---
title: When and how fast to scale Meta ads — scaling criteria and the 80/20 budget split
slug: meta-ads-scaling-rules-and-budget-split
summary: Smart Marketer's 2025 scaling rules (scale at CPA 50% below target for 7 days, +20% every 3 days, watch weekly frequency >2), the 80/20 prospecting/retargeting split with new-vs-existing exclusions, and the Pretty Farm Girl case (spend doubled, revenue +65% in 60 days).
domain: marketing
tags: [ecommerce, meta-ads, scaling, budget, media-buying, case-study]
related: [meta-ads-creative-testing-engine, evergreen-discount-collection-page, ecommerce-ad-creative-metrics, meta-ads-budget-math-for-apps, meta-scaling-winners]
sources: [yt:GR8JS8sQiL8]
updated: 2026-10-24
---

# When and how fast to scale Meta ads

## TL;DR
- **Scale when CPA is ≥50% below target for 7 days.**[^1]
- **Raise budget ~20% every 3 days** — doubling usually spikes CPA.[^1]
- **Weekly frequency above ~2** → consider throttling/pausing to avoid wearing out audiences.[^2]
- Budget split: **~80% prospecting** (excluding people who engaged/bought), **~20% retargeting** — up from the 60/40 Smart Marketer used to recommend.[^3]
- **Split new vs existing customers with exclusions** — Advantage+/CBO "prospecting" often mostly retargets existing customers.[^4]
- Set goals from **back-end customer value** (what a customer is worth over the first year), not in-platform ROAS.[^5][^6]

## Details
**Who.** [[people/molly-pittman]] (Smart Marketer agency), July 2025, sharing the criteria from a media-buyer call "last week"; she notes they've scaled ~20% per 3 days since she started Meta ads in 2012.[^1]

**Pretty Farm Girl case** (family farm brand selling tallow skincare, riding a hot trend): in 60 days the agency **doubled ad spend** and grew **revenue ~65%** (Molly also says 60%).[^7][^8] In order, per media buyer Dominique:[^4][^9]
1. *Funnel fix* — separate new and existing customers with proper exclusions, so you know whether you're actually acquiring.
2. *Creative overhaul* — fresh formats; the founder was strong on camera but they relied on her too much.
3. *Message testing* — a product almost every woman could use is a trap: find pockets by age, interests, skin type, daily life and write to each.
4. *Budget* — they were judging by in-Facebook ROAS; working out monthly customer value let them set more aggressive CPA goals.
5. *Landing pages* — moved from product pages to homepage and evergreen discount collection pages ([[evergreen-discount-collection-page]]).
They also weren't spending 80% on new customers and didn't know it.[^10]

**When 80/20 shifts.** Brands with big organic traffic (SEO, organic social, TV, famous founder) have large retargeting pools and may need more bottom-funnel spend.[^3]

**Before the next scale push**, have new creative variants ready so the algorithm has fresh inputs.[^2]

## Disagreements / caveats
- Molly says every client has nuances; these are baselines for "e-com, non-e-com".[^11]
- The revenue lift coincided with a trending category (tallow) — not all of it is attributable to the account changes.
- App-UA notes in this base ([[meta-ads-budget-math-for-apps]]) size budgets by the ~50 events/week learning threshold; Smart Marketer's rules are CPA-headroom-based. Compatible, different lens.

## See also
- [[meta-ads-creative-testing-engine]] — producing the winners you scale

[^1]: yt:GR8JS8sQiL8 @ 37:16 — Molly Pittman
[^2]: yt:GR8JS8sQiL8 @ 37:16 — Molly Pittman
[^3]: yt:GR8JS8sQiL8 @ 30:03 — Molly Pittman
[^4]: yt:GR8JS8sQiL8 @ 12:56 — Molly Pittman
[^5]: yt:GR8JS8sQiL8 @ 15:36 — Molly Pittman
[^6]: yt:GR8JS8sQiL8 @ 31:15 — Molly Pittman
[^7]: yt:GR8JS8sQiL8 @ 12:03 — Molly Pittman
[^8]: yt:GR8JS8sQiL8 @ 06:31 — Molly Pittman
[^9]: yt:GR8JS8sQiL8 @ 14:35 — Molly Pittman
[^10]: yt:GR8JS8sQiL8 @ 31:15 — Molly Pittman
[^11]: yt:GR8JS8sQiL8 @ 38:50 — Molly Pittman