With the introduction of improved campaign measurability, marketers have realised they might have let their addiction to installs get the better of them. The industry was tired of watching their advertising budgets drain into the bottomless pit of vanity metrics and began decrying the death of the old cost per install (CPI) strategies back in 2015.
Although the industry started down the path of addiction recovery and accepted it had a problem it hasn’t stopped some marketers from thinking that sticking to a rigid cost per acquisition (CPA) will save them from falling into the old trap of chasing unprofitable installs.
The issue is, human behavior is a lot more complicated than short-sighted CPA optimisation strategies take into consideration. Not all new users are equally valuable to an app’s long-term bottom line, and user behavior can vary widely after the initial CPA is met (repeat vs one-time users).
That’s why it is important to go beyond a fixed CPA model and instead optimise towards a model that takes into consideration customer lifetime value (CLTV) to uncover the optimal bidding strategy that can scale with the maximum potential for ROI.
So, how do mobile marketers predict the future behaviour of new app users and optimise their buying to target their highest value audiences? With a deep-dive into data and a flexible media buying model. Currently, this is best achieved through programmatic platforms and leveraging a multi-touch attribution model to go beyond fixed CPA and instead operate with effective dynamic bidding.
Going down the programmatic route
Programmatic platforms track and use identifiers and events like device ID, view, click, cookie, and fingerprint data to track the conversion journey of app users and associate these identifiers with in-app user behavior (or events). The algorithms process long lists of input data including device type, device vendor, operating system, publisher, ad format, ad size, time of day, geographical information and many other data points, to then determine the users with the highest probability of conversion, supported by lookalike modeling.
Using this input data, the platform calculates the relative value of each impression in combination with historical data and then determines if and how much to bid for each impression.
By using an intelligent programmatic platform, each conversion and subsequent purchase helps to create detailed portraits of an app’s most valuable users, and those portraits are then used to identify the best potential new users who will have similar LTVs and maximum ROI.
When armed with the understanding that not all users are equally valuable and the ability to quantify those different values and identify CLTV, the fixed CPA model betrays its inherent limitations. Fixed CPA reduces the number of publishers willing to accept campaigns, reduces the scale and reach campaigns can achieve, and additionally, limits access to many high-quality publishers and formats (that often tap the highest quality audiences).
Dynamic CPM
To avoid these limitations, the optimum bidding model for programmatic LTV optimisation is dynamic CPM (dCPM). dCPM bidding is what allows programmatic buying to have the flexibility necessary to find the highest quality users, and optimise on LTV. The “dynamic” aspect essentially means that a campaign’s max and min ad spend will not be set in a strict sense but instead change on an impression-by-impression basis, based on what the machine learning algorithms calculate to be the optimal price to pay for different inventory.
This model takes into consideration the value of the users that can be reached and the ultimate goal of hitting an ideal acquisition cost on an aggregate level, meaning bidding might be a bit higher when chasing premium inventory like video ad space to reach high LTV audiences but will be lower when it spots a deal or a good opportunity to scale.
CLTV optimisation takes “performance-based” marketing to the next level and allows advertisers to fully utilise the rich data sets offered by programmatic platforms. dCPM bidding offers a much more effective alternative to the fixed CPA models and the benefits make the transition worth it.
It might be painful to admit that there isn’t a magical fixed CPA that can save marketers from underperforming ad spend but it was never that simple.
In the fast-paced mobile advertising industry, we have to constantly optimise our strategies, find better targeting, leverage more meaningful insights, and improve performance. Failure to evolve is failure, period.
