Personalization: One-to-One Marketing With Millions

natural language processing

by BRADY EVAN WALKER

Our Roadblock to Personalization

One-to-one personalization at scale is hard, but it’s not impossible.

For all businesses, the most valuable insight into audiences lies well outside customer relationship management (CRM) systems. 80% of customer behavior data isn’t found in your database and doesn’t adhere to formal data models. It’s the dark matter and the black gold of the Internet. It’s called Unstructured Data.

As Alissa Lorentz wrote for Wired, the proliferation of data has created “a data mining gold rush that will soon have companies and organizations accruing Yottabytes (10^24) of data.”

If that’s hard to swallow, chase it with the following quote from Phil Fernandez, Chairman and CEO of Marketo:

That wealth of data can bring us closer than ever before to our customers, and give us the ability to build real relationships with every customer based on what they’re saying – and not saying. It’s the Holy Grail for marketers.

Whoever can conquer unstructured data can spot patterns, create meaning, and create solutions for some of the marketing world’s most significant hurdles. These hurdles include analytics, social media monitoring, sentiment analysis, content creation, product development, targeting, and one-to-one personalization.  

A fully featured cognitive system organizes content automatically, tagging each datum with key attributes and creating a view of the customer that evolves at pace with the individual’s evolving online and mobile habits and interests.

What Personalization Really Looks Like

“Remember that a person’s name is to that person the sweetest and most important sound in any language” — Dale Carnegie

Carnegie might be right when it comes to meeting strangers at a party, but topping an email with someone’s first name, while a step in the right direction, is far from a complete picture of one-to-one personalization.

Forrester Consulting defines personalization as, “The ability to present customers with relevant products and offers that accurately reflect their stated and unstated needs.”

In other words, it’s about knowing your customers, not how marketing enterprises traditionally “know” a customer based on data points like age, gender, location, spending habits, etc., but knowing a customer the way their friends do.

And what better way to do that than a structured, scalable approach to social media monitoring that interprets activity and logs it as data alongside the rest of your CRM!

personalization

A structured approach to social media monitoring, for instance, can clue a retailer in on a customer’s evolving lifestyle.

A pregnancy announcement made via Instagram with a baby bump selfie, properly structured with image recognition and natural language processing and fed into your CRM could cue a cascade of special offers, discounts, and congratulatory messaging.

And how many times have you seen in your own social newsfeeds a friend asking for suggestions on things to do and see on their next trip?

For travel and hospitality, the automated ability to transfer this unstructured data into an actionable context will be invaluable. Commingling insight drawn from social and blog posts with account data, travel companies can develop a truly personalized experience.

With image recognition, a cognitive computer could learn where a traveler is based on their Instagram feed and offer real-time, nearby shopping, dining, and sightseeing recommendations.

The wealth of travel blogs, travel-themed social feeds, and sites like TripAdvisor and Lonely Planet, when understood and analyzed by an intelligent computer, will provide travel and hospitality marketers the tools for expertly informed, true personalization.

For example, a honeymoon to Paris. With the ability to extract meaning from unstructured data, a cognitive computer can scan possibly millions of honeymooner travel memories, which data can filter through customer profiles to reach a set of travel recommendations, promotions, and offers most likely to jive with a particular couple’s plans and personalities.

The more you know about the customer, the more effective the personalization will be. Extending hyper-personalized offers will increase the quality of customer interaction and the likelihood of conversion.

This level of personalization can also serve as damage control.

If an online retail customer has returned more items than they’ve kept, then your deep analysis of their unstructured social media data can pore over their activity on Twitter, Pinterest, Instagram, Facebook, and Google Search to create a unique, personalized fashion profile.

With that data, your automated marketing system can offer promotions only for the products most likely to appeal to that specific shopper. It can even rearrange your product page to suit their taste, burying the products it knows won’t match the customer’s tastes.

 

personalization

Quality analytics technologies can read, understand, summarize, and visualize unstructured data en masse. This makes it easy for marketers to identify trends within topics like demographics, locations, and products.

The most powerful aspect of cognitive marketing systems is unsupervised learning. Not only will this help marketers optimize engagement with current audiences, but it could open up possibilities in markets outside of their current scope.

The Future

With smart data, automated marketing systems can decide, in real-time and at scale, how to meaningfully interact with customers.

Cognitive marketing systems could potentially scale in such a way that a company could generate a new, hyper-informed, always-on marketer for every single customer and every single potential customer.

The Content is King Paradigm is crumbling under its own weight. Marketers must realize the internet has ceased to be a monarchy. Relationships rule the kingdom, but for a one-to-one with millions, we must turn to Data as our closest advisor.

personalization

Join us next week as we dig into the other big hurdle into wrangling Unstructured Data: Image Recognition.