News Stay informed about the latest enterprise technology news and product updates.

How to make your voice-of-the-customer program matter in 14 steps

At the Text Analytics Summit, Chris Jones of Zynga presented 14 steps on how to ensure a voice-of-the-customer program delivers.

Getting executives to rely on data when making decisions can prove to be a stumbling block for some organizations, but according to Chris Jones, senior manager of text analytics for Zynga, leaning on too much data can also be problematic, especially when it comes to understanding the voice of the customer.

Businesses may assume a customer profile compiled with data -- both structured and unstructured -- provides enough information to serve the customer well. The assumption, Jones said to those gathered at the seventh annual Text Analytics Summit held in Boston last month, can lead organizations in the opposite direction and create distance between a business and its customers.

“Talk to your customers,” said Jones, a relative newcomer to Zynga, “which seems [obvious], but I see this all the time.”

This nugget of advice was one of 14 steps he presented on how to take a voice-of-the-customer program from pilot to production.

Text analytics sounds great, but how do I take this to my company? How can I get people excited about it?” he asked at the beginning of his presentation. “In the end, it comes down to business value. Other than that, your bosses don’t care.”

Step 1 and 14: Feedback is a gift
Jones started his list with more philosophy than practicality: Feedback is a gift.

“I really want to hammer this one home,” he said.

In fact, he believes in this statement so passionately, it made the list twice, framing the talk as Step 1 and Step 14.

Themos Kalafatis, a predictive analytics consultant who was in attendance for the summit, said there is wisdom in this approach.

“Getting someone to actually understand the power behind text analytics is the biggest challenge,” he said. “Some companies don’t want to get into this.”

Step 2: Start small
Similar to advice given when implementing a business intelligence project in general, Jones said to start small rather than become overwhelmed by large projects that require juggling lots of data.

“Be nimble, be quick,” he said, “and move on from there.”

Step 3: Get points on the board early
Starting small can help illustrate an immediate return on investment by getting points on the board early.

“No one wants to hear ‘in three years … we’ll know if this paid for itself,’” he said. “You need to show the value of the tools quickly.”

Step 4: Find your sweet spot
Jones warned that while software products may sound as though they can solve every problem an organization may run into, the fact of the matter is they can’t do all of it well. Instead, Jones said, find the text analytics software’s sweet spot.

“Sometimes you may be getting great data from social media sources or your call center, and sometimes you’ll say, ‘wow, the quality of the data we have isn’t very good,’” he said. “You need to figure out where the software makes sense.”

Daniel Eskenazi, the vice president of research and development for a media startup called EcQuant, has run into this very problem.

“Vendors claim they have good software, but that’s not necessarily true,” he said during a break.

Still, Eskenazi said while his company is evaluating options -- specifically, text analytics software that can deliver on speed and accuracy -- it’s looking to make an investment soon.

Step 5: Find your cheerleaders
“I love text analytics,” Jones said. “I love voice of the customer.” But, he said, he’s just one person in an organization, and to help get this message across, finding cheerleaders is necessary.

“Find the people who are willing to embrace these tools and who will find value from these tools,” he said. “They will become very important.”

Step 6: Understand the voice of the customer
Jones said organizations may ignore data rather than linking pieces together. For example, an organization may collect customer data for both call centers and Web surveys, but the data never meets. This can cause gaps in understanding the voice of the customer.

“You have a lot of sources to leverage, and you need to ask, ‘How do these things tie together?’ to understand the voice of the customer,” he said. “Tools are good about telling you where to focus, but they’re not good about telling you what to go do.”

Step 7: Take inventory
While lots of data is a good thing, it’s also important to think about how the data is presented. One way of doing this is to take inventory of, for example, all areas representing the voice of the customer and then work on creating a consistent message.

Think of this as pinpointing the big takeaway from across teams, Jones said, rather than presenting different views in different formats.

Step 8: Keep your customers in mind
While it’s good to work on mining customer feedback, Jones said it’s also easy to get carried away. Tools such as SurveyMonkey are readily available and easy to use, but without a little planning, they can be a turnoff to customers.

“Be respectful of your customers in terms of how many times you’re interrupting their experience with your product and services website,” he said.

Be cognizant of the number of surveys a customer may encounter, he said, ensure any duplicable questions are removed, be consistent, and don’t be lazy.

“Don’t ask questions you already have data to in your data warehouse,” he said. “If you don’t know how many times a customer is coming to your website through Web tracking, then shame on you.”

Step 9: Know your audiences’ needs
Once the text analytics program is up and running, reports will need to be generated. Who gets those reports? Who is going to care about the data?

“The answer is everyone,” Jones said, “but the reporting needs are very different.”

Some employees, such as executives, may require PowerPoint presentations to see the big picture; others may want accessible and scalable data on a specific category. Rather than bog down some employees with too much data while others are left wanting more, talk to who will be receiving reports and figure out what’s needed.

“The level of granularity in the report will change [based on the audience],” he said.

Step 10: Link it together
Data from text analytics needs to be compiled across teams to understand the voice of the customer -- from call centers to Web surveys to social media interactions. But, Jones said, data from those sources shouldn’t sit in a silo. It also needs to be linked back to an organization’s structured data as well.

Step 11: Go drinking with your partners
Silos can impact more than an organization’s data collection and analysis. When employees don’t cross units, an “us versus them” mentality may form, potentially leading to delays and miscommunication.

“You’re going to need these people,” he said. “It’s kind of like a SWAT team. How do we get this done without being stuck in the process, without this taking years of IT requests? You don’t have to drink Guinness, but I recommend it.”

It may not be beer, but find something to help break barriers.

Step 12: Talk to your customers
Analytics can produce chasms between an organization and its customers. Don’t forget the benefits of simple interaction, he said.

Step 13: Change the experience
Once a text analytics program is up and running, don’t become complacent, Jones said. Instead, think about how to continue to grow the program by taking on larger projects. Jones specifically mentioned discovering real-time uses to help change a customer’s experience with an organization.

Dig Deeper on Text analytics and text mining

Start the conversation

Send me notifications when other members comment.

Please create a username to comment.