A Meta account that has run for several years usually accumulates inherited rules: optimise for views before cart events, isolate every interest in a separate ad set, duplicate a winner at a higher budget, or build a separate structure for retargeting. Some practices once served a purpose, and none should be declared universally dead. In 2026, ask three better questions. What business constraint does the setting solve? Does it fragment data and budget? Can a controlled test still prove its value?
Account warm-up and event sequencing
- Warming an account with view-content or add-to-cart campaigns
- Waiting for a pixel to become warm before launching
- Optimising for add-to-cart before purchase
If the commercial objective is a purchase, shallower events may generate cheaper signals while steering delivery towards people who browse or add to cart more readily. Optimise for the deepest event that occurs often enough to learn from. When a new account lacks data, reduce test scope, improve tracking and strengthen creative rather than treating low-intent optimisation as a mandatory ritual.
Audience fragmentation and funnel structures
- Defaulting to interest targeting
- Assuming every product needs lookalike audiences
- Using fixed TOF, MOF and BOF structures
- Creating separate retargeting audiences and campaigns
- Running large audience-testing structures
- Testing dozens of interest stacks
- Building one campaign or ad set for every interest
- Splitting gender or age brackets by default
- Manually excluding many audiences in every ad set
Meta's current guidance emphasises broader audience opportunity, account simplification and automated delivery. Segmentation can still be appropriate for legal restrictions, language, geography, inventory, margin, materially different creative needs or separate budget ownership. Otherwise, too many ad sets leave each with fewer conversions and less budget. Establish a simpler baseline, then test only the segments supported by a real business reason.
Duplication, budgets and bid controls
- Duplicating a winning campaign at a much higher budget
- Using inflated cost caps to control spend
- Using tight bid caps to protect performance
Duplicating campaigns can restart learning and create overlapping delivery. Cost caps and bid caps need to reflect conversion value, allowable acquisition cost and sufficient delivery volume. A tight constraint may suppress spend; a loose one may provide little control. Record budget changes, learning status, customer quality and at least one complete purchase cycle rather than judging a strategy from one day's CPA.
Placements and daily intervention
- Micromanaging placements every day
- Frequently editing ad sets in response to short-term volatility
Meta recommends broader placement coverage and automated delivery, while brand safety, aspect ratio, regulation and landing experience may still require controls. Confirm that the creative is legible and tracking is accurate in each placement, then compare automation with necessary restrictions. Frequent edits make it difficult to tell whether a change came from the market, the creative or the operator.
A practical retesting checklist
- Write down the problem the inherited setting was meant to solve
- Confirm whether that constraint still exists
- Create a simpler baseline
- Change one material variable at a time
- Review delivery, conversion quality and economics together
- Allow the test to cover the buying cycle
- Document the conditions instead of turning one result into a universal rule
WEPR audits Meta account structure, creative supply, event tracking and landing-page conversion, then designs controlled tests with clear risk limits. The work removes complexity that no longer serves the business and tests each setting against commercial outcomes.
Are interest targeting and separate retargeting campaigns unusable in 2026?
No. They can remain useful under specific product, market and data conditions. Use a simpler structure as the baseline and test whether segmentation adds incremental value.
Source：Meta Performance 5 official guidance
Source：Meta Advantage+ Audience official guide