Overview

The campaign had two main goals. First, CIFRA aimed to attract new users from Tier-1 ads regions, where competition is fierce and acquisition costs are high. Second, the team wanted to validate and optimize their HubSpot + Meta Ads integration powered by Zapier. The setup ensured that every registration was sent directly from HubSpot into Meta Ads in real time, helping the algorithm learn faster, target the right users, and maintain data accuracy. This case shows how to connect Meta Ads to HubSpot effectively for scalable growth.

Challenge

Tier-1 markets are some of the hardest places to run ads. High purchasing power comes with high competition, and small inefficiencies can multiply into wasted spend very quickly. CIFRA’s marketing team knew that success would depend on both precise audience targeting and a flawless technical integration.

Without clean, real-time data, the Meta algorithm would be forced to optimize on incomplete or distorted signals. This would slow down the learning phase, inflate the cost per registration, and limit the campaign’s potential to scale. The challenge was to demonstrate that their infrastructure could deliver accurate event tracking during a real campaign and keep acquisition costs at a level that made continued growth possible.

Approach

Let’s start with the campaign setup. Why did you use Lookalike audiences, and why three ranges?

We decided to test three different Lookalike ranges: 1%, 1–3%, and 3–5%. Each was designed to strike a different balance between quality and reach. The 1% Lookalike represented users most similar to CIFRA’s best existing customers and was expected to generate results quickly. The 1–3% range offered a moderate expansion, widening the pool without straying too far from the core profile. Finally, the 3–5% Lookalike targeted a broader audience with the potential for incremental growth once the narrower groups had proven their efficiency. This tiered approach allowed us to compare performance across different depths of similarity and to identify opportunities for scaling.

What advantages did real-time registration data provide?

This HubSpot → Zapier → Meta CAPI setup became the campaign’s backbone. Because registrations were sent straight into Meta Ads the moment they happened, the algorithm had the live feedback it needed to fine-tune targeting. Without this setup, some of the data would have been lost or delayed, leaving the algorithm to optimize on incomplete information. Instead, Meta was able to learn faster, identify high-quality users more accurately, and stabilize campaigns more quickly. This was especially critical in Tier-1 markets, where every wasted impression comes with a higher price tag.

Dynamics

Before analyzing audience performance, the team validated and optimized the HubSpot + Meta Ads connection through Zapier, fine-tuning data mapping to ensure that every registration reached Meta Ads without delay. This gave the algorithm clean signals that directly shaped Lookalike audience performance.

Which Lookalike group started delivering results first?

The 1% Lookalike was the first to generate consistent registrations. It required only a few days of training before it began performing, and it continued to deliver steady results for roughly a week and a half.

What about the wider Lookalikes?

At the start, both the 1–3% and 3–5% groups looked inefficient. For nearly two weeks, they delivered very few registrations, which could easily have led to the conclusion that they weren’t viable. But instead of cutting them too early, we allowed more time for the algorithm to learn.

How did the 3–5% Lookalike evolve?

After about a week and a half, the 3–5% group suddenly stabilized and began to generate steady registrations. What had initially looked like wasted spend quickly turned into an effective growth driver. This shift was a reminder of how important it is to respect the learning phase in Meta Ads, particularly with larger audience ranges.

At what stage did both 1% and 3–5% perform efficiently together?

In the final phase, both groups were fully trained and performing side by side. The 1% Lookalike kept quality high, and the 3–5% Lookalike delivered extra volume without pushing costs up. The mix struck the right balance between precision and scale.

Results

What was the final CPL and number of registrations?

The campaign ended with a cost per registration of $7.3–7.5 and a total of more than 700 new users.

Which success metrics did you track beyond CPL?

We monitored the entire user journey to understand where optimizations were needed. From impression to click, CTR gave us insights into how well the audience and creative aligned. From click to landing, CPC allowed us to control the cost of traffic. And from landing to registration, CR and CPL served as the ultimate KPIs, measuring conversion efficiency and acquisition cost.

How did the results compare to your expectations?

We originally expected the 1% Lookalike to be the primary driver of results. While it did perform well, the biggest surprise was the performance of the 3–5% group. After its slow start, it became just as effective as the 1% audience, giving us a much larger reach than anticipated without pushing costs up. This exceeded our forecasts and validated the scalability of the campaign.

Insights

The campaign highlighted how critical a clean Zapier Meta Ads integration is for paid acquisition in Tier-1 markets. Real-time event transfer through the HubSpot → Zapier → Meta Ads pipeline ensured that no data slipped through the cracks and the algorithm trained on precise signals. Without this layer, the cost per registration would have risen immediately, and campaign stability would have been compromised.

Another key takeaway was the importance of patience during the learning phase of Lookalike audiences. At first, both the 1–3% and 3–5% groups looked inefficient, delivering almost no registrations. However, once the algorithm had enough time to stabilize, the 3–5% audience became one of the strongest performers. This reinforced the idea that early performance is not always a reliable indicator of long-term potential, and that disciplined testing over time is essential.

The campaign proved that positive ROI is possible even in the toughest markets. With clean data flow and disciplined audience testing, CIFRA kept the cost per registration at $7.3–7.5 — low enough to make growth sustainable in high-priced Tier-1 auctions.

The only gap the team flagged was the 10–15% of events missing from the ad account. Optimizing event_id logic, unifying fields, and introducing retry-queue monitoring would further minimize data loss, accelerate algorithm training, and lower CPL even more.