The account is working well. You double the budget. A week later ROAS has halved.
This is a wall almost every brand hits. And it usually gets read as “the platform broke”. What broke isn’t the platform — it’s the way the increase was made.
What is happening?
Three things at once.
1. The learning phase restarts
When you change an ad set’s budget significantly, Meta’s algorithm re-enters learning. Through that period delivery is unstable and costs rise.
A 20% increase usually doesn’t trigger it. A 100% increase definitely does.
2. The cheap part of the audience runs out
The platform always starts with the person most likely to convert. As the budget grows it has to reach less willing people.
That is more expensive by definition. Within the same audience, the people you reach at 3,000₺ a day and the ones you reach at 6,000₺ are not of the same quality.
3. The creative pool isn’t big enough
As budget rises, frequency climbs fast. Two creatives can spend 3,000₺; they cannot spend 6,000₺ — the same people see the same ad twice as often and fatigue.
The right rate of increase
The rule that works in practice:
- At ad set level: 20-25% every 3-4 days
- At campaign level (CBO): up to 30% every 3 days
- If you need a jump: instead of raising budget, duplicate into new ad sets
If you want to double, that’s roughly a two-week job. Done in a single day, you lose two weeks anyway.
What must be ready before the budget
Three things should be in place before you raise budget:
1. A creative pool. If the budget is doubling, you need at least twice the creative. The rule is crude but useful: one active winning creative for every 1,500-2,000₺ of daily spend.
2. An audience that can widen. You can’t scale on a narrow interest-based targeting. Targeting needs widening before the budget rises — broad targeting plus strong creative generally scales better than narrow targeting plus weak creative.
3. Operations. Can you ship if orders double? Is there enough stock? Will customer service hold? Growing the advertising and then delaying delivery comes back as returns and bad reviews.
Horizontal scaling
Vertical scaling (raising the same ad set’s budget) jams at some point. Horizontal scaling is usually more stable:
- Opening the same creative to new audiences
- Feeding the same audience new creatives
- Adding new placements (Reels, Advantage+ placements)
- Adding a new platform (Google or TikTok if Meta has jammed)
With horizontal scaling the learning phase stays confined to the new structure and the working structure isn’t touched.
How much of a drop is normal?
A degree of ROAS decline while scaling is normal and acceptable — as long as contribution rises.
An example:
| Before | After | |
|---|---|---|
| Spend | 100,000₺ | 180,000₺ |
| ROAS | 4.0 | 3.2 |
| Revenue | 400,000₺ | 576,000₺ |
| Margin (35%) | 140,000₺ | 201,600₺ |
| Contribution | 40,000₺ | 21,600₺ |
Here the ROAS fall is too steep: revenue rose but the money left over shrank. That scaling failed.
In the same table, if ROAS had held at 3.6 the contribution would be 46,800₺ — that is a success.
So the limit on scaling is set not by ROAS but by your break-even point. Keep growing while contribution keeps rising; stop when it starts falling.
Summary
- Raise 20-25% every 3-4 days; don’t jump
- Grow the creative pool before the budget
- When vertical jams, scale horizontally
- Decide on contribution, not on ROAS
- Make sure operations can take it