3. The Facebook Algorithm UNLEASHED

Facebook Algorithm: Key Components and Optimization

Understanding the Facebook Algorithm

• The Facebook algorithm is based on three key parts:

  • Advertiser Bid

  • Estimated Action Rate (EAR)

  • User Value

1. Advertiser Bid

• Contains information about:

  • Audience

  • Bid strategy: automatic, manual

  • Ad set history

  • Campaign objective: website conversion, traffic, reach, etc.

  • Optimizations

2. Estimated Action Rate (EAR)

• Facebook's prediction of a user taking a desired action.

• Stability and the Auction:

  • Optimizing for purchases: if the purchase rate is low, campaign stability suffers because Facebook isn't getting enough data.

  • Downstream funnel analysis: Facebook looks at clicks, landing page engagement, page load speed, and funnel metrics.

• Example: Optimizing for "Add to Carts"

  • High add-to-cart rates indicate a successful ad and audience match, providing Facebook with more data to optimize.

  • Balancing the objective: driving add-to-carts might not directly lead to purchases.

3. User Value

• User experience factors:

  • Ad relevance (quality metrics)

  • Landing page speed

  • Sales funnel conversion

  • User history (site visits, bounce rate)

• Data visibility:

  • Facebook Pixel enables Facebook to see similar data to Google Analytics: browser type/version, device, visits, location, IP provider.

Total Value in the Auction

• The combination of bid, estimated action rate, and user value determines auction wins.

• Long-term impact:

  • Consistently losing auctions increases costs over time.

  • Facebook aims to align advertiser goals with user goals.

  • Positive signals: user interest and purchases.

  • Negative signals: ad reports, users returning to Facebook immediately after clicking the ad.

Pixel Implementation Considerations

• Pixel delay:

  • Delaying the pixel fire to remove bad traffic can negatively impact Facebook's assessment of site loading speed, leading to higher CPMs.

  • Facebook measures the time between the ad click and the pixel fire to determine loading speed.

• Alternative: Delay the pixel event fire (e.g., view content, lead) instead of the actual pixel itself.

Importance of Estimated Action Rate and User Value

• Instability in results often stems from issues with EAR and user value.

• User value factors: ad relevance, landing page speed, landing page quality.

• Graduation Testing: Aims to maximize EAR and user value for optimal bidding and cost efficiency.

Facebook's Valuation of Users

• Facebook assigns a value to each user based on various factors (campaign setup, ad formats).

• Cost per thousand impressions (CPM): $CPM = \frac{Cost}{Impressions} * 1000$

• Not all users are equal:

  • Example 1: John Smith clicks many ads but rarely buys, resulting in a lower cost per impression.

  • Example 2: John Doe rarely clicks ads but is an active online buyer, making him more expensive to reach but also more likely to convert.

Setting Objectives and Optimizations

• The cost in the auction reflects the value to the advertiser and the optimizations in place.

• High-value customers may cost more to reach but can be more profitable.

• Finding the right pixel event to optimize for is critical (e.g., registration vs. purchase).

BPM Method and User Value

• Creating user value first leads to advertiser value.

• Key elements: messaging, offer, avatar, creatives, copy, landing page conversion, persuasion techniques.

• Benefits: lower costs, better quality customers, increased spending.

• Flywheel effect: building user value informs the pixel and ad account, helping find more ideal users.

Signals from User Behavior

• Post-click signals: high click-through rates, low "x-out" rates (users hiding the ad).

• Relevance and Quality: targeting relevance, ad quality, landing page experience, sales funnel.

Structured Approach

• The BPM method: A structured approach provides consistency and scalability.