Optimizing for LTV? Great Decision! Here’s What you Need to Know

February 9, 2022
4 min
Ben Malka

So you’re gearing up to take your growth campaigns to the next level, by shifting to an LTV-optimized approach. 


That’s a great decision, especially when you consider how some of the biggest DTCs, such as Ipsy/BoxyCharm, have already made the shift, and seen amazing results.


LTV-optimized campaigns yield amazing results. Think +200% uplift in ROAS. But as with all good things, they come to those who wait. These amazing results do not happen overnight.


Also, there are some technicalities you best be aware of beforehand, because ad networks can initially behave in an unexpected way once they’re fed with new signals.


I’m basically here to tell you that these behaviors are not a cause of alarm. They are to be expected initially, and will smooth over, over time.


Okay, let me explain.

The Nature of LTV-optimized Campaigns

Optimizing for LTV is different from optimizing for purchases, or even value optimization. That is because it looks at long-term value, instead of short.


To conduct LTV-optimized campaigns, you have to signal ad networks of users who are expected to have high lifetime value over time—either by events based on data analysis (finding the correlation between actions done in the first few days and high value) or by using predictive LTV models. 


LTV-optimized campaigns ultimately present outstanding results, because they are wired very differently from the ground-up, compared to traditional campaigns.


LTV-optimized campaigns target users based on their predicted future value. This refers to people who purchase within the conversion window, but also, and even more importantly, beyond the purchase window. We’re looking to optimize for long-term value, as opposed to the commonly-used ‘immediate value’ (purchases within conversion windows,). We’re interested in optimizing for what users will yield over time.


Initially, it might seem a little weird. I mean, why target some users, when you can target all users? In this case, it’s a matter of quality over quantity. You don’t want just any and every customer— you’re trying to target the most valuable ones that are expected to spend the most.


Here’s a visual representation of what I mean.

When a signal is sent to focus on the most profitable users, it is done because they will be worth the most in the long-term. That is because loyalty-focused marketing is far more strategic than hit and runs. The fact that it is also a sustainable form of marketing is one of its best perks, especially for brands that turn to predictive UA for greater efficiency.

What to Expect After Launching the Campaign

By now, I’m sure you know that we live and breathe LTV-optimized campaigns.


So based on our experiences, let me explain what takes place behind-the-scenes once ad networks receive signals marking users with high LTV.

Phase One: Campaign Learning Period

The ad network will begin the process of learning the new signal. This is because it is not something the network optimizes for by default. 


It will take time for each ad network to learn your signal.


During the learning period, you might experience fluctuations in CAC. It can get high, or it might even be unstable. These fluctuations should stop once the learning period is over—with a stable budget. 


On that note, since these are expected behaviors, we suggest avoiding making any changes to the budget throughout the learning period. 


Also don’t freak out. It’s a part of the process. As you continue reading on, you’ll see how it all ultimately works out.

{{banner}}

Phase Two: Trends to Expect 

Once the ad network learns the signal, the campaign's performance metrics will be stable.


CAC might be higher than what you are used to, but as the budget grows and stabilizes, the CAC will lower, and stabilize as well. The first graph you’ll see below illustrates exactly that. 


Correlatively other performance metrics will be higher as well. resulting in an overall higher ROI. Collectively, it’s a series of wins.


The following graphs further demonstrate the expected behavior over time.


Over time, LTV-optimized campaigns bring in more revenue, 

ending up with higher ROI than the original campaigns with lower CAC.


CAC fluctuation over time: The CAC of the predictive campaign is similar to that observed in the benchmark campaign. After stabilizing, the CAC of the predictive campaign outperforms the benchmark.

Correlation between spend and CAC: Lowering spend causes a CAC inflation (and vice versa).

ROAS uplift over time: As cohorts mature over the course of 3-5 months, the ROAS of the predictive campaign becomes significantly higher than the benchmark campaign.

Higher ARPU: Predictive campaign ARPU is higher than the benchmark campaign from the get go.


Higher retention: Over time, retention rates for predictive campaigns are significantly greater than those observed in the benchmark campaigns.

The Main Takeaway: CAC is Important, But ROI is Crucial

At the end of the day, you and your team should understand that greater focus should be placed on your ad spend ROI. 


The incidental rise in CAC is an expected behavior, which is not a cause of concern when launching LTV-optimized campaigns. Because again, the ad network learns and adjusts to the new signals you send it, and even after it learns, you should still expect a similar or higher CAC compared to your usual campaigns. 


The LTV-optimized approach focuses on a smaller and more targeted group, whose value is higher-than-average.  


Over time you will see that the campaigns which optimize for maximum LTV show exponentially higher ROI. 


When you think about all the curveballs being thrown at us marketers from ad networks and operating systems, the utilization of LTV-optimized campaigns acts as the ultimate saving grace for growth teams behind DTCs.


Ben Malka
Product Marketing Manager

Stay Updated

Join our Newsletter.


Get our best content on growth marketing in your inbox
Something went wrong. Please try again.

Request Sent. We'll be in touch!
Something went wrong. Please try again.