[Study] How Much Can You Save by Turning Off an Ad Campaign Based on LTV Predictions

The best way to evaluate the effectiveness of an advertising campaign is by using ROI.

ROI is the return on investment made in a product, ad campaign, or other activity. Any ROI value greater than zero reflects net profitability, while a negative value is a clear sign of a loss-making ad campaign.

The return on investment can be calculated using the following formula: 

ROI = (Income – Advertising Costs) / Advertising Costs * 100%, where income is the difference between revenue and the cost of making a product or providing a service.

What if your ad campaign is already running and you're spending money on it, but your profit won’t be calculated until after users make purchases?

In this case, you can use LTV predictions to forecast how much a particular user will spend in your app or on your website over their lifetime with your brand.

This way, you can estimate the revenue that users attracted by the ad campaign will bring, and consequently calculate the ROI and effectiveness of an ad campaign that is already running.

First, let’s take a closer look at how LTV predictions work.

How LTV predictions work

LTV predictions are based on data about purchase history and user behavior in an application or on a website.

The longer an ad campaign runs, the more events take place, enabling more accurate LTV predictions. Go to our documentation to learn more about the accuracy of LTV predictions in various scenarios.

To generate an LTV prediction, seven days of running an ad campaign are enough. During this period, sufficient data is accumulated to decide on the campaign’s effectiveness without spending a large portion of the budget.

But is this data enough to resolve that the ad campaign must be disabled, and how much money can be saved by doing so on the seventh day? These are the questions we seek to answer in our study.

Methodology

Objectives 

Objective 1. Understand the feasibility of using LTV predictions to make decisions regarding the effectiveness of ad campaigns.

Objective 2. Identify potential cost savings that can be achieved by disabling an ineffective ad campaign at the right time based on LTV predictions.

To achieve our study objectives, we analyzed two metrics:

For Objective 1, we looked at the variance between the predicted LTV and the actual LTV.

For Objective 2, we measured the potential cost savings that could be achieved by utilizing LTV prediction data about expected losses to disable the ad campaign.

What ad campaigns did we study? 

We selected seven ad campaigns across four major gaming projects. We chose these projects because they get the most value out of LTV predictions due to a large number of ad campaigns with varying budgets, creatives, and targeted territories.

Timing of ad campaigns 

2021–2022. We selected this timeframe to ensure a significant time gap between the ad campaigns and our analysis, enabling us to make a more robust comparison between the predicted LTV and the actual LTV.

Metrics

• Predicted LTV – the projected revenue that the app will generate during the period of the ad campaign, calculated using MyTracker's machine learning models
  
• Predicted ROI – the return on investment of the ad campaign calculated based on the 180-day LTV prediction.
• Actual LTV – the actual revenue generated by the app during the ad campaign period.
• Actual ROI – the return on investment of the ad campaign calculated based on the actual revenue.
• Net Loss – the potential cost saving that could have been achieved by disabling loss-making ad campaigns based on LTV predictions (seven days after launch). It was calculated using the following formula: Total Advertising Costs – Advertising Costs for 7 Days.
  

Results of the study

Ad campaign 1 – App 1

The predicted ROI of this ad campaign was negative right from its launch in December 2021. Over the following months, it continued to deteriorate, yet the marketing team failed to decide to disable the campaign, thereby generating losses.

Ad campaign performance

• Ad costs – US$ 261 000
• Net loss – US$ 99 000

Ad campaign 2 – App 1

Another ad campaign for App 1 also showed negative ROI from its very outset, with further deterioration over time. This campaign was not disabled either, resulting in multi-million dollar losses.

Ad campaign performance

• Ad costs – US$ 215 000
• Net loss –US$ 91 000

Ad campaign 3 – App 2

Unlike previous cases, this ad campaign for App 2 showed promising results in March 2022.

Ad campaign performance

• Ad costs – US$ 40 000
• Net loss – US$ 7 500
  

Ad campaign 4 – App 3

The ad campaign for App 3 lasted three months and was also making losses right from the outset.

Ad campaign performance

• Ad costs – US$ 52 000
• Net loss – US$ 24 000

Ad campaign 5 – App 3

In this case, the marketing team did a better job of scrutinizing the ad campaign’s performance but still failed to leverage LTV predictions. Although they decided to disable the campaign 60 days after its launch, rather than waiting for six months, these 60 days were long enough to generate sizeable losses.

Ad campaign performance

• Ad costs – US$ 13 600
• Net loss – US$ 4 600

Ad campaign 6 – App 4

Similar to the previous case, the decision to disable the ad campaign was made after 60 days, which was too late and also led to losses in the millions.

Ad campaign performance

• Ad costs – US$ 24 700
• Net loss – US$ 7 000

Ad campaign 7 – App 4

The marketing team of this ad campaign run for App 4 also relied on the actual data over a 60-day period to evaluate the campaign’s performance.

Ad campaign performance

• Ad costs – US$ 28 000
• Net loss – US$ 11 800

Conclusions

In this study, we reviewed seven ad campaigns with different budgets and approaches to advertising management. All of them show that:

• Waiting for an actual ROI of an advertising campaign can result in losses in the millions.
  
  While evaluating ad campaigns post-completion can yield accurate results, it does not prevent wasteful spending on ineffective ads.
  
  
• Leveraging LTV predictions to calculate ROI makes it possible to evaluate the effectiveness of an advertising campaign just seven days after its launch.
  
  Even where the decision to disable the ad campaign was made after 60 days, it was way too late and resulted in millions of dollars in losses. By using LTV predictions, it is possible to assess the effectiveness of a campaign at an early stage and avoid wasting ad spend.

How to predict LTV

Predict on your own

You can make your own prediction using instructions put together by our experts. To do this, download our free Ebook How to Predict a Mobile App’s LTV on Your Own.

Use our future LTV calculator for subscriptions

You can also make use of our free future LTV calculator for subscriptions – just enter your subscription parameters at the campaign launch and get a prediction.

Enable MyTracker’s free prediction of revenue

Use MyTracker’s LTV predictions to avoid wasting your ad budgets. This is what an LTV prediction report by MyTracker looks like:

Connect MyTracker and set LTV predictions using our documentation or request a demo to receive expert guidance from our team.