AI Can Elevate Brand Loyalty in The Age of the Customer

One quote that was offered during Friday’s Loyalty360 webinar, “How Can AI Help Marketers Customer Insight Challenges,” presented by Manthan, was from Global Bank: “We are drowning in data and starving for insight.”

All three webinar speakers−Brandon Purcell, senior analyst, Forrester; Ujwal Dhoot, vice president marketing and e-commerce, Charming Charlie; and Varij Saurabh, director, customer analytics, Manthan−believe that artificial intelligence can help marketers optimize promotional strategies with better and more effective customer insights. 

The Age of Information dominated the 90s and early 2000s, but since, roughly, 2010, it has been The Age of the Customer, Purcell noted.

“We’re interested in taking action and capturing insights,” Purcell said. “When I think about customer insights, I think in terms of the insights lifecycle. Insights begin their lives as raw data–be it transactional, behavioral, demographic, structured, unstructured, or any and all of the above. We use analytical techniques to turn data into insights. But the lifecycle doesn’t start there. We need to take action on the insights to capture value from our customers. This is a continuous learning process. The effectiveness of our actions becomes more data that retrains our models, creating better insights and more effective actions.”

Prior to installing Manthan’s solution for retail intelligence, fashion accessories retailer Charming Charlie depended on spreadsheets to combine disparate systems data for reporting purposes. This process required the availability of technical resources to perform labor-intensive data manipulation to deliver retail reporting. As a result, Charming Charlie’s IT team was performing data dumps to support numerous reports required across the business and consequently, reports did not reflect real-time data.

Now, members of its merchandising and operations teams can build their own dashboards and reporting views specific to their requirements without placing an IT request. Manthan’s tools didn’t just improve Charming Charlie’s reporting processes but transformed how the company analyzes data.

Dhoot said Charming Charlie has reduced email unsubscribers by identifying their propensity to unsubscribe (by 20 percent), as well as improving cross-selling with derived preferences and reducing churn by targeting customers who are likely to churn.

“We changed their behavior,” Dhoot said. “We’ve increased engagement via personalized offers. We observed the repeat purchase rate was very good for the loyal customer base. However, the basket size and value was stagnant for a long time. We needed to nurture the loyal customer base to try other products. We used a context-based recommendation engine, using clustering and association rule mining techniques to derive item representation vectors that adapt to a dynamic assortment.”

Customer insights are the “gold buried in your data.”

“The shift toward AI will really transform marketing,” Purcell noted.

Dhoot said that, currently, AI functions better with CRM because that is the area that holds the most customer information.

Saurabh said AI represents a “glimpse of the future, how a marketer envisions the workplace to look like, and what tool sets will the marketer need.”

Saurabh offered his building blocks of customer marketing:

  1. Acquire a unified view of the customer, and understand their tastes and preferences
  2. Predict their value to your business, purchase behaviors, response propensities
  3. Build engagement strategies that are informed by the analytics–channel, offer, timing,
  4. Continuously improve outcome through testing, learning and optimizing
“AI is changing the way these building blocks were set up, data was analyzed, and campaigns created,” Saurabh said. “People today are very comfortable speaking and chatting with a machine. There are many examples today such as Amazon Alexa, Apple Siri, and Microsoft Cortana. They are also on the move. AI is helping create new interfaces to how analysts and marketers see the data. AI is able to understand and correctly interpret the questions, and also proactively surface most relevant insights.”

Saurabh talked about some examples of things the Manthan team has worked on during the past couple of months:

1. We send out a lot of email to customers, so managing unsubscribes is very important
2. Customers get bombarded with messages from hundreds of avenues every day, so we need to cut through the clutter and provide personalized recommendations
3. The average non-subscription based business has an attrition rate of over 50 percent, and customer churn is the single biggest challenge to profitability

He discussed how the team went about solving for each of these scenarios:

Send out emails daily

Different content

Got all opens/clicks/unsub/buyer data

Analyzed for performance, predictive scoring on likelihood to take action

Implemented and saw a 20 percent-plus reduction in churn for top 40 percent of the customers

Build extension to predictive model to automate learning and implement based on ongoing customer engagement changes

Models will change quickly over time, due to rapidly changing customer preferences and behavior

Have 10+ categories of products

We know a customer’s likelihood to buy X after Y category

We wanted to explore being relevant to the customer, standing out amongst all the junk they get so wanted to work on a recommendation engine

Too early to call results, but we plan to create many templates like this for different use cases
  • Shop by color
  • Same category of product as previous purchase
  • Different category with affinity to buy
  • Segmentation based on $ LTV
  • Segmentation based on likely to buy with offer
Churn is the single biggest challenge for any business, Saurabh added.

Meanwhile, Purcell said that customer analytics uses customer data and analytic insight to design customer-focused programs that win, serve and retain customers.

“When we asked users how important a particular data source was to business strategy, we see that the more traditional data sources from packaged apps like ERP or CRM that include planning data or customer data is more important than data from less traditional sources like video data, sensor data, weblog data, or social network data,” he said. “There is a lot of customer insight that you are leaving on the table with these less traditional sources.  As a result of the reliance on more traditional sources, analytics has always optimized for binary, one-dimensional customer behavior like purchase and marketing response from internally sourced data. Traditional analytics does not support creating insights on newer forms of data that add to the context around customer behavior.”

Purcell said that no one method is going to be able to achieve results for contextual marketing and personalization, so the next step is to uncover the dependencies between these methods so that your analytics toolkit has the right mix of methods that achieve a wider range of applications.

There remains, though, gaps between data and action.

“There’s been a real disenchantment with buzzwords (Big Data, predictive analytics, business intelligence, real-time activation, and optimization),” he said. “I believe that help is on the way. AI will truly help marketers stitch together disparate data.”

Purcell defines AI as the theory and capabilities that strive to mimic human intelligence through experience and learning.

“AI is constantly learning from new data and optimizing toward a certain target,” he explained. “I think AI can link all of these disparate phases of the customer lifecycle. Marketers are sitting on a ton of marketing content. What AI promises is to take that unstructured marketing content and actually tag it, turn it into data, and find out what messages resonate best with certain customers. At the core of AI is the ability to think and learn. It will eventually orchestrate and optimize the entire customer journey.”

Artificially intelligent solutions combine all three components (sense, think, and act).

“Marketers really need to get on board,” he said. “AI is transforming the way customers interact with brands online.”

Dhoot said that technology is changing the way we do business, not just in retail, but across all verticals.

“This will change in the future, where art and science will meld and be equally important,” he said. “The rate of innovation will only accelerate. We need to get comfortable with change, it’s the only constant. History has taught us this, and the rate of innovation is accelerating, not slowing down. A lot has been said about machines taking over our jobs, I don’t think that’s the case. Just like in merchandising, machines are a good starting point, but we need to be their guide and curation expert. Roles are going to be very dynamic in the future. We need to understand who responds to what and how we can change their behavior.”

Recent Content