Case Study
How Klar Achieved 32% Higher ROAS with Predictive LTV Signals
Klar is the leading digital financial services platform in Mexico, providing services including credit cards and loans for consumers
32%
increase in
ROAS
25%
reduction in CAC
AI Decisioning Platform
Stop letting ad platforms guess who your best customers are
Turn your first-party data into AI-powered predictive signals that train Google, Meta, and TikTok Ads to acquire your best customers.
Powering growth at the world's most ambitious companies
THE CHALLENGE
Your ad campaigns are optimizing for the wrong users
Most teams are asking for users who convert.
Not users who stay, spend, and grow.
The Gap
Platforms have a 7-day window to observe conversions and optimize campaigns on them. But for most businesses, a customer's true value won't be clear for weeks or months.
The Stand-In
So Teams fall back on Proxy Events like an install, a signup, a first transaction. but that only tells platforms to find more people who convert, not more people worth keeping.
The Result
Campaigns keep running, but the user mix quietly concentrates around the wrong people. By the time it shows up in cohort quality, the damage is months in the making.
Voyantis Acquire
The AI Decisioning Platform for performance marketing
One platform that predicts who your best customers will be, engineers your signals for every ad network, and continuously recalibrates so performance compounds over time.
Data Gold Mine
- First party
- Declarative
- Engagement
Customer Preferences
- User mix constraints
- Strategic beliefs
- Optimization guardrails
Business Context
- Performance
- Growth goals
- KPIs
Your Best Customers
High Predicted Lifetime Value
High Return On Ad Spend
See how the world’s most innovative companies are winning with Voyantis
Case Study
How Semrush Reduced CAC by 17% by Identifying High-Value Users Earlier
Semrush, a leading online visibility platform, helps businesses grow through SEO, content marketing, competitor research, and social media tools.
21%
increase in pROAS
15%
decrease in CAC
Case Study
How Lennar Spots High-Intent Homebuyers Early with Predictive Value-Based Bidding
Lennar is one of the largest homebuilders in the United States, constructing tens of thousands of homes annually across more than 20 states
40%
increase in lead-to-appointment rate
15%
decrease in cost-per-appointment
Case Study
How Upside Acquires More Revenue-Generating Users at Lower Cost with Voyantis
Upside is one of the largest cashback marketplaces in the US, connecting consumers with brick and mortar retailers through personalized offers on gas, groceries, dining, and more.
44%
increase in ROAS
15%
decrease in CAC
Built for Value-Based Optimization
Every major ad network. One platform.
Whether you're just getting started with value-based strategies or already running them in-house, Voyantis helps growth teams get more out of every dollar they spend on Google, Meta, and TikTok.
Performance Max
tCPA
App Campaigns
Search
Demand Gen
tROAS
YouTube (Video Action)
Display
Max Conversion Value
Max Conversions
Advantage+ Shopping Campaigns
Advantage+ App Campaigns
Value Optimization (VO)
Manual Sales Campaigns
Purchase Optimization
Manual App Install Campaigns
App Install Optimization
App Install
Value Optimization (VO)
Web Conversion
Product Shopping (Catalog)
Complete Payment
Under the Hood of AI Decisioning
Predictions generate insight. Signal engineering drives performance.
Voyantis uses agentic AI to make decisions about every user-level signal it generates, autonomously accounting for your business, your goals, your platforms, and your users. The result is a signal built for performance, not just insight.
Why Teams Choose Voyantis
This isn't a model you maintain. It's infrastructure that evolves with you.
Successful signal engineering takes years of iteration across thousands of campaigns, millions in experimentation, and constant upkeep. Most teams only discover how much is involved once they're already in it.
Prediction
Signal Engineering
Platform Optimization
User Quality
Operational Cost
Building in-House
What teams discover
Most LTV models are built for reporting, not real-time campaign activation. Having one isn't enough.
Without signal engineering, raw predictions get treated as ground truth — and small errors get amplified at auction scale.
Each platform has its own signal requirements, timing logic, and learning behavior, and they change without notice.
Without continuous debiasing, the user mix quietly shifts toward people who won't drive the business.
Getting this right requires 3–5 senior hires, 12–18 months to first results, and a permanent operational commitment.
Voyantis AI Decisioning
What's ready when you are
Models are purpose-built for activation and auto-retrain as your business, funnel, and user mix continuously evolve.
Autonomous signal engineering encodes, times, and calibrates predictions for how each platform actually learns.
Separate model variants deploy per ad network, each tuned to that platform's learning algorithm and updated automatically as platforms evolve.
Continuous debiasing runs in the background, keeping platforms pointed toward the users who will actually drive your business.
Your team stays focused on your product while Voyantis handles the activation layer.
4.3
B
Predictions Created
700
m
Signals Sent
300
m
Users Analyzed
Go Live Within Hours
No new system to learn.
No workflow to change.
Voyantis runs via API, connects to your existing data infrastructure, and dispatches signals directly to ad networks. Your campaigns, media buying, and workflows stay exactly as they are.
Beyond Acquisition
The same engine that acquires great users now keeps them too
Voyantis Engage brings LTV predictions and AI decisioning to your entire customer journey, from activation and retargeting to re-engagement and win-back campaigns.
