6 paid-ad results, each credited and linked to the source.
All 111 budget and scaling results in this hub are marked as wins, and most show spend rising while efficiency held or improved.
Among the 21 that state a ROAS change as a percentage, the median lift is +44%; among the 24 that state a cost-per-result change, the median cut is about 24%.
Treat that with care: 72 of the 111 (65%) are vendor-reported, and only 4 are controlled split tests.
A result lands here when the main thing that changed was how much was spent, how fast it grew, or where the budget went: raising daily budgets, moving money between channels or campaigns, phasing spend around a sale, or cutting spend that wasn't adding sales.
Most members also tag a second lever, most often measurement (34), audiences (30) or campaign structure (27).
That overlap matters. Many of these stories changed budget and something else at the same time, so the budget move gets only part of the credit.
You can compare this lever with others in all results.
Every one of the 111 members is filed as a win.
There are no lost or flat results here, which says more about what gets published than about scaling.
Agencies and ad platforms rarely write case studies about budget increases that went badly.
To estimate typical size, we pulled every member that states a percentage change.
For ROAS (including incremental and Day 0 ROAS), 21 members give one: the median is +44%, ranging from +5% to +858.8%.
For cost per result (CPA, CAC, CPL, CPI), 24 members give one: the median is a 24% drop, ranging from 2.9% to 70%.
For context, see how this change ranks against others.
Evidence is mostly weak: 66 are before/after comparisons, 29 are single campaign results, 11 are platform or vendor studies, 4 are split tests and 1 is a snapshot.
The clearest practitioner examples come from X posts. @runtraffic reported a YouTube account at Spend +$682k, ROAS 0.76 → 1.51, the rare case where ROAS rose as spend climbed.
The same poster shared Demand Gen campaigns with a combined ROAS 1.63 on $183k over one month while they were still limited by budget.
@thedanielbudai described a regulated ecommerce brand going from Revenue $4.5k → $505k/month at ROAS 2–2.5x by raising budgets in 10–15% steps.
The 2–2.5x is the peak month's ROAS; the first month ran at a loss on purpose.
On the fast end, @runtraffic took a lead gen Search account to Spend 5x in 3 days, CTR −0.25 pts, but shared no lead cost.
Among agency stories, Pilothouse reported Meta conversions +430%, CPA −40% on +223% spend for Hestan after running acquisition all year ahead of Q4. These are all before/after or campaign results with no control.
Meta appears on 47 members, TikTok on 39, Google on 35 and YouTube on 20 (results can list several platforms; 5 list none).
The evidence mix differs a lot by platform:
| Platform | Results | Vendor-reported | Before/after | Campaign | Study or split |
|---|---|---|---|---|---|
| Meta | 47 | 22 | 33 | 12 | 1 |
| TikTok | 39 | 32 | 20 | 12 | 7 |
| 35 | 15 | 24 | 7 | 4 | |
| YouTube | 20 | 12 | 9 | 5 | 6 |
Meta is where independent agencies dominate, so see Meta budget and scaling tests for the less promotional set.
TikTok is mostly TikTok for Business case studies (26 of 39), many about GMV Max and Smart+, so read TikTok results knowing 32 of 39 are vendor-reported.
Google results lean toward Search and Performance Max budget moves; see Google budget and scaling tests.
YouTube has the most studies (6 of 20), many from Think with Google, in YouTube results.
There are no LinkedIn members yet; LinkedIn results covers other levers.
61 of the 111 members don't name an ad format, which fits a lever that sits above creative.
Where a format is named, video ads lead (32), followed by partnership ads (14), UGC ads (14) and catalog ads (10).
By industry, apparel (19), general ecommerce (17), beauty (13) and home (12) make up 61 members.
Lead gen, SaaS and finance together account for 9, so the evidence for scaling non-ecommerce accounts is thin.
We hand-checked the 26 members that state both a spend increase and an outcome.
In 21, the outcome grew faster than spend or efficiency held.
Examples: Barton Watch Bands (ROAS 1.49 → 2.14 with spend +248%, Janky Patel newsletter), Grüns (22X media spend, 40X revenue, vendor-reported by Northbeam) and Naked Copenhagen (revenue +92% YoY with ad spend +57%, vendor-reported by Think with Google).
If your real goal is efficiency rather than size, the wider playbook is in how to increase ROAS.
No member is marked lost, so the warnings sit inside the wins.
In 5 of the 26 hand-checked cases, output grew more slowly than spend.
Flighted's Collective client went from about $30,000 to $150,000 a week on Meta (5x) while weekly qualified leads went from 150 to 350+, so cost per qualified lead works out to roughly double (5x the spend for about 2.3x the leads).
Hillside Shilajit raised spend 356% for 263% more revenue, and Uniqa raised budget 50% for 45% more quotes (both vendor-reported).
Other wins carry a cost elsewhere: the 5x lead gen jump cut CTR by 0.25 points, a seven-figure jewelry account saw CVR −14%, and Hestan's returning customer rate fell 6%.
Laura Geller raised GMV Max spend 50%+ for a 10% total revenue lift (vendor-reported).
Scaling probably won't work when campaigns aren't limited by budget, when you already capture most of the searches in your market, or when ROAS is near break-even, because the extra spend buys pricier results.
In this hub's Meta results, several of the accounts that held efficiency consolidated campaigns first, then raised budget in steps and added fresh creative.
Barton Watch Bands went from ROAS 1.49 to 2.14 while spend rose 248%. These are before/after results, not controlled tests.
None of the 111 results test step sizes directly.
One Google Ads result scaled with 10–15% budget steps, and a Search account jumped 5x in 3 days with a 0.25-point CTR drop.
Smaller steps are the safer default, then judge each step on what the extra spend bought.
The Google members here mostly raised budget on campaigns that were limited by budget, or moved spend from brand to non-brand and Performance Max.
Of 35 Google results, 24 are before/after comparisons and 15 are vendor-reported, so test on your own account.
Not always. In 21 of 26 members that report both spend and outcome changes, the outcome grew as fast or faster than spend.
In the other 5, it grew more slowly, and these are published wins, so failed scale-ups are under-represented.