--- title: Broad targeting vs interests, lookalikes and retargeting on Meta (2023 → 2026) slug: meta-targeting-broad-vs-interests summary: How targeting advice flipped from Smart Marketer's 2023 70/15/15 interest/lookalike/broad split to 2025–2026 broad-by-default — and the remaining cases for interests (forcing a new avatar, lead gen), geo-only (local), customer exclusions, and Advantage+ as audience research. domain: marketing tags: [meta-ads, targeting, broad-targeting, lookalikes, retargeting, advantage-plus] related: [meta-value-rules, meta-audience-segments-and-remarketing, meta-andromeda-gem-lattice, meta-account-structure-2026, meta-persona-angle-creative-dimensions, meta-consolidation-vs-sprint-approach] sources: [yt:x1b4h42HTG8, yt:BIVnOWcGnfA, yt:kn4ByanAe-s, yt:cnWpC1g5qek, yt:aNCUFwQw524, yt:Ha0gOItSZaU, yt:tcQ_1mm_bHU, yt:C0R5YJiWpig, yt:du6rDgRf06g, yt:sjd711834hc] updated: 2026-10-24 --- # Broad vs interests, lookalikes and retargeting ## TL;DR - **2025–2026 default: broad** (no detailed targeting), automatic placements, optimise for the purchase — creative does the targeting.[^1][^2][^3] - Dara Denney: Andromeda/GEM made interests and even prospecting-vs-retargeting splits "pretty much useless"; she runs no retargeting on most brands even at $500K–1M/month.[^4][^17] - **Exclude existing customers** from top-of-funnel campaigns, or Meta may quietly retarget your warm base.[^6] - Interests still have two uses (Molly Pittman, 2026): **forcing Meta toward a new avatar** (e.g. a bodybuilder supplement moving into older women doing yoga) and big **lead-gen launches** (a planned ~$50K lead-gen push targeting marketing tools/gurus).[^7][^8] - Local businesses: **geo radius only** (her restaurant: 50 miles), no interests.[^7] - **Detailed targeting and lookalikes are only suggestions** for most performance goals now; they "can't hurt" but you can't prove they help (Jon Loomer). Restrict age/gender/placements only for age-restricted goods — otherwise use **value rules** for proven problems ([[meta-value-rules]]).[^18][^19] - Remarketing happens algorithmically: define audience segments and Meta's breakdown typically shows 10–30%+ of budget already reaching warm people ([[meta-audience-segments-and-remarketing]]).[^20] ## Details **Why broad works now.** Meta's unified models can match the right ad to the right user, and Meta reads creative, copy and Page signals to find buyers — "creative is the new targeting".[^9][^10] Josh Brampton's caveat: a niche-looking ad still only reaches the niche that consumes that content; to reach other buyers you need ads that look like *their* feed.[^11] **The 2023 baseline (for contrast).** Smart Marketer's 70/15/15 rule: ~70% of top-of-funnel budget on **interest** clusters (aggregated to 5–20M audience for tests, 15M+ for scaling), 15% lookalikes, 15% broad. Buyers spent ~30% of their day on off-platform interest research; accounts ran 26–95 ad sets differing only by interests. Lookalikes were coming back: for Ella + Mila a 1% LAL got 100 customers at $37 CPA vs 5–7% LAL 331 customers at $34 — larger LALs need 7–14 days. Best LAL seeds: purchasers, email subscribers, top-20% customers by spend.[^12][^13][^14][^15] **Advantage+ as research.** In 2023 an Advantage+ campaign revealed many buyers over 40 for a product the team assumed was young-skewing; new campaigns for that avatar took the client from $10K to $15K/day.[^16] **Low-budget rule (Ben Heath, Loomer).** At ~$100/month: broad, at most interests as suggestions, no hard constraints, no warm audiences — a small budget can't recover from a too-narrow bet. Loomer's low-volume recipe: target by **location (country) only**, keep all placements.[^21][^22] **2020 baseline.** Savannah Sanchez's 2020 ecommerce structure tested on a 1% purchaser lookalike plus broad-interest audiences (min. 2M) with interest expansion on, then scaled winners across 5% lookalikes and interest stacks — the era before broad-by-default.[^23] ## Disagreements / caveats - Molly in 2023 ("interest targeting is still where the money is") vs Molly in 2026 ("broad… no detailed targeting… hurts me in the chest") — the platform changed, so treat the 2023 split as obsolete except for the avatar-forcing use.[^12][^3] - **Retargeting:** Dara runs none; Smart Marketer's Andromeda layout keeps one win-back campaign (see [[meta-account-structure-2026]]). - Dara's "no awareness/engagement campaigns" has an exception for brands at multiple seven figures/month, where Meta lift tests showed delayed conversions.[^5] ## See also - [[meta-persona-angle-creative-dimensions]] — targeting through creative - [[meta-andromeda-gem-lattice]] - People: [[people/molly-pittman]], [[people/dara-denney]] [^1]: yt:kn4ByanAe-s @ 18:10 — Molly Pittman [^2]: yt:kn4ByanAe-s @ 19:26 — Molly Pittman [^3]: yt:kn4ByanAe-s @ 27:07 — Molly Pittman [^4]: yt:x1b4h42HTG8 @ 00:00 — Dara Denney [^5]: yt:BIVnOWcGnfA @ 03:04 — Dara Denney [^6]: yt:kn4ByanAe-s @ 27:58 — Molly Pittman [^7]: yt:kn4ByanAe-s @ 1:05:04 — Molly Pittman [^8]: yt:kn4ByanAe-s @ 1:06:06 — Molly Pittman [^9]: yt:x1b4h42HTG8 @ 05:44 — Dara Denney [^10]: yt:kn4ByanAe-s @ 08:50 — Molly Pittman [^11]: yt:aNCUFwQw524 @ 03:05 — Josh Brampton [^12]: yt:cnWpC1g5qek @ 24:05 — Molly Pittman [^13]: yt:cnWpC1g5qek @ 27:20 — Molly Pittman [^14]: yt:cnWpC1g5qek @ 28:52 — Molly Pittman [^15]: yt:cnWpC1g5qek @ 29:44 — Molly Pittman [^16]: yt:cnWpC1g5qek @ 41:55 — Molly Pittman [^17]: yt:BIVnOWcGnfA @ 01:33 — Dara Denney [^18]: yt:Ha0gOItSZaU @ 01:33 — Jon Loomer [^19]: yt:tcQ_1mm_bHU @ 04:02 — Jon Loomer [^20]: yt:Ha0gOItSZaU @ 03:56 — Jon Loomer [^21]: yt:du6rDgRf06g @ 02:22 — Ben Heath [^22]: yt:C0R5YJiWpig @ 04:56 — Jon Loomer [^23]: yt:sjd711834hc @ 04:44 — Savannah Sanchez