Skip to content
English
Contact us
  • There are no suggestions because the search field is empty.

Meta Best Practices

Stage: Specialist Playbooks

Type: Playbook — Platform-recommended best practices

Reading time: ~18 minutes

Positioning note: These are platform-recommended best practices — advice directly from Meta. They have not been independently validated through Fospha Deep Dives. For Fospha-validated, proprietary best practices built from real customer data, see Module 4.2 (Google Deep Dive).

Advantage+ and Automation

Meta's Andromeda retrieval engine now rewards automation. Manual campaigns miss creative-user matches and undermine learning. Implement Advantage+ Shopping Campaigns (ASC) to automate creative, audience, and budget optimisation.

Allow at least 2 weeks for the model to stabilise after enabling Advantage+. Manual setups can still be used for testing new markets or creative hypotheses, but automation should be the default.

Signal Health: Pixel and CAPI

Meta's ML relies on high-fidelity conversion data. Low Event Match Quality (EMQ) reduces match quality between a brand's CRM and Meta's user graph.

Red Flag: Weak signals or known Pixel/CAPI errors. Audit both Pixel and CAPI setup, ensure deduplication and full parameter passback. Target EMQ of 6 or above. Re-check after any site or GTM change.

Account Structure: Do Not Segment

Each ad account is an isolated data ecosystem — learnings do not transfer between accounts. A temporary CPM improvement after migration is likely exploration, not genuine efficiency.

Best Practice: Keep historical accounts active. Focus on creative rotation and structural consolidation first. Migration = last resort. New accounts take 2-3 weeks to regain stability.

Learning Phase Management

Meta needs approximately 50 optimisation events per week per ad set. Fragmented structures dilute signal density and stall learning. Ad set stuck in Learning for 7+ days = red flag.

How to Fix: Consolidate ad sets to concentrate signal. Alternatively, optimise for higher-funnel event or increase budget. Consider Advantage Campaign Budget (ACB). Avoid major edits that restart learning.

Budget Stability

Budget changes reset learning by altering auction expectations and pacing. CPM spikes are often caused by instability, not competition.

Best Practice: Keep budgets steady for 7+ days post-launch. Scale gradually in 10-20% increments. Plan pacing 2-3 days before peak. Avoid simultaneous creative or bid edits alongside budget changes.

Creative Fatigue

As engagement drops, quality ranking declines, leading to lower ad score and higher CPM. High CPM + low CTR/ROAS = classic creative fatigue.

Best Practice: Refresh creative frequently. Rotate via arcs or sequences instead of relaunching entire ad sets — preserves learning continuity.

Audience Management

Overlapping or narrow audiences cause internal competition, reducing budget efficiency and raising CPMs.

Best Practice: Merge similar ad sets and zoom out on granular definitions. Use Advantage+ Audience or broad targeting with strategic exclusions. Exclude existing customers when testing top-of-funnel.

Placements: Go Broad

Meta's auction optimises across all surfaces. Restricting placements limits inventory and drives up CPM.

Best Practice: Enable Advantage+ Placements and allow the system to find lowest-cost results. Test exclusions gradually rather than restricting from the start.

Pre-Peak Strategy

1. Seed the Model Early: Index more budget to Reach and ThruPlay campaigns in October. These impressions expand retargeting pools and improve Estimated Action Rate for peak auctions. Frame as auction priming and audience expansion — not direct ROI.

2. Pace Your Budgets: Sudden surges overwhelm pacing and cause CPM volatility. Ramp budgets gradually (15-20% daily) 2-3 days before campaign start.

Test and Scale: Keep It In-Situ

Transferring a strong ad set to a new campaign restarts learning even if assets performed well originally.

Best Practice: In-situ testing — only rotate creatives once learnings complete. During peak, prioritise stability over frequent testing. Small tweaks beat full rebuilds.

Quick Reference: 11 Red Flags

  1. Not using Advantage+ (ASC) — missing auction opportunities
  2. Weak Pixel/CAPI signals — EMQ below 6
  3. Segmented or migrated ad accounts — learnings lost
  4. Ad sets stuck in Learning 7+ days — not enough events
  5. Frequent budget edits — causing auction resets
  6. High CPM + low CTR — creative fatigue
  7. Overlapping or narrow audiences — internal competition
  8. Limited placements — restricted inventory
  9. Under-investing in TOF pre-peak — missing audience priming
  10. No budget pacing plan — CPM volatility at peak
  11. Splitting test and scale campaigns — restarts learning