I relaunched a proven Meta campaign with the same creatives and audience. CPI jumped from around $1 to as high as $8. Here’s what I changed and what happened next.

Meta Paid Marketing | Why CPI Can Change Even with the Same Creatives, Audience, and Budget

One thing I learned from running Meta app install campaigns is that using the exact same creatives, audience, and campaign structure does not guarantee the same CPI.

Sometimes, even when nothing seems to have changed, performance can look completely different.

Here is one example.

1. The Same Setup, but a Different CPI

About a week ago, I received an email saying that my Meta Ads account had been restricted.

I never found out why. After searching Reddit, I noticed that many people had experienced similar issues. Most of them also said that Meta never gave them a clear explanation.

Since there was not much I could do, I created a new ad account and launched campaigns using the same creatives and the same ad set structure.

However, the campaigns did not optimize properly during the first week.

These were not new or untested creatives. They had already been classified as winning creatives in my previous account, consistently delivering a CPI of around $1.xx.

After launching them in the new account, however, the results were completely different.

• Most ad sets showed a CPI of around $3.xx
• One creative even reached a CPI of $8.xx

I could not simply leave the campaigns running like this and hope that performance would eventually improve.

So I combined all the creatives into one new ad set. I also merged the separate $30 daily budgets into a single $60 daily budget and restarted the learning process.

After that, CPI gradually stabilized at around $1.xx–$2.xx.

This experience showed me that even with the same creatives and audience, results can change depending on the ad account, campaign structure, and learning process.

2. What My Winning Creatives Have in Common

My definition of a winning creative is fairly simple.

A winning creative:

• Generates a high number of installs
• Achieves a ROAS of 200% or higher

These are the creatives that are worth keeping and potentially scaling.

3. Why Is Meta Spending Only Half of My Budget?

Updated on Feb 24

At one point, my campaign had the following results:

Metric	Result
Daily budget	$60
Actual spend	Around $30
Installs	18
CPI	$1.98
CPM	$4.57
CTR	1.01%
Frequency	1.08
Bid cap	None
Optimization goal	App Installs

My CPI was $1.98, but I had only generated 18 installs, and Meta had spent only around half of the daily budget.

My first thought was:

If Meta spent more, would it not bring in more new active users?

Looking at the metrics, there was no obvious delivery problem.

• A CPM of $4.57 was very low, which suggested that the campaign was not losing heavily in the auction.
• A CTR of 1.01% was reasonable.
• A frequency of 1.08 did not indicate audience fatigue.
• There was no bid cap limiting delivery.
• The campaign was optimized for app installs.

This suggested that the campaign was not failing to spend because of high competition.

The more likely issue was conversion volume.

Meta app campaigns generally need enough conversion data to optimize confidently. At that point, the campaign had generated only 18 installs, so the algorithm may have been delivering conservatively.

Based on the CPI, spending the entire budget should have produced approximately:

$60 ÷ $1.98 ≈ 30 installs

Based on CPM, the estimated spend was:

9,398 impressions ÷ 1,000 × $4.57 ≈ $42.94

The numbers suggested that Meta could theoretically spend more, but the algorithm was choosing not to use the full budget.

Possible Solutions

1. Reduce the Budget

Meta tends to spend more efficiently when the budget matches the available conversion volume.

Instead of maintaining a $60 daily budget, reducing it to around $40 could create a more appropriate balance between budget and conversion density.

2. Add New Creatives

The low install volume might not have been caused by the audience itself.

It was possible that the existing creatives had already reached most of the users who were likely to respond to them.

In that case, increasing the budget would not solve the problem. Adding new creatives could give Meta more opportunities to find new responsive users.

3. Test Broader Targeting

An audience size of 3.6 million should be large enough.

However, for app install campaigns, broad targeting combined with Meta’s automated targeting can sometimes perform better than manually narrowed audiences.

Action

My next action was simple:

• Create new creatives
• Upload them to the campaign
• Monitor whether install volume and budget utilization improve

4. Spending More Does Not Always Improve Performance

Another lesson I learned is that a larger budget does not always lead to better results.

One of my campaigns had been generating a CPI below $1, with a daily budget of around $30.

Then, because of my mistake, the daily budget suddenly increased to $200.

After that, CPI increased to around $1.09.

The result was still reasonable, but it was worse than before.

My assumption is that Meta had to spend the much larger budget within a single day, but the available audience was not large or responsive enough to absorb that increase efficiently.

As a result, the campaign may have expanded beyond the most relevant users and started reaching people who were less likely to install.

This experience reminded me that budget is not simply about how much money I am willing to spend.

It is also about whether the available audience and creative performance can support that level of spending.

A campaign needs a budget that allows Meta to reach the right users without expanding delivery too aggressively.

Finding the right budget can be just as important as finding the right creative.

** IG @zzangsonga**