Creating Unique User Experiences Through Hyper-Personalization

We’re living in an experience economy. According to the U.S. Chamber of Commerce, 67% of customers switch brands not for want of improved product features or lower price, but due to a lack of personalized engagement.1 Given this scenario, personalization cannot be an afterthought.

Marketers have always tried to reach out to customers through targeted advertising. With thousands of companies using this technique, consumers are bombarded with promotional emails, messages and calls, most of which go unopened and unanswered.

Cutting through this glut of communications requires four things: (1) reaching the right consumer, (2) with the right product, (3) at the right time and (4) through the right channel. The large volume of data that retailers and brands can collect through different touch points provides them an opportunity to engage with their customers in a highly personalized manner. Personalization done right can also be highly attractive to customers if it helps them select the perfect product in a hassle-free way. If the use of personal data is transparent and clearly beneficial, then customers should be comfortable sharing that information.

Hyper-personalization means moving beyond the simple recommendation engines pioneered by websites such as Amazon, where customers are offered additional products based on their recent purchases and similar purchases made by others. Hyper-personalization uses a much broader understanding of consumers’ behaviors, activities, context and preferences to maximize the impact of marketing campaigns, product designs and experience designs.

To create a hyper-personalized experience, the first step is to understand the customer’s genome: Who is the buyer? What is she buying? Where is she buying from? How often does she buy the product? And so on. This information, collected through different touch points, can be used to build data models that reflect the buyer’s unique characteristics.

Hyper-personalization deep dives into consumer behavior, preferences and activities to help create effective product and experience designs

The next step is to use these data models to derive useful insights about the customer and then use these insights to create tailored offers that strike a chord with the consumer and cut through all the marketing noise. On top of this, firms can go a step further in hyper-personalization by creating unique content for consumers.

While consumer goods and retail sectors are the obvious places for hyper-personalization, it can also be used in many other consumer industries such as finance, telecom and healthcare.

Customer genomes

Hyper-personalization requires a thorough understanding of the consumer’s preferences and buying patterns. The first step in building consumer DNA is collecting data about the customer’s buying behavior. This information can be collected through many different touch points. For example, when the customer visits a retailer’s website, the products viewed, the products added to the cart and the range of prices browsed can all be stored.

Each of these activities provides a unique strand of customer information. Collecting such information from different interactions and combining them creates the consumer’s DNA or genome and gives insights into the consumer’s unique buying characteristics. Figure 1 shows how different activities and interactions provide information about the customer to create the shopper’s genome.

Figure 1: Each interaction adds up to make a unique consumer genome

Each interaction adds up to make a unique consumer genome

These genomes give simple insights that can provide behavioral and contextual information. For example, they can show the number of times a customer has viewed a product before buying, purchasing frequency, general basket size, rate of returns, the kind of products that the customer returned, the customer’s favored physical stores, etc.

While many firms already use analytics to understand customers and to provide recommendations on different channels, they struggle to create a single customer view. Each channel uses its own data set. Quite often, data obtained from one platform is not used for other channels.

For example, if a customer buys clothes on an apparel brand’s website and visits their retail store the next time, their preferences from the previous purchase are not used to recommend products despite the fact that the data lies somewhere within the company. The genome provides a single customer view. With this single view of the customer, all channels can then communicate consistently with the consumer and also feed new data collected to enrich the genome.

The insights from a genome can also be used to effectively manage interactions with consumers. Customer profiles based on genomes can be used by retail store associates when interacting with the customers. They can also help to segregate customers based on their choices and to develop targeted marketing campaigns for those subgroups. Advanced analytics applied to genomes can provide insights on trending products in a region, how much more people would be willing to pay for the products, and what product quantities the firm should produce in the next cycle.

More sophisticated data models based on these genomes yield further insights such as probability of purchase, customer lifetime value, whether the customer is profitable or not, etc. The models can also help identify customer likes and dislikes, such as favorite product lines, whether the customer would be attracted by discounts and in what range, whether a positive review would influence the buyer to purchase, preferred channels of purchase, etc.

Figure 2: A genome brings together multiple data sources allowing insights that can be deployed across channels

A genome brings together multiple data sources allowing insights that can be deployed across channels

Targeting campaigns

Establishing genomes enables much faster targeting of marketing campaigns to customers. Hyper-personalization can occur at scale when marketers build campaigns from strands of genomes, targeting small groups of customers that could respond positively. For instance, a marketer can select specific strands such as probability of purchase, price paid, location, browsing history and age to build a specific targeted campaign with a higher response rate.

Perhaps the retailer or brand is sponsoring a sports event and will launch a localized campaign selling co-branded apparel for fans around the time of a specific game. Using the genome as a foundation, combined with prebuilt analytic data models, the marketer can select a small group of customers who are highly likely to purchase these products if provided with a discount voucher or a chance to win special seats at the game.

The difference between this campaign and a broad-based advertising or marketing campaign is that hundreds, if not thousands, of these micro campaigns can be run throughout a year. They give brands and retailers the ability to reduce the cost and increase the speed of campaign building and deployment.

Unique tweaks

To make these campaigns even more personal, contextual communication can be added to the process. Within a subgroup that receives the same campaign, the visual assets, pricing or language can be customized based on specific customer profiles. Consider two people, Jim and John, who are searching for the same television set on a retailer’s website. Both will be part of a group that is sent marketing material about this TV. But their genome models also give the following information:

  • Jim likes traveling, visits stores before placing an order online, does not use discount codes and will pay more for faster delivery.
  • John buys lots of superhero-themed products, likes gaming, searches on multiple websites before buying, waits for sales to make the actual purchase and reads a lot of customer reviews.

While both men can be targeted by the same campaign, the retailer can take their differences into consideration when creating messages for them.

For example, Jim could receive an email with the TV in the advertisement displaying holiday destinations. There could also be a pointer to the nearest store where he can look at TVs, recommendations with more expensive options and the promise of same-day delivery for just a few additional dollars.

Meanwhile, John could find an advertisement with the TV showing an Avengers movie trailer. Product rating and positive customer reviews could be mentioned, and there could be a discount code for use with a specific credit card for the purchase.

Though they look very insignificant, these subtle personalized expressions can influence buying decisions. For example, a leading lifestyle goods company used Infosys’ genome solution to gain insight into their customers’ buying behavior. The company used genome data to send personalized messaging to their customers in order to increase the effectiveness of their marketing campaigns. Active customers rose fourfold in a short span of time. Their repeat buyer rate shot up from 8% to 23%, and online purchases increased by 30%.

Art and science

Data science can bring together thousands of data points across hundreds of variables to enable insights never before imagined. But this does not replace the art of good marketing, branding and messaging. Effective hyper-personalization does not rely just on the data analytics, but on marketers who are empathetic with their customers and are also technically proficient with these new tools.

Effective hyper-personalization involves empathetic marketers who know how to use data analytics efficiently

One way to ensure good results is to include “thick data” analysis in the campaign process. Such qualitative, customer-centric data enriches traditional analysis by adding context to better reveal customer motivations and pain points. Thick data provides insights into how brands and products can enhance and support the everyday emotional lives of customers. Understanding this can provide strong qualitative steering for marketers as they design campaigns and product strategies.

The practical implementation of a large volume of hyper-personalized campaigns requires that companies build a sophisticated library of creative assets. Based on data models, AI-based systems can then automatically recommend and assemble this information into personalized campaigns involving the product type, flavors, colors or associated imagery that gets fed into the messaging. To work at maximum efficiency, hyper-personalization will feed off an ever-evolving set of designed creative assets.

As with so many areas of business that are being sped up and grown through artificial intelligence and automation, hyper-personalization will not replace marketing departments. But it will reshape how they design, implement and deploy campaigns in a more data-led, technology-driven, but creatively agile way.


  1. We believe brands are only as good as their worst customer experience., October 2019, Horizontal Digital


Related Stories

Connect with the Infosys Knowledge Institute

Opt in for insights from Infosys Knowledge Institute Privacy Statement