Question from Reddit user:
Hi everyone, what’s the right way to choose a conversion event to optimize for in 2025?
Before you were supposed to get 100 events like add to carts within 30 days I believe when you could optimize for Initiate Checkout. Has this gone away and you can just optimize for Purchase on new accounts??
Thanks
Answer from Nabil:
The short answer is:
The old “100 events in 30 days” for a mid-funnel event like Add to Cart is largely an outdated and unreliable rule of thumb for optimizing in 2025, especially for new accounts.
While the underlying algorithm still needs conversion data to exit the “learning phase” – which is typically around 50 optimized events per Ad Set per week – the best practice now is to almost always optimize for your lowest and most valuable conversion event, which is usually Purchase, even on a new account, or Lead if you are a service business.
The key factor is no longer a strict conversion count for a higher-funnel event, but rather the quality and completeness of the data you send to Meta, which is best achieved by implementing both the Meta Pixel and the Conversions API (CAPI) for a redundant tracking setup.
The long answer is:
You are correct that the idea of optimizing for a mid-funnel event like Add to Cart or Initiate Checkout until you had enough volume to move to ‘Purchase’ used to be a very common strategy.
The official Meta guidance on the learning phase has fluctuated over time, sometimes mentioning 10 events, other times 50, but the current generally accepted goal for an Ad Set to exit the “learning phase” and optimize efficiently is to receive approximately 50 conversions for the chosen event per week.
However, the biggest shift in 2025 is the power of the algorithm and the degradation of browser-side tracking due to privacy changes like iOS 14.5 and ad blockers.
Because of this, the general consensus is to trust Meta’s machine learning and optimize for the final outcome you truly want, which is Purchase or Lead.
If you optimize for Add to Cart, the algorithm will get very good at finding people who add to cart but may never buy, which isn’t your goal.
The priority now is to maximize the quality and quantity of your actual Purchase data signal, and the best way to do this for both new and mature accounts is through a redundant tracking setup.
This means sending every event, including Purchase, via two different methods: the traditional Meta Pixel (browser-side) and the Conversions API (server-side).
The Conversions API acts as a reliable backup, ensuring that conversions that the Pixel might miss due to network issues, page loading errors, or browser privacy restrictions are still correctly tracked and sent to Meta.
For new accounts, this superior tracking dramatically increases the chances of hitting the 50 weekly conversion threshold and providing the algorithm with clean data, even with a lower initial volume, so you can and should optimize for Purchase right away.
An excellent way to implement this robust, redundant tracking, especially for accounts without an in-house development team, is by using a server-side tagging solution like Google Tag Manager (GTM) Server-Side in conjunction with a service like Stape or by hosting GTM in a Google Cloud Platform environment.
This setup allows your website to send data to your own server-side GTM container, which then securely forwards the event information, including crucial customer parameters like email and phone number (which are hashed for privacy), directly to Meta’s Conversions API.
This server-side method bypasses many of the browser-based tracking limitations, providing a much higher Event Match Quality score and thus a stronger signal for Meta’s algorithm to optimize your ad delivery and attribute conversions correctly.
By combining the Meta Pixel and the Conversions API in this manner, using unique event_id for deduplication, you give your campaign the cleanest, most complete data possible to successfully optimize for Purchase from day one.
About The Author
