This content is part of the Essential Guide: Customer data management tools shape personalization, boost CX

6 tips for effective customer data mining

Digging into customer data can improve sales opportunities -- but how do you balance that against data privacy concerns? Get insights from data professionals.

Retail and marketing organizations are collecting massive amounts of data on customers, but they're not always getting all the use out of this information that they could. As new privacy regulations promise to constrain the use and sharing of private data, it's becoming increasingly important to use the data wisely.

In fact, according to a recent Adobe survey of nearly 13,000 marketing and advertising professionals, the biggest opportunities B2C companies saw for 2019 was the use of personalized, data-driven marketing. In addition, 55% of organizations said the biggest shift they're seeing is the better use of data for more effective audience segmentation and targeting. In second place, at 42% in the Adobe survey, was improving customer intelligence and insights for a holistic customer view.

The following are some expert tips for more effective customer data mining.

1. Utilize the past to enable the future

Shopping recommendations are typically the first priority of retail website designers. Here, as with many other data mining challenges, the more data and the easier it is to get to, the better.

Stephen Lamb, senior director of digital marketing and media at Chico's Distribution Services LLC, a Florida-based women's clothing retailer, said his company has an internal data lake in which it gathers all information that might be needed, such as browsing and purchase histories.

"There are customer analytics in a single view of the customer file," he said. "There are store analytics and web analytics about her shopping habits that also help us paint that picture."

Then, he said, data scientists can write a query to the data lake to find customers who like to buy at a discount. Customer data is then correlated to that of other customers and to product data, including the text of customer reviews.

Such customer data mining can help a company predict when customers are likely to buy again and time the release of marketing messages to make them more effective.

Similarly, Chico's uses data mining and analytics to predict when a customer is about to stop shopping with Chico's and become a lapsed customer.

"We would serve different messages," he said.

The data infrastructure platforms to make all this happen are getting better all the time, according to Ben Lorica, chief data scientist at O'Reilly Media, an online learning company for enterprises.

In addition, the platforms are getting better and collecting and analyzing the data in real time, while preserving privacy and security, he said.

2. Speak to customers in their own language

Salespeople know to adapt their pitch based on who the customer is and what they're looking for. The same is true for the digital channels.

Chico's Lamb found that his company's customers might prefer different communication channels or to be contacted at different times of day or different days of the week. "We're always looking for the best way to speak to our customers how she wants, where she wants," he said.

Chico's opted for a commercial platform, Vibes, which allows for personalization of mobile communications.

"Mobile messaging is probably one of the more critical things that we have to connect to the customers," Lamb said. "Mobile messaging is a very personal experience for her and is something we take very seriously."

Vibes also has built-in artificial intelligence, including natural language processing powered by Amazon Comprehend, which allows brands to analyze conversations with their customers to help predict mobile customer growth, churn and engagement trends.

3. Democratize data

As with Chico's personal shoppers, customer data mining can be very powerful when conducted by people who can make the most out of the information.

Too often, though, access to data is concentrated to a limited number of employees, said Robert Johnson, CTO of Interana, a customer data platform company.

"A common misconception is that only data scientists are able to find or understand this information," he said. "The most successful organizations we see are the ones who are able to get this visibility into the hands of the business people who already have context about their products and customers."

Marketing and advertising experts, in particular, should be as close to the data as possible, he said. "They are the ones who know what analysis and what questions will drive the most value to the business."

4. Prepare for the end of cookies

Over the last two years, both Safari and Mozilla have restricted the use of marketing tracking cookies, turning them off by default in the majority of cases. Now, Google is getting ready to tackle cookies as well.

"Cookies are the way every other third party collects data about consumers across the internet," said Dave Mariani, founder and chief strategy officer of AtScale, a data virtualization platform vendor. Mariani previously ran data and analytics for Yahoo's audience and advertising businesses. "If cookies go away, only Google and Facebook will have the ability to see consumers across the internet because they have the presence, the Facebook login, the Google login, the Google services, that sees you everywhere."

One of the biggest enablers of cutting-edge and effective customer data mining is the move to the cloud.

Retailers will now be limited only to the data that they collect on their own users on their own sites.

In addition, the General Data Protection Regulation and the new California Consumer Privacy Act, will make it more difficult to share data with third parties. To respond, retailers will have to get more proactive about customer data mining, Mariani said.

"When the customers go and visit the merchant website, put something in their basket and go through the checkout, the merchant can get that signal back and immediately present them with another offer," he said.

The key is to tighten the feedback loop, to continue to work with that customer right in that moment.

"Some people will call it more creepy," he said. "But you're going to see more of that."

5. Embrace the future

Too often, analytics is a backward-facing function, said Waleed Ayoub, CTO of Rubikloud Technologies, which makes AI software to help retailers make intelligent decisions.

"As a result, shoppers are given cookie-cutter personas that do not evolve along with their shopping behaviors," he said.

But machine learning can help companies make predictions about future behaviors and create new marketing opportunities.

"In fact, you can even map a set of actions to increase the chance of customers doing certain things," he said. "This is what's usually called 'building a customer lifecycle.' In fact, with artificial intelligence, you can model this behavior down to the individual customer level, even in very data sparse conditions."

He said retailers are already using this approach to help customers move from low-margin, high-frequency categories to high-margin, low-frequency ones. "Tactics that leverage machine learning in this way have real and measurable impacts on overall business performance."

6. Look to the cloud

One of the biggest enablers of cutting-edge and effective customer data mining is the move to the cloud. Cloud platforms offer easy storage of massive amounts of data and allow for quick integration with third-party analytics products and services.

"A lot of the new tools are cloud-first tools," said Sanjay Manandhar, CTO of ISM Connect, a digital marketing platform that serves more than 100 million customers and more than 100 venues, including NASCAR, the Washington Redskins, Minor League Baseball and concerts and conferences. "The old days of business intelligence are quite far behind us. Now, there are a lot of good tools."

Some companies build their own platforms to address issues such as latency and cost, he said. "But we are quite comfortable right now going with Amazon tools."

And while the company has its own machine learning experts and computer vision team, there are also off-the-shelf cloud-based tools that are very appealing.

"Something like Google Analytics can work without much effort, and you can overlay things on top of it," he said.

Dig Deeper on Data science and analytics

Data Management
SearchAWS
Content Management
SearchOracle
SearchSAP
Close