3 Tips for Perfecting Lifetime Value Optimization

Customer lifetime value optimization ties your marketing goals to your business goals. Here’s what you need to do to make the most of this important metric.

These best practices can help you optimize customer lifetime value.

Increasingly, savvy insurance carriers are recognizing the value of optimizing their customer acquisition for customer lifetime value.

If you’re unfamiliar, lifetime value-to-customer acquisition cost is a metric that compares the amount of money a carrier earns from a customer over the course of their business relationship to the amount of money that the carrier paid to acquire them. It’s superior to less sophisticated metrics like cost-per-bind or cost-per-enrollment because it takes into account the size of the consumer’s premium and how long the customer is likely to stay with the carrier.

But from a technical perspective, getting LTV-to-CAC right isn’t always easy, and it doesn’t happen overnight. In order to effectively optimize for customer lifetime value, you need to make smart decisions about how you collect data and how you adjust your bidding strategy for best results.

Here are three best practices for optimizing for customer lifetime value:

1. Build a model that’s inclusive of as much data as possible

The bedrock of lifetime value optimization is building a predictive algorithmic model that can accurately assess how much each consumer is likely to be worth to your business. By reviewing how much long-term value different kinds of shoppers bring to your bottom line, you can craft an algorithm that assigns a score to each shopper based on how much you think they’ll be worth to you. Then, you can optimize your campaigns toward your desired LTV-to-CAC target.

Crucially, you’ll want your model to be inclusive of as many key variables as possible. This way, you’re able to make the most accurate prediction about how valuable each shopper is likely to be for you.

If you’re an auto insurance carrier, this means building a predictive model that reflects the variance in lifetime value that occurs across customer variables such as credit score, currently insured status, accident history, location, homeownership status, and the number of vehicles owned. If you’re in health, life, or Medicare insurance, you’ll want to look at the kinds of policies a certain kind of shopper buys and how long those shoppers tend to stay enrolled. The more precise your model, the more effectively you’ll be able to optimize for LTV-to-CAC.

Remember, the effectiveness of your lifetime value optimization is as much about the quality of the data you put into your model as it is the accuracy of your model itself.

2. When in doubt, adjust your lifetime value scoring model—not how the model is implemented in your bidding strategy

Even the best predictive models take some tweaking, and it’s inevitable that you’ll run into moments where you’re not quite hitting your lifetime value goals. For instance, you might find that despite your best efforts, you’ve acquired customers who almost exclusively do not own homes, and that’s dragging down your overall lifetime value metric.

When this happens, carriers sometimes overcorrect by adjusting their bid modifiers inside the MediaAlpha platform—instead of simply adjusting how a variable is weighted inside the lifetime value algorithm on the front end. For instance, they might decide to remove certain kinds of non-homeowners from their targeting, or to increase their bids for all homeowners. But when you implement this approach, you might find yourself optimizing for homeownership instead of the LTV-to-CAC you’re actually trying to maximize. In this example, you might wind up acquiring too many homeowners who don’t deliver long-term value, or missing out on renters who otherwise show signs of being great customers for you.

Instead, it’s best to adjust how a variable is weighted inside your customer lifetime value algorithm on the front end. For instance, if you find that you’re acquiring too many non-homeowners and it’s hurting your performance, you can simply increase the percentage by which your algorithm multiplies a consumer’s lifetime value score when they’re a homeowner.

3. Refresh your model every 4-6 months

Insurance is a fast-moving industry, and trends in customer acquisition are always changing—particularly if you find yourself operating inside a hard market. That’s why it’s so important to regularly update your lifetime value model based on the actual results you’re seeing from the customers you acquire.

For instance, you might notice that consumers with higher credit scores are underperforming for you in comparison to how they’re weighted in your model. This means it’s time to adjust your predictive model to account for this new data point. Generally speaking, we recommend refreshing your model every 4-6 months.

Want additional information about getting lifetime value right? We’re happy to chat.

Lifetime value is the best metric for insurance advertisers to optimize toward, so it’s important to make sure that your model is well-equipped to deliver the results you desire. Once you have the right model, MediaAlpha’s managed service team of top-flight data scientists and industry experts can help you implement your customer acquisition strategy on our platform at scale.

If you have any questions about adjusting your model or executing your strategy, our team is more than happy to lend a helping hand. Just reach out to your account manager to set up a meeting. If you’re not yet a MediaAlpha client, you can schedule a time to speak with us on our website.