Skip to Main Content
A green circle in front of a person
Insight

Make it Personal

How AI and machine learning are creating custom products and experiences for the masses

If you’ve ever shopped on Amazon, you’ve experienced firsthand the company’s recommendation algorithm that upsells and cross-sells products based on your tastes and personality. Perhaps you’ve even purchased one of the recommended products. If so, you aren’t alone. After its deployment, Amazon experienced a 29 percent increase in sales over the same quarter in the previous year.

Starbucks, on the other hand, launched a mobile app that combines its loyalty program with personalization, giving customers the ability to fully customize, preorder and pick up drinks from any location. Currently, the app drives approximately 22 percent of all U.S. sales. These are just two examples of the impact personalization can have on companies, yet it’s a strategy that is gaining ardent support across industries. In a study by Researchscape International, the majority of respondents (87 percent) reported a measurable lift in their personalization efforts and more than three quarters (77 percent) believed personalization should be a bigger priority for their organization.

It’s not just companies that can benefit from personalization. According to a Gartner study, the majority of respondents felt that personalized messages geared toward helping consumers gives them a better deal, and nearly half feel it saves them time.

Author

Burton McFarland

Principal Data Scientist

Percentage of customers who consider help type important; Help me get a better deal: 62% - Save me some time: 49% - Provide information I didn't have before: 46% - Make the purchase process easier: 45% - Make the purchase process less confusing: 44%

FIGURE 01

Personalization with the highest impact on consumers

What is personalization?

Essentially, personalization is a way to match the right digital content to customers. Companies are using it to present products and special offers in ways that align to their customers’ preferences and habits. It has made brands more valuable and relevant, enhanced customer engagement and satisfaction and has had a profound effect on sales and top line growth.

Often blurring the line between marketing and user experience, personalization can take many forms, including bespoke product recommendations, offers, discounts and search results, as well as creative that is matched to individual tastes and preferences.

Each of these examples is very different, requiring different data, approaches and technology to enable. However, with the right orchestrated approach, they work together to become a reinforcing experience for end users, giving them exactly what they want throughout their entire journeys.

The top 3 personalization strategies

There are three main approaches driving personalization: top down segmentation, trigger programs and AI, and machine learning (ML). Most businesses are already engaged in the first two, but the trend toward leveraging AI and ML is quickly picking up speed. Moving forward, AI-driven personalization will become increasingly common across every business with a direct path to customers.

Top down segmentation and rule-driven

Segmentation and rule-driven personalization is the process of leveraging human experts to create customer segments and sending different content to those segments. It is based on demographic splits (age, gender) or simple purchasing behavior (frequency of purchases). This manual process requires each experiment to be A/B tested.

Trigger programs rely on low-level customer behaviors to trigger marketing messages. Examples include Browse/Cart abandon messaging and retargeting. Trigger programs allow individual product personalization, but the target audience is often very small.

Trigger programs
AI & machine learning

AI and machine learning approaches involve gaining insights from large quantities of data to predict what customers want, when they want it and their preferred method of communication. This could include predicting which products they are most interested in or what content they are most likely to view.

Methods in AI and machine learning have proven to be most effective in scaling personalization to all customers, providing a repeatable way to create the most engagement.

The modern approach to personalization

Today, the most cutting-edge personalization techniques leverage deep learning. Most personalization platforms offer some form of out-of-the-box modeling along this spectrum. However, no platform offers everything, and almost none have options for deep learning. You will get the best results by combining custom-built machine learning and deep learning models along with capabilities from personalization and customer journey platforms. Doing so enables reuse of existing marketing and journey orchestration platforms, while also incorporating bespoke models for more accurate predictions. It also creates powerful flexibility for enabling customized personalization and integrating it into existing platforms across the enterprise. 

circles

The building blocks of personalization at scale

To get started with AI-driven personalization, you will need to put several technical components in place. The good news is that most organizations have already made many of these investments. In fact, they are the same pieces needed to create and distribute digital content to customers. They include:

Data collection

It all starts with collecting and managing good customer data, such as historical transactions, marketing data and click-stream analytics data from web and mobile.

Content pipeline

Personalization at scale assumes a fast content pipeline is in place. This includes content management systems’ quality control and reuse, as well as a tagging and instrumenting process.

Marketing and e-commerce delivery systems

How you reach your customers is key and involves software tools for delivering marketing content, orchestrating customer journeys, segmenting audiences, A/B testing and tracking results.

The final piece needed is the infrastructure for model training and deployment. This new addition into the technology stack allows data scientists to create the AI and ML models necessary for accurate predictions. These models generate insights about customers that can be fed into marketing and content delivery systems to create unparalleled customer experiences.

As trends, personalization and customization aren’t necessarily new. But the ability to do so at scale—holistically and cost-effectively—can only be achieved through AI and ML. By leveraging custom-built machine learning models, you can make the most accurate predictions possible about customer behavior—and deliver experiences that are timely, relevant and compelling. 

Start a conversation
Burton McFarland
Burton McFarland
Principal Data Scientist