Blog · Tracking & Analytics

How Long to Run a Marketing Test Before You Decide

Stop killing your ads too early. Learn how to calculate the right marketing test sample size for your salon to get reliable data and more bookings.

iGrowSalons · 2027-02-09

How Long to Run a Marketing Test Before You Decide

You launch a new ad for your salon on Monday. By Wednesday, it’s only brought in one lead. You panic, turn it off, and declare the offer a failure. Then you build another one, and the cycle repeats. You’re making decisions based on noise, not data.

I don't cut hair. I run ads and build funnels for salons, and I've watched a few hundred of them try to get new clients. The most common mistake I see isn't bad ad creative or a terrible offer. It's pulling the plug before the test actually means anything. You are starving the algorithm and making emotional choices.

To know if a campaign actually works, you need statistical significance. In plain English, that means you need a large enough marketing test sample size to prove the result wasn't just a fluke. Today, I'll show you exactly how to calculate that number so you can stop wasting money on premature decisions.

Why Three Days Isn't a Test

When you launch a campaign on Meta or Google, the platform goes into a learning phase. It is actively testing your ad against different pockets of the audience to see who responds. If you kill the ad in 72 hours, you didn't run a test. You just interrupted the machine while it was still figuring things out.

Meta specifically defines creative fatigue as an audience seeing the same creative too many times, causing engagement to drop and costs to rise [1]. But on day three, fatigue isn't your problem. Your problem is a lack of data. You haven't reached a valid marketing test sample size.

Consider this: the median landing page conversion rate is 6.6% across industries, and 10%+ is considered good [2]. If you send 20 clicks to your landing page and get zero bookings, you haven't proven the page is broken. At a 10% conversion rate, 20 clicks should yield two bookings. Getting zero could just be a bad afternoon. You need more volume to know the truth.

The Math Behind a Valid Test

You don't need a degree in statistics to figure this out. You just need to understand how much data is required to make a confident decision. Let's look at a worked example for a typical salon campaign.

Suppose you are running a $10/day ad test to find your winning offer, like we discuss in The $10/Day Ad Test That Finds Your Winning Offer. You are testing two different images.

Here is the formula we use to determine if we have enough data:

Target Leads Needed = 100 / Expected Conversion Rate

If you expect a 10% conversion rate on your landing page, you need at least 1,000 clicks (or impressions, depending on what you are testing) to get 100 leads. Why 100? Because in local lead generation, 100 conversions is the baseline for statistical confidence. Anything less, and a single lucky booking skews your entire percentage.

Worked Example: The Balayage Campaign

Let's say you are running a balayage special. Here are your numbers:

• Ad Spend: $30/day
• Cost Per Click (CPC): $1.50
• Daily Clicks: 20
• Target Conversions for Significance: 50 (a lower threshold for smaller local budgets)

If you need 50 conversions to know if the ad is a winner, and your landing page converts at 10%, you need 500 clicks. At 20 clicks a day, it will take you 25 days to reach a valid marketing test sample size.

If you kill that ad on day five, you only have 100 clicks and maybe 10 leads. That is not enough data to declare a winner or a loser.

How to Set Your Testing Timeline

Here is the exact process we use to structure tests for our done-for-you clients. Follow this checklist before you launch your next campaign.

1. Define the metric you are testing. Are you testing the ad creative (Click-Through Rate) or the landing page offer (Conversion Rate)? You can only test one at a time. If you change both, you won't know what caused the result. For more on split testing, read Split Testing a Salon Landing Page Properly.
2. Determine your baseline. Look at your historical data. What is your average cost per lead? What is your average landing page conversion rate? If you don't know, assume a 5% conversion rate to be safe.
3. Calculate the required volume. Use the formula above. If you need 50 leads to make a decision, and you convert at 5%, you need 1,000 clicks.
4. Divide by your daily budget. If clicks cost $2, you need $2,000 in ad spend to get 1,000 clicks. If your budget is $50/day, the test will take 40 days.
5. Set a hard rule and don't touch it. Write the end date on your calendar. Do not log into Ads Manager and turn it off because you had a slow Tuesday. Let the test run until you hit the required volume.

Marketing Test Sample Size Benchmarks

Use this table as a quick reference for how much data you need before making a decision.

If you are struggling to decide between a lead form and a landing page, check out our breakdown in Lead Form Ads vs. Landing Pages for Salons.

When to Actually Kill a Campaign

There is a difference between being patient and being foolish. While you need a proper marketing test sample size, you shouldn't burn cash on a campaign that is clearly broken from day one.

We have a specific framework for this, which we cover in detail in When to Kill a Salon Marketing Campaign. But here is the short version:

Kill the campaign early IF:

• You have spent 3x your target Cost Per Acquisition (CPA) and have zero bookings.
• Your Click-Through Rate is below 0.5% after 2,000 impressions.
• Your landing page conversion rate is below 2% after 200 clicks.

If the numbers are borderline, let it run. The algorithm often finds cheaper conversions later in the week.

Quick reference

Save this checklist before you launch your next marketing test:

• Identify the single variable you are testing (creative, copy, or offer).
• Calculate the required number of clicks or impressions based on your baseline metrics.
• Determine how many days the test will take based on your daily budget.
• Commit to the timeline and do not pause the ad early unless it hits the emergency kill criteria.
• Use the Sample Size Calculator to remove the guesswork.

Want to make this math effortless? Grab the Sample Size Calculator — it's the exact template we hand to done-for-you clients. It tells you exactly how many days to run your ads based on your specific budget.

FAQ

How long should I run a Facebook ad before changing it?

You should run a Facebook ad for at least 7 days before making any changes. This allows the platform to exit the learning phase and gives you a mix of weekday and weekend data. Making changes earlier disrupts the algorithm and invalidates your test.

What is a good sample size for a marketing test?

For local salon marketing, aim for at least 1,000 impressions to test ad creative, or 300 clicks to test a landing page. This provides enough data to ensure your results are statistically significant rather than just random luck.

Why did my Facebook ad stop getting leads?

Ads often stop getting leads due to creative fatigue, where your audience has seen the same image too many times. Meta notes this causes engagement to drop and costs to rise. You need to rotate in new creative while maintaining a valid marketing test sample size.

How much should I spend on a Facebook ad test?

A common approach is the $10 to $20 per day test. The total spend depends on how quickly you want to reach statistical significance. If you need 300 clicks and clicks cost $1, you need to spend $300 total to complete the test.

Make Data-Driven Decisions

Stop guessing and stop killing your ads too early. Use the math to dictate your decisions. Grab the Sample Size Calculator — it's the exact template we hand to done-for-you clients. Plug in your budget, and it will tell you exactly how long to wait before you judge your next campaign.

References

[1] https://www.facebook.com/business/help/1346816142327858 [2] https://unbounce.com/average-conversion-rates-landing-pages/