You can adjust all of your cookie settings by navigating the tabs in this window.
When Nanigans was founded back in 2010, it was founded under the principle that performance marketing at scale was inherently broken. At the time, there were a bunch of “digital advertisers” that were trying to understand all these new performance KPIs including CPM, CPC, CPL, CPA…etc. As we looked at all of these metrics, with CPA the most advanced metric being adopted by the general marketing public at the time, we knew it was time for Nanigans. Since 2010, we’ve been honed in on solving for lifetime value (LTV), and more specifically predictive lifetime value for performance marketing at scale.
While the concept of lifetime value (LTV) is not new, being able to accurately predict from a performance marketing perspective and aggressively bid in real time towards lifetime value is…so what does lifetime value based bidding look like in today’s performance marketing world across desktop and mobile?
Take the E-commerce example above, in which John and Cindy are each making a purchase today. John is purchasing a t-shirt and a hat, while Cindy is only purchasing a t-shirt. At first glance, it appears that John is a more profitable customer as his purchase price is larger. For advertisers bidding to CPA or single purchases, future ad budgets would optimize towards more customer profiles like John’s…and that’s where it gets scary.
By focusing ad spend on CPA or first time purchases instead of lifetime value (LTV), advertisers would not only be missing out on millions but actually paying for lower value customers. In this example, Cindy is married, has two children and has plans to purchase additional products (besides just her t-shirt) past “today.”
As you can see, over the next 18 weeks, Cindy plans to make additional purchases for herself, children and husband and when compared to John is clearly the obvious choice for customer acquisition comparatively. So why don’t more advertisers focus on customers like Cindy with lifetime value (LTV) in mind?
The main reason that a customer profiles like Cindy are avoided compared to John is that Cindy costs more to acquire as a customer compared to John. Cindy, however, is more valuable over time than John, so actually advertisers should be willing to pay more to acquire her as a new customer. That’s right, you need to increase your CPA and focus solely on lifetime value as pictured below!
Customers that have the highest lifetime value (LTV) often cost more upfront. Once acquired, however, their value over time is much higher than customers who are acquired via CPA based bidding at a lower upfront cost. If your frontline metrics increase, however, such as your CPA then your boss is going to start to take notice. So how do you justify the increase in CPA? By incorporating predictive lifetime value based models into your marketing analytics.
Our goal at Nanigans is to scale performance-marketing budgets for the advertisers that use our software. The only way we’re able to do this, is if our advertisers are generating revenue, which is only possible if consumers are not only responding to product offerings they’re seeing, but actually purchasing. In performance marketing, purchase revenue is our indication that consumers had a positive experience. When we don’t see purchases from a particular segment, that’s our indication that there’s something wrong and we need to update the current strategy in place.
Referring back to the example above featuring John and Cindy, you’ll recall that the goal was to acquire as many customers like Cindy as possible, and to avoid John at all costs (as he had no intention of making future purchases). How would one do this you ask? There are a few components that go into finding customers like Cindy.
Maturity Curves: Once a customer like Cindy has been acquired, it’s imperative to instantly start tracking her purchase behavior via maturity curve analysis.
By utilizing maturity curves, advertisers can immediately start to determine whether or not the customer they acquired was of high value. In addition, early profitability recognition positively influences predicative modeling algorithms, so that similar customers can be targeted in real time using affinity models.
Affinity Models: Also known as look-a-like models, affinity models will bid to potential customers who have similar profiles to those of your most profitable customers.
In this broad example displayed above, the high affinity target includes females, aged 35 while the low affinity market is males, aged 30. The problem for advertisers is that there are only so many 35-year-old females that can be targeted, so increasing volume and scale becomes daunting. By building affinity models, however, new correlative demographic targeting opportunities (that produce similar purchasing behaviors) become apparent and allow advertisers to both quantitatively and qualitatively scale with confidence. For instance, while an initial target customer may be been women, aged 35, analytical insights may have also shown that within that segment there was a reoccurring characteristic of “Tough Mudder” enthusiasts.
By testing affinity models, an advertiser could now test women, aged 30 who are also Tough Mudder enthusiasts. This testing methodology typically incorporates numerous demographic and psychographic variables along with thousands upon thousands of segment tests in order to allow advertisers to scale with lifetime value (LTV) at the forefront of their marketing efforts.
Cohort Analysis & Revenue Tracking: Of course it’s important for you to track revenue and identify high performing cohorts so that you can continue to spend with confidence. By examining high performing cohorts, advertisers are able to understand historical performance so that similar learnings can be applied to current campaigns at scale.
Expected Revenue Tracking: While revenue to date allows advertisers to continue spending with confidence, predictive revenue allows them to scale, and scale with certainty. Per the example below, high-level metrics like spend, CTR and revenue to date are of interest, but by understanding expected revenue over time today, capitalizing on future growth happens in real time.
So how do you advertise today? Are you focused on CPA? Single purchases? If so, how confident are you that you’re spending your budgets as efficiently as you could be? By strategizing, bidding and optimizing to predictive lifetime value, you can be sure that you’re protecting yourself against low future values, and maximizing all of the investments that you can be making today, that will pay off the biggest in compounding future lifetime value.
If you wan to learn more about predictive lifetime value marketing at scale, please reach out to us at email@example.com
Performance Advertising News Roundup: Mobile In-Game Advertising Takes Off, Facebook CMO Weighs In on Rebrand, Diversity
Performance Advertising News Roundup: Late Thanksgiving’s Impact on Retailers, Five Marketing Misfires