At first glance, anonymous personalization seems like a perfect oxymoron. After all, if one were to make an anonymous donation to a charity, finding a personalized thank you card in the post might be a tad upsetting. Upon further examination, however, anonymous personalization in the context of customer experience is not only quite real, it is a worthy pursuit that enhances customer experience and drives revenue.
First, why do it? The simple answer is that customers expect the brands they engage with to know who they are at an individual level. Consider a Harris Poll survey commissioned by Redpoint where 39 percent of all customers – up from 37 percent in 2019 – said they will not do business with a company that fails to deliver a personalized experience, which customers define as knowing who they are across channels – website, in-store, mobile app, call center, etc. While that expectation solidifies the more a customer does business with a brand, it is present from the first interaction.
Two Sides of Anonymous Personalization
Before exploring how a brand can personalize an interaction with a first-time website visitor, or perhaps the user of an in-store digital kiosk, let’s get to the bottom of what we mean by personalization and the purpose of anonymous personalization besides satisfying customer expectations.
In the context of customer experience, Gartner defines personalization as the process of creating an individualized experience – a tailored interaction between two parties – with a view of enhancing the experience of the participant. Anonymous personalization adds a degree of difficulty with the potential of obviously not knowing much of anything about the audience member, and certainly not any personally identifiable information (PII).
The two parties involved in the equation are the unknown visitor and the brand. Done well, anonymous personalization should ensure a positive outcome for both sides. For the brand, that outcome may be measured by metrics tied to revenue, such as driving retention, reducing churn or signing up more loyalty members. For the customer, a positive outcome related to the enhanced customer experience might be a more user-friendly website navigation, fewer steps, or more tailored content.
In a dual value exchange, there are many ways to measure success depending on the desired outcome. Whether it’s brand awareness where success is measured by click-through rate, or a goal of capturing first-party data where success is measured by a formerly anonymous customer completing an online form, it is important that a solution that offers anonymous personalization be able to capture and track the success metrics. One reason for doing so is that it helps convert anonymous customers into known customers, while also justifying anonymous personalization by demonstrating the tangible benefits.
Self-Optimize Anonymous Personalization
As an example of anonymous personalization, consider a landing page hero image for a sporting goods company. The first time a visitor arrives at the website, anonymous personalization begins with the creation of a digital footprint. Without any PII, the brand has a first-party cookie ID and device information, which may then be filed away for future use along with any behavior from the online session – pages viewed, time on a page, clicks, etc.
One method of first-touch anonymous personalization is to perform a/b testing on a hero image. The sporting goods company, for example, could randomly select one of four hero images – a bicyclist, a hiker, a skier and a kayaker, and let that cycle of images run for an hour, day, week or another set timeframe. Through continual monitoring, the brand can determine which image was most successful in reaching the desired outcome – clicks, signs-ups, purchases, etc. After selecting a winner, the brand can then use that image for every new first-time visitor, minus a hold-out group. But knowing that audiences and preferences change, perhaps 20 percent of first-time visitors constitute the hold-out group, and that audience is again divided into four groups with each group receiving a different image (bicyclist, hiker, etc.). A/b split testing continues, self-optimizing the outcome with each iteration to determine whether to switch out the hero image.
Expand a Digital Footprint with Real Time
With a real-time decisioning engine, anonymous personalization can also be accomplished during an online session based on a first-time visitor’s behavior during the session. If an anonymous device ID zeros in on cycling products, for example, and that information is captured as part of a new digital footprint, the brand begins to build an affinity score which can then be used to personalize content while the session is ongoing – including updating the hero image and a product recommendation carousel.
Using insight to adjust personalization rules might be extended to how a first-time visitor comes to the website. The visitor might come from a link, for example, in which case a brand can identify the traffic source through a URL identifier. This information will also be included in a digital footprint and used to deliver personalization during an online session.
The concept of anonymous personalization applies to multiple channels. An in-store kiosk is one example where personalization is similar to the first-time website visitor. On a mobile app, a first-time visitor may have opted into location sharing, in which case a personalized experience might entail notification of a new store opening or another type of personalization based on the anonymous ID breaking a geofence. Even the time of day of an anonymous visit to a website or the breaking of a geofence might be important for a digital footprint, perhaps triggering an urgent call to action before the expiration of an offer.
As a digital footprint expands, a brand might deliver anonymous personalization based on a number of combinations of a/b testing combined with a traffic source, website behavior, location, time of day, etc.
Get Predictive with Machine Learning
The above examples of anonymous personalization fall into what’s described as rules-based personalization where the marketer sets rules for the content an anonymous visitor will see based on certain behavior. Another type of anonymous personalization is based on predictive analytics.
In this form of anonymous personalization, a machine learning model will segment an audience based on the desired metric or outcome, and however the audience divides according to the algorithm will determine the content an anonymous visitor receives. As an example, a model might analyze what landing pages a group of anonymous visitors looked at, how long they interacted with a page or what specific products they looked at, among other variables. A first-time visitor may then be placed into a certain segment based on commonalities between that visitor’s expanding digital footprint and those of previous first-time visitors.
With machine learning embedded into a platform with a decisioning engine, self-learning models can be built on the fly. As with the a/b testing, machine learning segmentation self-optimizes for the desired outcome. If, for example, a model predicts that a visitor has a high propensity of becoming a VIP member, the brand promotes material it knows is most relevant to other VIP members.
The brass ring, of course, is when an anonymous visitor becomes known, either through providing a name, completing a form, signing up for a loyalty program, making a purchase, etc. At this point, the digital footprint combines with everything else that is known about a customer to build a Golden Record that provides a brand with a single customer view.
Because the now-known customer has been receiving a personalized experience from the start, future personalized interactions do not seem out of place, as if the brand is just now figuring out who the customer is. Rather, the consistency of interactions as a customer proceeds along an unknown to known journey breeds familiarity; this brand knows me, and this brand cares enough about me to provide me with a consistently relevant, personalized experience.
For a video presentation from Will Stuart-Jones on the value of anonymous personalization, click here.