Most explanations of Meta's advertising system begin and end with Andromeda. That is now incomplete. Meta replaced three layers of its ad system in roughly twelve months — the retrieval engine, the shared knowledge model, and the real-time ranking process. Each change shifts more of the matching work away from the targeting panel and toward the creative itself.
Three systems, one shift.
These are Meta-reported platform averages, not performance guarantees for an individual advertiser.
Four checks before you build your next campaign.
If one of these is weak, fix it before spending.
Problem diversity
Do your ads address different homeowner problems, or the same problem reworded?
Good: sold-price report + renovation guide + buyer-demand updateAudience as suggestion
Are you treating location as a constraint and leaving the rest to Meta?
Good: suburb targeting + Advantage+ expansionCreative as signal
Does each ad contain enough specific information for the algorithm to match it?
Good: "Four-bedroom Baldivis homes — see recent sales"Evidence over assertion
Does each ad demonstrate local knowledge with verifiable proof?
Good: three recent sales with dates and typesStop over-targeting.
The old approach was to build a precise audience: an age range, several property interests, a postcode and an income proxy, then show everyone in that audience the same appraisal ad. The new system makes that approach counterproductive.
Advantage+ audience can use your inputs as suggestions and search more broadly when it predicts better performance. Advantage+ is now the default for eligible leads campaigns.
Give Meta different problems, not different colours.
Changing the background colour on an appraisal graphic does not create another targeting option. Changing the homeowner's problem does.
These ads support the same commercial objective — more listings. They approach it through different homeowner situations.
Track what happens after the lead arrives.
Meta can only optimise toward the outcomes it receives. If the only signal is "form submitted," the system searches for more people likely to submit forms — not more people likely to become listings.
If your CRM can send downstream outcomes back to Meta through the Conversions API for CRM, the system can optimise toward leads more likely to become qualified. Meta reports advertisers using this setup saw an average 15% reduction in cost per quality lead.
Stop fighting the algorithm with over-targeting.
The algorithm moved strategy out of the targeting panel and into the ad itself. Blockwise helps turn distinct seller propositions into on-brand creative, prepare the campaign and lead form, and bring Meta leads into a clearer review path.
The practical details.
Do I still need to set a location target?
Yes. Your geographic service area is a genuine business constraint. What you should remove is the layering of speculative interests inside that location. Let the creative identify the homeowner situation; let the location setting define where you can operate.
Does this mean I should run 50 ads at once?
No. A local agent spending a modest daily budget should start with six to twelve genuinely distinct assets — different problems, different proofs, different offers — and let the system learn which messages create real conversations.
Is "creative is the new targeting" just a marketing phrase?
It reflects a real architectural change. Andromeda retrieves ads by reading creative signals. Advantage+ audience treats your audience inputs as suggestions. The content of your ad now plays a larger role in determining who responds. Audience settings have not disappeared — they have been demoted.
Should I select the Housing Special Ad Category for real estate ads?
Yes, when the campaign relates to housing. Meta requires advertisers to select the relevant Special Ad Category, and failing to do so may result in ad rejection. The algorithm changes do not remove compliance obligations.