Deepix AI: scaling faster without fighting the algorithm

Challenge

Scaling Without Disrupting the Algorithm

The team wanted to automate budget decisions without constantly interfering with Meta’s optimization. They needed a tool flexible enough to define their own performance logic, thresholds and timing, while still leaving room for the platform algorithm to do its job.

Slow Reaction at Scale

With Meta spend reaching $50K+ per day, manually monitoring 30–50 campaigns made it difficult to react quickly to changes in performance. Scaling opportunities could be missed, while underperforming campaigns could continue spending before someone had a chance to intervene.

Optimization Across Time Zones

A large portion of the account’s spend comes from the US, so significant campaign activity happens while the team is offline. Without automation, performance changes could go unnoticed for hours, delaying both scaling and optimization decisions.

Solution

Bïrch automation complements the Meta and TikTok algorithms with an additional layer of decision-making, giving the team capabilities and control they wouldn’t have from platform optimization alone:

Flexible Automation Around the Algorithm

Deepix AI doesn’t try to automate every decision Meta makes. Instead, the team defines the boundaries for automation: CPA benchmarks, scaling aggressiveness, buffer zones and timing. Bïrch also lets them set a frequency cap for each task, controlling how often the same action can trigger for an individual ad, ad set or campaign. Their advice is to trust the data, stick to the rules, and avoid making emotional decisions or changing rules too frequently.

Meta: Tiered Scaling With a Performance Buffer

Deepix AI uses several CPA-based rules that respond differently depending on how far performance is from target. Strong CPA triggers larger budget increases, while performance closer to the target triggers moderate increases. A buffer zone prevents unnecessary changes when CPA is still within an acceptable range. As performance worsens beyond the buffer, budget cuts become progressively stronger.

Deepix AI Auto-Scale & Budget Cut Strategy

A dynamic, multi-tiered automation system that automatically scales or cuts ad set budgets based on real-time CPA relative to target.

TikTok: Faster Rules for a Faster-Moving Algorithm

The team takes a more aggressive approach on TikTok because it spends and generates conversions faster than Meta. When CPA beats target, Bïrch can increase budgets more aggressively. Because performance can also deteriorate faster, the team uses stricter downside controls. Rules are based on today’s performance rather than hourly windows.

Rules That Run Without Constant Monitoring

Once the rules are set and validated, Bïrch handles routine budget changes while the team is offline. The team can review automated actions in Bïrch logs and Ads Manager, making it easier to verify that rules are behaving as expected without manually checking every campaign throughout the day.

Results

Up to 10,000 Automated Actions Per Month

Bïrch handles up to 10,000 automated actions per month, covering routine scaling and budget reduction decisions.

11 Hours Saved Per Week

The team estimates that automation saves approximately 11 hours per week, previously spent monitoring and optimizing campaigns.

More Time for Growth Work

With routine campaign monitoring automated, the team can spend more time launching campaigns, working with creators and partners, and handling other growth activities instead of checking dozens of campaigns throughout the day.

Scaling and Optimization Without Constant Oversight

Budget increases and reductions can happen according to predefined performance rules while the team is offline. This gives the team a way to keep campaigns moving without having to react manually to every performance change.