Book a demo

Book a meeting with one of our marketing consultants and see how we can help you improve your ad spend and increase your LTV.

Oops! Something went wrong while submitting the form.

What if targeting customers with higher retention was possible?
Ipsy embraced Voyantis' value predictions and achieved a 27% ROAS increase on Meta ad campaigns.

Case Study
Retail, CPG
Boxycharm was acquired by Ipsy


Diverse Customer Base: Ipsy's subscription for beauty products, draws in a variety of customers, each exhibiting distinct spending patterns, creating a spectrum of spending and retention segments.
Missing Signals: Ipsy, hindered by limited foresight into future customer value, was optimizing campaigns based on cost per subscription (CPS). Even initial package purchases didn't predict if a customer retains.
Voyantis' Help: To elevate Ipsy's campaign performance, Voyantis supplied predictions of individual users' lifetime value, forecasting their likelihood to retain.
A/B Testing: Ipsy set up an A/B test on Meta to assess the impact of aligning campaigns with Voyantis' predictions.
Game-Changing: Voyantis’ value-based strategy significantly elevated Ipsy's campaign efficacy: ROAS surged by 27%, and retention rates improved by 33%.

"After we understood we have a significant lifetime value variance within our subscriber base, we worked with our Facebook and Voyantis teams to create a prediction model that uses the Conversions API to enable us to optimize our acquisition investment on Facebook for lifetime value and not just to first-time purchase value."

Alessandra Sales's profile image
Alessandra Sales
VP of Growth

The Challenge

Campaign funnel with/ without Voyantis | VisualCampaign funnel without Voyantis | Visual
Image with the number 1
Ipsy operates on a monthly subscription for beauty products, allowing subscribers the flexibility to cancel anytime.
Image with the number 2
There's significant variability in subscriber retention, leading to diverse customer lifetime value segments.
Image with the number 3
Ipsy struggled to predict individual customer retention periods, resorting to using CPS (cost per subscription) as a proxy metric.
Image with the number 4
Recognizing the limitations of CPS, Ipsy sought a more effective method to pinpoint customers likely to retain longer and contribute greater value.
Campaign funnel with Voyantis | Visual

The Pilot

1. To test Voyantis' prediction efficacy, Ipsy conducted an A/B test on a video ad campaign targeting US females aged 18-65, setting the same budget cap on the test and control group. 

2. Both groups were set on the same Meta bidding strategy (Highest Volume) for a fair comparison. The control group focused on CPS (cost per subscription), whereas the test group was optimized for predicted LTV.

3. While Voyantis’ LTV prediction provides a continuous value, the input Meta requires is binary (high/low value only) so Voyantis and Ipsy set a value threshold to signal to Meta what customers are considered “high value”.

"Using this prediction model, we were able to improve the 12-month return on ad spend by 30%"

Alessandra Sales's profile image
Alessandra Sales
VP of Growth

How does it work


Processing & AI prediction model

Voyantis developed a custom prediction model, utilizing Ipsy's distinct customer data from Ipsy's onboarding quiz and historical retention records to formulate precise Lifetime Value predictions..


Feedback to Meta

LTV predictions were sent via Voyantis’s Orchestration layer to Facebook CAPI to focus the campaign on high LTV users on Meta.



Voyantis crafted the ideal approach for transmitting predictive signals to Meta's Value-Based Bidding, ensuring early, accurate, and rich enough feedback to Meta.


Accuracy verification

The model is constantly tested and verified by the Voyantis AI engine to generate the highest accuracy of future predictions. Voyantis constantly update the model to meet Meta’s latest ad model upgrades. 

The Impact

The results of the pilot clearly indicate the advantage of using value predictions for improving campaigns outcomes.

Efficiency boost

Focus on better customers: Voyantis predictions identified those customers that are likely to have higher retention and therefore higher value, which increased the retention rate at the end of 9 months by 33%

Lasting Impact

Significant ROAS jump: By focusing on customers with better retention, the overall value of each customer is higher and with a similar CAC - ARPU increased by 15% and the ROAS jumped by 27%.

Retention rate at the end of 9 months
6 months ARPU
12 monthsARPU

Your data, our AI, +20%ROAS