Daniel Victorino
Mahta Bio reduced customer acquisition cost by 35% by validating three creative approaches with Galaxies Synthetic Personas before running any paid media investment. Validation took 48 hours. The winning version went live already optimized, without a paid learning period.
Here are Mahta Bio’s surprising numbers:
35% reduction in campaign CAC Mahta Bio case | 48h to validate three different creatives Galaxies platform | R$70k additional effective budget per month Estimate with R$200k/month investment | 0 days of paid media optimization Campaign launched with validated creative |
Before spending one real on media
Mahta Bio’s media budget was finite. The team had three different creative approaches for the new acquisition campaign and needed to choose one, leaving the other two in the drawer.
The usual choice would be to run all three in an A/B test for two to four weeks, let the algorithm learn, spend budget during the optimization period on creatives that do not convert, and only later discover which one worked after the money was gone.
Mahta Bio did it differently. Before running anything, it submitted the three versions for validation with Synthetic Personas representative of the target audience. Forty-eight hours later, the team knew which version was more likely to convert, which objections each route created, and which message should be refined.
The result was a 35% reduction in CAC in the first campaign, eliminating part of the waste normally generated by the initial learning phase of media algorithms.
The situation: three creatives, one budget, zero certainty
What problem did Mahta Bio need to solve? Choose, among three different creative approaches for an acquisition campaign, which would generate more conversion for the target audience without spending media budget on learning. The team needed evidence before launch, not after the campaign had already consumed budget. |
Mahta Bio operates in a competitive market with short campaign cycles. Each month with CAC above target directly affects the company’s growth cost. The marketing team knew the three creative approaches had potential, but did not know which one would generate more intent among the priority audience.
Version A: focus on immediate functional benefit, with a direct message about the product result.
Version B: focus on social proof, testimonials, and other users’ results as the main element.
Version C: focus on competitive differentiation, highlighting why the product was different from alternatives.
On platforms such as Meta Ads and Google Ads, A/B tests require a learning period, budget consumption, and weeks until performance metrics stabilize. The question was: how can the winning creative be identified before spending the budget?
Creative validation in 48 hours with Synthetic Personas
How were Synthetic Personas used to validate Mahta Bio’s creatives? The three creative versions were submitted to Synthetic Personas representative of Mahta Bio’s target audience in the Galaxies platform. In 48 hours, the platform identified the version with higher declared conversion intent, the main objections, and the adjustments needed before launch. |
The process was direct. Mahta Bio’s marketing team described the target-audience profile and the hypotheses behind each version. Galaxies created Synthetic Personas representative of that audience and the team submitted the three versions for evaluation.
The questions asked to the personas were objective: which ad catches more attention? Is the message clear? What would make you click? What would prevent you from clicking? If you saw this ad in your feed, what would your reaction be?
Within 48 hours, the team had answers to all these questions and identified which creative generated higher conversion intent by persona before investing any paid-media budget. Version A was the clear winner in purchase intent and message clarity.
With that data, the team made targeted adjustments to Version A, addressed the credibility objection the analysis had identified, and went live with a creative that had already been validated with a representation of the real audience.
What are Synthetic Personas and how do they apply to real audiences?
Synthetic Personas are behavioral representations created with artificial intelligence to simulate how consumer groups tend to decide, react, and behave when exposed to different stimuli.
They are built from aggregated and anonymized real data, without representing specific individuals or reproducing answers from real people.
In practice, they allow teams to validate messages, test hypotheses, predict objections, and simulate market decisions before campaigns, launches, or strategic changes happen in the real world.
Unlike traditional personas, which are usually static descriptions based on interviews and hypotheses, Synthetic Personas are dynamic, simulatable, and behavior-oriented.
This makes it possible to explore scenarios, identify decision patterns, and reduce uncertainty before real investment in media, product, or communication.
The result: 35% lower CAC, with no optimization period
What was Mahta Bio’s result with Synthetic Personas? CAC 35% lower than the company’s historical averages, already in the first campaign after synthetic validation. No paid media optimization days. The version that went live came directly from the synthetic validation stage, with only targeted adjustments before launch. |
The campaign went live with the adjusted Version A. No A/B test, no optimization period, no weeks of elevated CAC while the algorithm learned.
The result: CAC 35% lower than the company’s historical averages.
To put this number in perspective: with a monthly paid-media investment of R$200,000, a 35% reduction in CAC represents R$70,000 in additional effective budget per month. Without increasing investment. Without hiring more people. Only by entering the market with a validated creative route.
What this result means in practice Over 12 months, maintaining CAC 35% lower represents the equivalent of R$840,000 in additional effective budget for a company investing R$200,000 per month in media. The cost of validation with Synthetic Personas is a fraction of the budget that would otherwise be spent learning in-market. |
What this case teaches about campaign validation
What is the difference between testing creatives with Synthetic Personas and testing in the real market? Testing in the real market uses media budget and generates elevated CAC during the learning curve, which can last weeks. Testing with Synthetic Personas anticipates creative learning before launch, using research cost instead of media cost and delivering qualitative explanations for each result. |
The traditional campaign-launch model starts from a premise that is rarely questioned: you learn from the market. The creative goes live, you measure what works, adjust, and optimize. The cost of learning is treated as part of the media budget.
The Mahta Bio case shows there is an alternative: learn before going to market, with the real audience represented by Synthetic Personas, and reach media buying with the creative already optimized.
The performance difference is expressive. But the process change is simple: add a synthetic validation step before any media placement decision.
Dimension Traditional media A/B test Validation with Synthetic Personas | Duration 2 to 4 weeks 48 hours | Learning cost Media budget spent on suboptimal creatives Research cost, up to 93% lower | CAC impact Elevated CAC during learning curve Optimized creative from day one | Qualitative insight Limited to metrics Objections and reasons by persona | Decision moment After launch Before launch |
Frequently asked questions
How long does it take to see results with Synthetic Personas?
Mahta Bio received validation insights in 48 hours after submitting the three creatives. Creative adjustment and media launch happened the following week. The 35% lower CAC result was recorded in the first campaign launched with the validated creative.
Do Synthetic Personas replace traditional A/B tests?
They complement A/B tests with practical advantages: results in 48 hours, research cost instead of media cost, and qualitative insight into the “why” behind each result. The Mahta Bio case validated synthetically before any media launch and went live with an already optimized creative.
How was the 35% CAC reduction calculated?
Mahta Bio compared the CAC of the campaign validated with Synthetic Personas with historical averages from previous campaigns that had run without prior validation. The 35% difference represents the savings generated by entering the market with an already optimized creative, without a paid learning period.
How can CAC be reduced before launching campaigns?
By validating creatives with Synthetic Personas representative of the target audience before any media placement. Galaxies identifies which approach generates more conversion intent and which objections each version raises in 48 hours. The creative goes live already optimized, reducing CAC from the first day of campaign.
How can ads be validated before investing in media?
By submitting creatives to Synthetic Personas representative of the target audience before media placement. In Galaxies, the team describes the concept or shares the copy, and the persona evaluates message clarity, click intent, and main objections. The process takes 48 hours and costs a fraction of the media budget that would be spent learning in the real market.
How can creative performance be predicted?
With Synthetic Personas, teams can estimate declared interest, message clarity, main objections by segment, and performance comparison between versions before any media placement. It is not a CTR prediction, but a map of how the audience tends to react, with granularity by audience cluster.
Do Synthetic Personas work for Meta Ads?
Yes. The flow is direct: the team submits creatives to personas before creating campaigns in Meta. Personas evaluate which approach converts better for each audience segment, which copies generate more click intent, and which visual or textual elements create objections. The winning version enters Meta already validated.
How can waste be avoided during the algorithm learning phase?
Meta’s learning phase consumes budget with below-expected performance because the algorithm still does not know which creative works best for which audience. Synthetic Personas solve this by anticipating that information before media placement, reducing learning time and the cost associated with it.
Other clients using Synthetic Personas for marketing decisions include Bradesco Seguros, Nestlé, and TikTok. See all cases at galaxies.com.br/pt/clientes.
See how to validate campaigns before investing in paid media with Galaxies Synthetic Personas
See all result cases
Galaxies