Customer lifetime value, or LTV, is a crucial metric that helps you understand how much revenue your app's average user generates.
Here's the formula:
This metric is critical when it comes to calculating the return on your investment into user acquisition, engagement, and retention.
Marketers are paying closer attention to LTV as the cost of attracting customers continues to rise. The success of their efforts, after all, depends largely on the user’s LTV. Often, attracting a user is much more expensive than their first purchase, which is why companies compare the cost of acquisition (CPA) with LTV. The higher the LTV, the better the odds of the promotion paying off.
Understanding how much revenue a potential user will generate over their entire activity period, rather than just the first transaction, allows you to quickly refine your marketing strategy and allocate your budget wisely.
But businesses don't always have enough data to calculate LTV, or they may not have analysts who can. To address that challenge, we automated the process and implemented predictive models in Yandex Direct. This innovation enables advertisers promoting mobile apps to garner more post-installation conversions and higher revenue, especially in pay-per-install campaigns.
Mobile app owners strive to attract users who are more likely to take specific actions, such as purchases, after installation. We focus on ensuring that the tools we have for advertising apps with Yandex Direct address this challenge effectively and continuously enhance their performance.
Within Yandex's mobile tracker, there's an existing score predicting LTV for app users. We utilized it to train our models and incorporated the probability of post-installation target actions into our predictions. The score will be the primary factor for selecting ads in automated strategies.
This new approach significantly boosts the number of targeted actions after installations, consequently increasing overall revenue. The impact is particularly notable in pay-per-install campaigns: during the testing phase, we observed a profit increase of up to 12% from acquired users.
More and more clients are evaluating CPA, ARPU, and ARPPU over long cohort distances and refocusing on LTV optimization. However, suitable optimization models aren’t yet available in most advertising channels.
As an agency looking to solve this client problem, we evaluate metrics manually, which greatly lengthens and slows down the process of optimizing campaigns and improving metrics. The ability to optimize app campaigns in terms of LTV will allow you to effectively work with Yandex Advertising Network and non-branded search traffic.
How predictive models work
Our predictive LTV model is a machine learning tool we use for analytics in AppMetrica and to train Yandex Direct algorithms. Depending on how users behave throughout their first day in one app, the model predicts how much time they'll spend in similar apps as well as how much revenue they'll end up earning for the app owners.
The model is trained on anonymized AppMetrica data from similar apps and the revenue or user retention data you send Yandex Direct. As a result, Yandex Direct adjusts bids in real time and attracts users promising a higher LTV for your app.
The model is trained using anonymized data from comparable apps, incorporating revenue and user retention data. Consequently, Yandex Direct dynamically adjusts bids in real time to attract users who potentially have a LTV for your app.
Standard purchase algorithms: targeting likely app installers
New Yandex Ads for App Campaigns algorithm: focusing on a high-quality audience driven by LTV
We constantly enhance our technologies and broaden the capabilities of App Campaigns in Yandex Direct to ensure that advertising gets maximum results within your optimal budget. The launch of the predictive LTV model is one of the significant steps in this direction.
Start attracting paying users with App Campaigns in Yandex
