Perfect Price Discrimination can be Achieved with Universe Testing and Programmatic Advertising

I can’t tell you the number of times I’ve walked into a meeting (or Zoomed into a meeting lately) and had some business unit leader tell me they want to double the amount of impressions we deliver on their behalf next year. Yes, increasing impressions and casting a wider net definitely helps generate more accounts, since all else held equal, pouring more people into the top of the funnel means more people land at the bottom of the funnel. But before we start investing heavily to get more people at the top, we need to pause to make sure it’s the right population. In these meetings, I’ve always stressed that marketing really boils down to solving two problems: how to identify a target consumer, and how to make that target consumer become a client.

Optimizing for conversion doesn’t help if the target isn’t accurate

Just blindly looking for additional customers doesn’t work though. It’s impossible to come up with a marketing strategy without even knowing what the market segment is. There needs to be a target customer so our average revenue per customer covers our cost per acquisition. Now, having worked in financial services supporting credit products, I may have a skewed view of customer targeting. For credit products like loans or credit cards, usually the consumer that expresses the highest demand for the product is also the consumer who is most likely to default on that obligation. It’s the reason why FICO dings your credit score for having hard pulls on your credit report, the very act of looking for an extension of credit may signal financial turbulence.

Hard Inquiries (in red) signal credit-seeking behavior

How do we Find our Target Customer?

Regardless of this industry-specific quirk, the goal of all e-commerce (or any business, really) should to be maximize revenue by increasing both the number of consumers, and the average lifetime value of each consumer. In my last post, I talked using metrics to optimize conversion funnels, so I feel pretty good about actually getting new customers once the audience is identified. What’s still missing however is who our audience is.

When I was working at Chase card services, we tackled the challenge of defining our audience by doing a universe test. A universe test here involved showing credit card ads to everyone in Chases’s prospect database, and approving well below the standard underwriting threshold. While a universe test like this guarantees that a portion of the res ponders will not generate enough revenue to cover their particular CPA, it does shed invaluable insight into who the target customers should be. In addition to defining the demographics and psychographics of each segment, this also tells us how much average revenue each customer segment brings to the firm. In other words, we know exactly how high the CPA can be for any given customer segment to still remain profitable.

How do we Implement Perfect Price Discrimination?

So now that we know who our target is, how to we activate them? If it’s possible to achieve dynamic CPAs for each customer segment, then in theory it’s possible to achieve a form of perfect price discrimination. In classical price discrimination, the cost of the good varies with the consumer’s willingness to pay. In other words, with perfect information, you can increase the revenue side of the profit equation. As a marketer, we can use a variation of price discrimination to instead dynamically lower the cost aspect of the profit equation to still increase revenue.

In the data-rich environment we live in today, DMPs such as Acxiom or Neustar provide us with the demographic and psycographic information necessary to segment the audience. Additionally, they also offer audience management solutions with programmatic advertisement vendors such as Google DoubleClick and Adwords. By leveraging the DMPs’ identity solution services, it’s possible to bid different amounts on each segment to differentiate according to profitability.

While perfect information is still quite a ways off, the information we are able to gather today already allow for pricing discrimination. However, with reputation mattering more than ever, implementing pricing discrimination on the consumer side is sure to backfire for discrimination purposes. Implementing it on the customer acquisition side is the safer method for buoying the bottom line without the accompanied reputational risk.

Marketing Analytics Professional | NYU Integrated Marketing Student