers.
Now helping us do that is a panel of 2 million people worldwide (including 1 million in the U.S.) who let us monitor their on-line usage.Then the content providers also tag their websites so we can verify and weigh our panel’s behavior to the full internet population.
We have also developed our own dictionaries and business rules to categorize content across the web, as well as for mobile applications and devices. For instance, when you visit a website, we capture the start and end time of a visit and the number of page views.
One of the biggest challenges is converting the raw web events into intelligence. If you have 350 server calls going back and forth between machines, how do you translate that to actual page views? It’s not easy, but we’ve built our expertise and methodologies in that for over 15 years.
Now in recent years you’ve invested deeper into the telecom industry specifically. What attracted you to do that? |
Yes, about 4 years ago we bought Nexius, a firm who specializes in measuring service quality in a carrier network. We recognized we could add value by combining the Quality of Service (QoS) metrics that Nexius provides with comScore’s own categorization data.
What distinguishes our approach is we’re agnostic as to the source of the network data. Other companies can profile subscribers by linking to proprietary probes. But we’ve said: why not work with any probe out there? By being DPI agnostic, we can capture voice, data and SMS quality.
So after calculating QoS, we combine that with comScore’s audience measurement and content categorization methodologies to capture the metadata. And this is very specific: it’s not just saying the user visited the Los Angeles Times, it’s knowing that she visited the technology, sports, or travel section of the site.
And our people are good at normalizing this data so when you look at the duration, page views, and megabytes consumed, it’s apples to apples across a multi-vendor network that includes equipment from Ericsson, Nokia-Siemens, and all the rest.
Believe it or not, this is not easy to do: connecting the CDR data with DPI data and CRM, but it allows you to make sense of all this data from a business perspective. And that’s huge.
Where are you apply your technology to use cases? |
One of the most valuable uses of our data is in customer care, such as telling whether or not a wireless phone is defective. We can look at the voice performance of a handset relative to other handsets of the same model in the same network region. And that enables us to pinpoint a device-specific issue from a network-specific one.
Now there are companies out there who capture device data by having an app resident on the device itself. That’s fine, except that the app is likely to be embedded in, say, only 7% of the phones, which means the care reps have trouble using a solution like that because of the swivel chair issue: having to know and train on too many screens.
With a multi-probe vendor approach like ours, however, you can compare the performance of over 90% of the phone models in a region. And covering the lion’s share of devices like that makes it cost effective to build standard processes and train agents on.
Can you give us an example or two of customer care cases where this is useful? |
Well, one carrier has over 38,000 customer care reps that access the tool on a daily basis.
Now imagine if a customer calls in and the care rep has no way of knowing whether or not there’s a problem with a particular handset. The rep might recommend a diagnostic routine with the customer on the line: “Click this, click that, and tell me what the screen says.” Or maybe you need to download some software to their device and then do 2 minutes worth of diagnostics.
So when you don‘t understand the health of a handset, it leads to some costly call handling time.
However, using our system, the operator runs a “defective handset analysis” on a regular basis so when the customer calls in, the care rep knows immediately whether the handset is the likely issue. And this can save a few minutes in average handling time which translates to hundreds of thousands of dollars in decreased care costs. Now, for certain specific problems, our customers claim they’ve sometimes reduced care handling by up to 35%.
Another care case is around customer disputes over mobile data overages. Say the customer is on a 5 Gigabyte plan and he used 6 Gigabytes of data that month. Well, many customers like that call to complain that their usage wasn‘t calculated correctly.
Now if the rep has no idea what’s driving the user’s consumption, they have no way of driving the conversation towards root cause analysis. However, with our solution in place, you have the amount of data consumed by application and protocol. So the care rep can then say, “May I have permission to look at the usage details on your account so I can assist you with this problem?”
And with the customer’s permission, the rep can reply: “Well, it looks like 80% of your data consumption came from the iHeart radio application. Do you use iHeart radio?” And the customer says, “Of course, I do. I love it and use it all the time.”
And it’s at that point where the conversation shifts because the customer has acknowledged using iHeart radio. And this means that instead of having a frustrated customer or offering the customer a credit on his next bill, the customer figures: “If I want to continue to use iHeart radio, I should probably sign up for a bigger plan -- or make sure I’m on WiFi when I use it.”
The payback from average handling time is clear, and I also think the operator gains points by having the intelligence at its disposal to resolve the issue right away. If you need to go back and forth — or contact the customer later on — the customer loses confidence in the carrier’s ability to execute. |
Absolutely. In fact, the challenge of delivering both these capabilities is having data available at a point where it can make a difference.
The data exists in the network. It’s all there, but how do you resurface that information to a care rep in a timely way and ensure the rep understands it and can act on it? That’s what we do -- take all the data and condense it down into metrics that help reduce average handle time and increase first call resolution.
Brian, thanks for the interesting use cases. Can you sum up by telling us what makes your approach most attractive? |
Well, certainly our ability to categorize the applications and websites the customer visits is one key advantage. Second, I think, is the ability to work with any probe in the network. And the third is that our system was designed from a business side looking back to get the data as opposed to having the data and trying to find a problem to solve.
Our carriers appreciate the fact that we didn‘t start off as a probe company. We started out with a business focus: what data do I need to answer certain strategic questions?
Maybe that sounds like a nuance, but it actually makes a big difference because when you approach the problem from a data-source-neutral and business-specific angle, you often end up with a more robust solution that’s more actionable in terms of moving the business forward.
Copyright 2014 Black Swan Telecom Journal