The 70/20/10 scaling budget split explained

Quick Answer

The 70/20/10 split allocates 70% of budget to proven, profitable campaigns; 20% to scaling experiments (geo expansion, new audiences, new creative angles in production); and 10% to pure experimentation (untested concepts, new ad formats, learning campaigns). It works because it forces discipline — you can't scale forever without experimenting, but you also can't experiment yourself to death without protecting the cash flow that funds it.

The Framework

1. The 70%: protect what works

70% of total spend goes to your stable, profitable campaigns. These are not where you experiment. Touch them as little as possible. The job is to maintain ROAS, not to optimise for theoretical upside.

2. The 20%: extend what works

20% goes to scaling-adjacent experiments — taking a winning ad set and trying it in a new geo, scaling a winning angle to new creator types, expanding placements. These experiments have one foot in proven territory.

3. The 10%: try what doesn't exist yet

10% goes to unvalidated bets — totally new creative angles, new formats (Reels if you've been in feed, etc.), new audiences you have no data on. Most of these will fail. That's the point.

4. Adjust the ratio for stage

Early-stage accounts (< $3k/day) often run more like 50/30/20 — more experimentation because there's less to protect. Mature accounts ($20k+/day) often run 80/15/5 — protecting the bigger base.

5. Don't let the 10% become a graveyard

The 10% bucket is supposed to recycle. New experiments enter, dead ones get killed. If you're running the same 'experiment' for 3 months, it's not an experiment anymore — it's a failing campaign you're refusing to kill.

6. Review the split monthly, not weekly

Weekly rebalancing creates churn. Monthly review gives you enough data to see what's actually working in each bucket and lets you graduate experiments out of the 10% into the 20% or even the 70%.

Real Numbers from the Field

A skincare account we managed had been operating at $7k/day with no formal budget structure. They'd swing between 'all-in on the winner' weeks and 'try everything' weeks. We forced the 70/20/10 split: $4.9k stable, $1.4k scaling experiments, $700 pure experimentation. Within 3 months their ROAS volatility halved and overall account ROAS improved 11%. The discipline of the 70% was the real win — it stopped them from constantly disrupting their best campaigns.

Frequently Asked Questions

Should the 70% include retargeting?

Yes, if retargeting is stable. Retargeting often deserves to be in the 'protect what works' bucket because it's less variable than cold.

What if my experiments are outperforming my stable campaigns?

Then graduate them. Move winners from the 10% into the 20%, and from the 20% into the 70%. The split is dynamic, not fixed to specific campaigns.

Is 10% enough for experimentation?

At $5k/day, 10% = $500/day. That's enough to run 3-5 small experiments simultaneously. At $20k/day, $2k/day is plenty. Below $3k/day, you may need more.

Who owns each bucket in a team?

Senior buyer owns the 70%. Mid-level owns the 20%. Junior or specialised testers own the 10%. This naturally lets people grow into more risk.

Does this work for B2B with longer sales cycles?

Yes, but use a longer review cycle — quarterly instead of monthly — and use lead quality not just CPL as your performance signal.

Stop Guessing What Your Ads Look Like at Scale

When you're spending £5k+ a day, you can't manually screenshot every ad on every device, in every placement, in every locale. Things break silently. Creators ship the wrong aspect ratio. Localised text overflows. Safe zones get clipped by Reels UI.

Pix-Vu lets you preview every ad creative across Facebook, Instagram and Reels placements before launch — desktop, mobile, all aspect ratios, all crop variants. Spot the broken creative before you spend £20k discovering it the hard way.

Try it free at https://pix-vu.com — built for media buyers who can't afford a single wasted impression.