Using Data for Loyalty Program Design

When More isn’t Better

At last count, 2.5 quintillion bytes of data are created each day. But let’s be honest, not many of us have any reference for that kind of volume. In fact, 90% of the world’s data has been created in the last two years. But in that short span of time, most businesses have been convinced of one thing: you must collect all of the data you can because more data means better outcomes.

Unfortunately, the harsh reality in the world of big data is that it isn’t proving to be as useful as we all thought, mostly because it’s really difficult to give all of that data meaning. This requires aggregation and analysis, usually by a third party, before feeding those insights into a marketing strategy to be used effectively.

At Clutch, we talk a lot about data-driven loyalty programs and personalization. This can mean a lot of things, depending on the end goal or industry involved. So, let’s lift the curtain behind this dark art of data analysis and explain how we use it to structure effective loyalty programs. What do we look for? How do we know when we’ve found it? Then, what do we do with it?

Historical Context

Before we set the structure for a loyalty or engagement program, we ingest a hearty amount of data to analyze. Ideally, we want to analyze about three years of data. This can be transactional data, product SKUs, behavioral information, or ideally all of the above. In it, we look for trends, or what we often refer to as tipping points, that present opportunities for us to motivate positive behaviors and mitigate negative ones.

Tipping Points

Here are a few common tipping points that can be incorporated into the design of a loyalty program. While indicators like average order value or number of visits can be broadly applied across different industries, the values used to structure a particular program are based on the unique insights derived from that brand’s data and objectives for the program. This customized approach is a departure from a one-size-fits-all loyalty strategy and has proven to generate higher ROI.

Spend

How much does someone need to spend before they become significantly more valuable to your brand? As you might expect, this is a metric that varies widely across categories. In any case, it’s a critical insight that can be used to build a strategy that gets more customers to that magic number.

Number of Visits

Getting a customer to try something once is often the easy part, but driving that second or third purchase is often where a pivotal change can occur. There is often a compounding effect that occurs after each subsequent purchase, making each visit significantly more important than the last, but the key is to identify the exact number of visits that precipitates a shift in future behavior.

Time Between Purchases

Often, the initial days or weeks after the first purchase are the most critical in terms of establishing an ongoing relationship with that customer. How much time can elapse between visits before the likelihood of the customer returning deteriorates? This is another objective we use to design a program that can get someone back in to shop before it’s too late.

Influential SKU Analysis

Here we look at individual products and their role in the shopping experience. Are there specific products that are tied to positive behaviors like basket size or frequency? If so, how can we design a program that motivates more customers to purchase that product?

Impact of Customer Identification

Overwhelmingly and across industries, when a customer identifies themselves to a brand, their value over time significantly increases. Whether by joining a loyalty or membership program, electing to receive their receipt by email, opting in to text messages, or using a unique offer, this is a priority of almost all loyalty programs. The act of identification not only drives other key business KPIs, but allows the program to build the customer profile and design communications more broadly around their unique preferences.

These are just a few examples of tipping points we look for when analyzing data and structuring a program, but there are often many more surprises in store when we dig deeper into the data. Those surprises can lead to new and creative ways to design a program in order to achieve specific objectives.

The important thing to remember is that the preliminary work is what will make a program successful, because it must be unique to the brand, their customers and their specific behaviors. When programs fail, it is usually because they are not a reflection of the end customer.

‍