They used to say that you are what you eat. Today, you are what you buy. You are the social media posts you click on. You are the websites you visit during the course of the day.
The digital signals you generate tell the story of who you are. When you add up all of these individual actions over time, they form what’s called a persona – a virtual reflection of your habits. When a persona contains enough relevant information, it can be used in many different ways.
Retailers and marketers have used personas for decades to finely tune their outreach to customers. The ads that appear in your email, the “customers also bought” feature, and even the placement of items on store shelves, are all built on the foundations of customer personas.
Personas are also routinely used to confirm identity. If your data history shows that you never leave the state and suddenly a first-class plane ticket to Tokyo appears on your credit card, that’s usually a strong indicator of unusual activity that may turn out to be fraud. In the absence of an actual identity verification process, businesses often fall back on personas for what’s called “risk-based adjudication.”
There are many other examples of data personas at work, from government benefits to credit applications, travel to insurance. While the practical applications of data personas vary widely, the principle remains consistent: personas connect different patterns of activity to generate insights into how individuals behave.
Limitations of Standard Data Personas
Personas are the building blocks of many critical business processes. But that doesn’t mean that they’re perfect. We’ve all seen irrelevant ads, blocked transactions, and unnecessary authorization steps, as indicators of misaligned personas. At best, these personas are deductive guesswork.
The value and applicability of data personas vary widely. Here are a few ways that standard data personas can actually hurt the businesses they’re trying to help:
Inadequate inputs: Most data personas are limited to the categories of information that happen to be available. The narrow scope of most data personas can lead to wildly inaccurate results. When customer behavior shifts due to an event that isn’t reflected in historical persona data, the lack of built-in context quickly renders it moot. Context is the key.
Insufficient history: Personas mature over time. As more digital signals are incorporated into the picture, context and insights become richer. Richer the context in turn leads to more accurate projections of expected behavior.
Static, rules-based processes: Personas are only as relevant as the processes they plug into. When decision engines are built around a single data element or factor, they can produce irrelevant results when the context surrounding that single factor changes.
Predictability: The connection between past actions and future performance isn’t always as simple or direct as a persona makes it out to be. Our actions are rarely based on an “if/then” decision tree, particularly if the data sources are too thin to capture relevant context. Forecasting behavior from narrow, single-use personas can be a perilous business.
Persona-Based Intelligence
Narrow data personas have a significant impact on business processes of all kinds. They help to create a new generation of higher-level personas which would deliver consistently accurate results, no matter the context. This approach is called Persona-Based Intelligence.
Persona-Based Intelligence adds depth and strategic perspective to standard data models. We have trained a set of highly detailed personas using massive data sets and innovative AI/ML technology. These personas draw from over 1,000 unique data elements to provide the context, predictability, and flexibility that standard personas cannot match.
The data behind these personas come from a variety of unique sources. It starts with open-source data processed with patented techniques to draw out connections and patterns that other AI/ML engines miss. That open-source information is combined with proprietary transaction data which is accessed through partnerships with key financial sector players. Finally, these personas are trained by exposure to client data sources, reflecting the lessons learned from engagements across multiple industries and use cases.
It’s these distinctive data sources and methods that produce the context and depth that any data model needs. Consumer patterns can (and will!) change for a wide variety of reasons – economic conditions, the weather, a global pandemic, political unrest, you name it. Consumers also change personas over time. This “drift” can flummox even the savviest of data models.
Ordinary personas break down when these patterns change because the data sources, they’re built on are generally narrow and limited in scope. Since they draw from such a broad foundation of data, these personas anticipate the impact of even minor changes in behavior.
These innovative personas also have a strong scientific foundation. Studies have been performed which helped to refine the links between shifting data patterns and actual consumer behavior. The result: these personas are tuned to notice the data elements which play a key role in driving new patterns.
How to Use Persona-Based Intelligence
Context-aware personas add value to any business process. They identify patterns of fraud hidden by complexity or factors beyond the scope of standard data inputs. They predict customer attrition with startling accuracy. They demonstrate the likelihood that customers will decide to purchase.
Every business process contains an ideal path – the decisions that lead to maximum revenue, higher customer satisfaction, and lower fraud risk. Persona-Based Intelligence not only identifies this ideal path but shows you how to get customers on it.
At an advanced level, businesses can also build ideal persona types that may not currently exist in historical data. Persona-Based Intelligence can then construct the means to move customers toward this new journey.
This is done by identifying the key decision points in any process and modeling customer behavior at each point. Using this detailed decision tree, we can then identify the points where a small reminder, a “nudge,” or digital intervention will make the most difference in the end result. This is a critical difference in this approach – most AI/ML simply provides insights from your data and relies on the client to propose the next steps. This radical new approach actually shows organizations how to produce the change that they want to see. Persona-Based Intelligence is undoubtedly the new frontier in the never-ending battle against fraud.
