Skip to content
Phil Graham Digital
  • Home
  • Podcast
  • AI Consulting
  • Facebook Ads
  • About
  • Testimonials
  • Contact
Phil Graham Digital

EP 495: Are You Testing Too Many Things at Once in Your Facebook Ads?

Next Level Social Media Ads Podcast
Next Level Social Media Ads Podcast
EP 495: Are You Testing Too Many Things at Once in Your Facebook Ads?
Loading
00:00 /
Apple Podcasts Google Podcasts Spotify iHeartRadio
RSS Feed
Share
Link
Embed

Subscribe: Apple Podcasts | Google Podcasts | Spotify | iHeartRadio

Are you testing Facebook ads, or just changing everything at once? In this episode, learn how to make your tests more useful, focus on the biggest variables, and stop creating results you cannot actually learn from.

This podcast is not affiliated with, endorsed by, or sponsored by Meta, Facebook, or Instagram. All trademarks belong to their respective owners. Opinions are based on Philโ€™s experience working with advertisers.

Show Notes:

๐Ÿงช Why running a lot of ads is not the same as actually testing them
๐Ÿ“Š How large ad sets can leave many creatives with too little reach to learn anything useful
โš ๏ธ The problem with changing too many variables at once and then not knowing what caused the result
๐ŸŽฏ Why a useful test should answer a clear question instead of just producing more versions
๐Ÿง  Philโ€™s take on testing one variable at a time, or two at the absolute most
๐Ÿช Why bigger variables like the offer, angle, image, or video are often strong places to start
๐Ÿ—‚๏ธ How separate ad sets can help give each test the reach needed for a cleaner comparison
๐Ÿ” Why knowing what happened matters less than understanding why it happened
๐Ÿ“ How tracking past tests can help advertisers spot patterns and make smarter decisions
โœ… Questions to ask before testing: What am I trying to learn, and will this setup actually tell me what caused the result?
โœ… Reminder: every business is unique, results vary, and there are no guaranteed outcomes

Transcript Disclaimer

This transcript was generated automatically using AI transcription tools and may include errors, omissions, or wording that differs from what was intended. I have not reviewed every word, and no part of this transcript should be interpreted as an income claim, success guarantee, or promise of results of any kind. The information is for educational purposes only.

Transcript:

Speaker 1:
Welcome to the Next Level Facebook Ads Podcast with Phil Graham. We help you master Facebook ads and give you an unfair advantage over your competition. Are you ready? Let’s go.
Speaker 2:
What’s up, podcast family? Welcome to the Next Level Facebook Ads Podcast episode 495. I’m Phil Graham.
We are literally five episodes away from 500 episodes every single week for over like nine years. How crazy is that? Well, I am so glad you’re here.
Welcome to November, 2026. It’s November 3rd. And I’ve got a great show for you guys today because I’m going to be talking about whether you might be testing too many things at once in your Facebook and Instagram ads and how that could hurt you and what you might want to be doing instead.
And make sure to check out next week’s episode, because we’re going to do a couple of weeks of optimizing and preparing for Black Friday and holiday ads through the holiday season because we are just about there, my friends. Before we jump into the content, though, if you need any help with your ads or your AI or your AI ads, whether that’s one-on-one coaching or you want somebody to run them for you so you don’t have to worry about it. For either one of those things, just go to my website, PhilGrahamdigital.com, or shoot me a DM on Instagram at PhilGrahamdigital, and we can set up a short Zoom call, and I would love to connect with you.
So here’s something really important. A lot of advertisers say they’re testing their ads, but a lot of times what they’re doing is they’re just running a lot of ads, but you should not confuse running a lot of ads or high ad volume with testing or learning. Time and time and time again, I’ll see a campaign with an ad set that’s got like 50 ads in there.
And when I look, it’s a complete mess. Five of the ads have a lot of reach, you know, 20 of them have a little bit of reach and like what’s the other 25 have like almost no reach. And you think you’re testing 50 ads.
You’re not. You might be testing five and you think you’re testing 50. Other times, even if you have a couple ads, maybe you’re doing it to where you just have a couple ads in there. That could be good. However, if you’re testing two ads against each other, another huge mistake I see advertisers make is changing five different things at once.
And that’s a big problem too, because no matter what happens, you will not know why if you changed so many things, if there’s so many variables. So let’s talk about what to do and what I recommend instead. If you’re truly testing ads, a test is really only useful if you can actually learn from it.
You have to be able to learn from it, right? Or else why did you do the test? The goal of testing is not to create 5,000 versions and they all are different, but you’re not learning because they’re not getting what they need in terms of budget.
Or you create two versions, but there’s like 5,000 different things within each version. There’s not one variable, there’s five. No matter what happens, you aren’t going to know why it happened, even if it’s amazing or if it’s bad.
If you don’t know why it happened, then isn’t that a problem? Did you really test if you don’t know why whatever result you got happened? You don’t know what created the result?
Now, there have been times when I might change a couple of things on an ad. I might change the creative and the hook, for example. In that case, that is two variables.
And if there’s a significant result from that, whether it’s really bad or really good, what I can do is I can do a further test and then I can isolate the two variables and do one ad that has the one variable and another ad that has the other, and then kind of figure out pretty quickly. So there are situations where you might want to do up to two variables. But many times I see people changing like tons of stuff, five, six variables, a different creative, different headline, different ad copy, different offer, different hook, sometimes even five different headlines and five different versions of ad copy.
And then it’s just a complete mess. It’ll take you practically it seems like to figure out why whatever happened happened. Now that doesn’t mean that I’ll never, ever, ever run an ad that has a bunch of differences compared to what I was running before.
Sometimes I will, but when I want to do a true test and really know what’s going on, and I really want to know why something is happening, I’m going to do one variable or two at the very, very most. And that will give me the ability to pretty quickly find out not only what happened, but more importantly, what caused that to happen. So I can either fix it or I can do more of it and benefit from it, hopefully even more.
Now, when I’m thinking about what variables to test first, when I want to test things, not always is this the case, but many times it is. I’m usually going to start with like big things first, whether it’s the offer or the ad angle. You know, there, there are times where in campaigns I’m running, I might change a word or two in the headline and run with that.
And sometimes that can make a big difference, but other times I’m changing something larger, like a completely different ad angle or completely different image or video, something that is kind of a bigger thing to test. And so what that looks like for you might depend on how much you’ve tested in the past, what your budget is. Like there’s a lot of variables there, but just remember, if you really want to test, test one variable, two at the most, have them.
What I recommend is having them in their own ad set so that you can pretty much guarantee they’re going to get reach that you need to see what’s going on. So if I wanted to test two ads, I would have a campaign with two different ad sets and inside each ad set would be one ad going to the same audience. And then between the two ads, there would be one difference between the two ads or two at the most, at the most two.
And then pretty quickly, I would know not only what’s working and what’s not, but why? And that’s really important. And then there’s something that goes beyond that.
In addition to just doing that, I also keep a tracking system so I can look back and see previous tests. What did I do? What didn’t work?
What did work? Why did it work? Why didn’t it work?
Why did it do amazing? I can look and I can learn from that. So I’m not only learning from doing it right now, I’m looking at my notes and my spreadsheets and my software that tells me what I learned before as well.
And a lot of that starts kind of blending together and you’re able to spot trends many times quicker because of that. And that’s a big deal. So just remember, running a lot of ads doesn’t mean you’re actually testing a lot of ads, much less testing them right.
If you want to test something, ask yourself, what are you trying to learn? And how can you guarantee, or at least as close to as possible, guarantee that you’ll know not just what the result was, everybody will know that, but you’ll know what caused the result because that is where the amazing stuff starts. Because once you know that, that can change so much of what you do moving forward strategically.
And as always, just a reminder, it’s always important to point out that every business is unique. Results can vary. There’s no guaranteed outcomes.
So always keep that in mind. So I hope that was helpful, you guys. Episode 495 is in the books.
Remember, next week, we’re going to start talking about Black Friday and holiday ad strategies. But I hope until then, you guys have a great week. If you want to get in touch, PhilGrahamdigital.com is the website.
and I will talk to you on next week’s episode. Peace out.
Speaker 1:
Thanks for listening to the Next Level Facebook Ads Podcast. Please remember to subscribe and share this with all your friends. For show notes, more tips, and to learn more about Phil, please visit PhilGrahamDigital.com/Podcast.

© 2026 & Services offered by Phil Graham Digital/GRAHAM GLOBAL ENTERPRISES LLC

  • Terms Of Service
  • Client Terms
  • Disclaimers
  • Privacy Policy
  • Contact
  • Home
  • Podcast
  • AI Consulting
  • Facebook Ads
  • About
  • Testimonials
  • Contact