/55-scaling-ads-global

name: 55-scaling-ads-global
description: "Use when the user has a winning ad and wants to scale it safely: readiness checks, vertical scaling at +20-30% per step, horizontal scaling into new audiences, creatives, and channels, learning-phase reset risk, how Advantage+ and Performance Max behave when scaled, and the signals that mean stop. Trigger on 'scale ads', 'increase ad budget', 'scale campaign', 'scale a winner', 'vertical scaling', 'horizontal scaling', or 'CPA went up after scaling'."
metadata:
  version: 1.0.0
  category: performance
license: MIT
triggers:
  - "scale ads"
  - "increase ad budget"
  - "scale campaign"
  - "scale a winner"
  - "vertical scaling"
  - "horizontal scaling"
  - "CPA went up after scaling"
  - "how fast can I scale"
output: "File .md — scaling roadmap by week, budget increase rules, pre-scale checklist, stop signals, and a creative refresh schedule"
related:
  - product-marketing-context-global
  - 19-ab-test-setup-global
  - 21-ads-audit-global
  - 03-performance-eval-global
  - 56-retargeting-plan-global
  - 52-account-structure-global
  - 57-next-ads-plan-global

Scaling Ads (Global)

Scaling means increasing spend while holding or improving efficiency. Scaling badly means CPA spikes and budget burns. **Hard rule: never increase budget while CPA is bad.** Fix the problem first (run `21-ads-audit-global`), then scale.

Information gathering

Ask up to 4 questions:

1. **Is there a proven winner?** How many consecutive days has CPA been stable? At least 7 are needed. 2. **What are the current numbers?** CPA, ROAS, frequency, spend, and reach against total audience size. 3. **Current daily budget and target daily budget?** 4. **Which platform, and how many spare creatives exist?** Never scale on a single creative.

Principles

1. **Bad CPA means no scaling.** Adding budget to a broken campaign accelerates the loss. 2. **+20-30% per step, never double overnight.** Sudden jumps reset the learning phase and CPA spikes. The safest cadence is +20% every 72 hours. 3. **Monitor 24-48 hours after every increase** before the next one. 4. **Do not edit a winning ad set while scaling.** Any significant change — budget shock, audience edit, creative swap inside the ad set — can reset learning. 5. **Always hold a creative reserve.** At least 2-3 fresh creatives ready before scaling. Scaling accelerates fatigue. 6. **Vertical first, horizontal in parallel.** Raise the winner's budget while duplicating audience and creative so the account never depends on a single point. 7. **Scaling raises cost per result even when done right.** Broader delivery means less efficient impressions. Decide in advance how much CPA erosion is acceptable at the higher volume.

Workflow

1. Readiness check — all must pass

• [ ] Running stably for at least 7 days
• [ ] CPA at or below target on at least 5 of the last 7 days
• [ ] Frequency below 2.5 (audience not saturated)
• [ ] Reach below 50% of the addressable audience (room left to grow)
• [ ] Not currently in the learning phase
• [ ] Tracking verified correct — never scale on bad data (`53-tracking-setup-global`)
• [ ] At least 2 fresh creatives in reserve

If any item fails, do not scale. Go back to optimizing or preparing.

2. Vertical scaling — raise the budget

| Mode | Increase | Cadence | Condition | |------|---------:|---------|-----------| | Conservative | +20% | Every 72h | CPA holding at target | | Moderate | +30% | Every 2-3 days | CPA clearly better than baseline | | Aggressive | +50% | Once only | Very stable campaign, risk accepted | | Never | 2x overnight | — | Guarantees a learning reset and CPA spike |

After each increase, monitor for 24-48 hours:

• CPA up by 20% or less: hold, let it stabilize, then increase again.
• CPA up more than 20%: HOLD, do not increase further.
• CPA up more than 40%: revert to the previous budget and wait 48 hours for stability.

3. Horizontal scaling — expand

1. **Duplicate the winning ad set into a new audience.** Keep the creative identical. Test the same demographic against a different interest theme, lookalike 1% -> 3% -> 5%, or broad. Exclude the original audience to avoid overlap. 2. **New creative on the winning audience.** Keep the audience identical, add 2-3 variants changing hook or format. Run these in the testing campaign, not inside the winning ad set. 3. **New channel with the winning message.** Port the proven hook from Meta to TikTok or YouTube (or the reverse). Keep the core message, change the execution. 4. **New market.** The cheapest volume expansion is often a second country rather than a wider audience at home. Check the cost baseline first: per `references/benchmarks-global.md`, a Tier 2 market can cost a fraction of Tier 1 CPM, but conversion value and AOV usually drop too. Re-run `10-reverse-kpi-global` for the new market before committing budget.

4. Learning-phase reset risk

| Action | Resets learning? | Safer way | |--------|------------------|-----------| | Budget increase up to 20-30% | Usually not | Step up gradually, wait 72h | | Budget increase above 30-50% | Likely | Break into several smaller steps | | Change audience, placement, or bid strategy | Yes | Duplicate into a new ad set instead of editing | | Add or remove creative inside the ad set | Possibly | Add in a new ad set or the testing campaign | | Pause and restart after a long gap | Possibly | Avoid pausing winners; throttle with budget instead |

5. Automated campaign types behave differently

Advantage+ (Meta) and Performance Max (Google) do not respond to scaling the way manual campaigns do:

• **They dislike frequent budget edits.** Large or repeated changes push them back into a learning state. Same +20-30% discipline applies, with a longer settling window.
• **They need signal, not structure.** When performance drops during scaling, the fix is usually more or better creative assets and cleaner conversion data, not more ad sets.
• **They cannibalize ma