how to test facebook ads post andromeda update — Transcript
Full transcript
- 0:00Hey guys, it's Mark here. In this video,
- 0:01I'm going to show you how to test
- 0:02Facebook Ads post Andromeda update.
- 0:04First, I'll be walking you through my
- 0:05personal Facebook Ads testing structure.
- 0:07Second, I'm going to show you how to
- 0:08actually set up this exact structure
- 0:11within your ad account. And third, I'm
- 0:13going to tell you why this structure
- 0:14works so freaking well. So, the
- 0:16Andromeda update basically flipped my
- 0:18entire creative strategy on its head
- 0:20because pre-Andromeda, I was very heavy
- 0:23on testing a [ __ ] load of variations.
- 0:26So, when you're testing ads, you can
- 0:27kind of put them into two categories.
- 0:29You can put them into a lot more, but
- 0:30generally the way that I think about it
- 0:31is you have your variations and you have
- 0:33your new concept. So, if you can't tell
- 0:36by the name, variations are very small
- 0:38changes to existing winning ads. New
- 0:41concepts are basically completely unique
- 0:43creative material that you're bringing
- 0:45into your ad account. So, like I said,
- 0:46it flipped it on its head. What I used
- 0:48to do was the majority of the ads that I
- 0:49was launching, roughly about like 80% I
- 0:51would say, was actually variations
- 0:53because once you find a winner, you can
- 0:55just create winner after winner after
- 0:56winner after winner and skyrocket your
- 0:58win rate because you're just making
- 0:59these tiny changes to your ads. And of
- 1:02course, new concepts would have been the
- 1:03other 20%. But now, this has basically
- 1:05been completely flipped to the point
- 1:07where my creatives are 80% new concepts
- 1:10and 20% variations. And when I'm testing
- 1:13new ads from like a brand new product,
- 1:15let's say, I am pushing much harder on
- 1:18testing new concepts instead of one
- 1:19concept with a bunch of variations. So
- 1:21first, let's talk about the structure.
- 1:22To launch new ads, I recommend typically
- 1:25using a CBO. CBO stands for campaign
- 1:28budget optimization and sounds like a
- 1:30big fancy word. All it means is that you
- 1:32set the budget at the campaign level
- 1:34instead of ad set level and Facebook
- 1:35will determine how to spend your
- 1:37campaign budget across the various ad
- 1:39set. Now, I'm going to come back to the
- 1:41actual budget and how you want to
- 1:42calculate how much you should spend. But
- 1:45inside of your CBO, you're going to have
- 1:47multiple ad sets. The number ultimately
- 1:49is up to you, but if you're launching
- 1:50something new, I recommend at least two
- 1:53or three ad sets in here. So again, back
- 1:54to a CBO, if I set the budget at a $100,
- 1:57it doesn't mean that it's going to
- 1:58perfectly spend $33 here, $33 here, and
- 2:02$33 here. No, of course not. It could
- 2:03quite literally spend $99
- 2:06in this one ad set and 50 cents here and
- 2:0850 cents here. Unlikely to happen, but a
- 2:10lot of times one ad set does cannibalize
- 2:11things, but that's actually okay. This
- 2:13entire philosophy of how I structure my
- 2:16ad testing centers around letting
- 2:18Facebook do the majority of the heavy
- 2:19lifting for you. So, inside each of
- 2:21these ad sets, the ad set level is
- 2:22typically where you set your targeting.
- 2:24So, who are you really going after? And
- 2:26what I do inside of these ad sets is I
- 2:28keep them completely broad, which means
- 2:30Advantage Plus placements. Typically,
- 2:32I'm running in terms of ages 18 to 65
- 2:35plus and I'm running male and female. Of
- 2:37course, if you're only selling to one
- 2:39demographic, then select the one that's
- 2:40most relevant to you. But if you're
- 2:41selling to both, keep it broad. Now, at
- 2:43this point, you may be wondering, "Mark,
- 2:44why do you have three ad sets if they're
- 2:46all going to be identical? They're all
- 2:47broad, they're all Advantage Plus, and
- 2:49they're all 18-65 plus. Everything is
- 2:51broad. Why do you have three of them?"
- 2:53Well, the reason I have three is because
- 2:54they're not actually identical. You see,
- 2:56inside each of these ad sets, if this is
- 2:58ad set one, this is ad set two and this
- 3:00is ad set three, I'm actually going to
- 3:01have three different creatives in each
- 3:03of these three ad sets. Again, I'm just
- 3:05using three. If you want to have more
- 3:07ads per ad set, you can. But here's the
- 3:09point that I'm trying to make. You want
- 3:11to segment these ad sets by ad concept.
- 3:14So, let's say ad set one is ad concept
- 3:16one, ad set two is ad concept two, and
- 3:19ad set three is ad concept three. And
- 3:21then inside, let's say you have three
- 3:23creatives each, which these three would
- 3:25actually be variations of that one
- 3:27concept, okay? I'll explain kind of what
- 3:29I mean, but this would be a total of, if
- 3:31we count the actual ads, 1 2 3 4 5 6 7 8
- 3:359 total ads for this ad launch. But
- 3:38these three ads are going to be
- 3:39variations of this one new concept. So,
- 3:43the reason why I said you can have as
- 3:44many ad sets as you want in here is cuz
- 3:46you could have five ad sets in here, but
- 3:48only one ad per ad set and each one is a
- 3:50new concept. I think the sweet spot is
- 3:52not one. I think it's like anywhere from
- 3:54three to 10, but [ __ ] I have ad sets
- 3:57running that have 20 plus ads in there.
- 3:59So, you know, 30 plus ads, 50 plus ads,
- 4:01some of them. There isn't a rule here,
- 4:03okay? I'll tell you what I recommend is
- 4:05anywhere from three to five ad sets,
- 4:07which means three to five concepts,
- 4:09completely unique concepts, and anywhere
- 4:11from two to five variations in each of
- 4:14those ad set. And by the way, this is
- 4:16mostly relevant to images and videos. I
- 4:18typically do my testing with images cuz
- 4:21you can typically just do it faster. But
- 4:23if you want to test with images and
- 4:24videos, you can. I would recommend
- 4:26segmenting them in the ad set or even in
- 4:29a completely different CBO because
- 4:31typically what I see is if you throw a
- 4:33bunch of images and videos in the same
- 4:35ad set, one of them will cannibalize
- 4:37everything else. Like all the videos
- 4:38will get spent and the images will get
- 4:40no spend, which of course you don't want
- 4:42cuz you still want to make sure you test
- 4:43those images, right? So, I'll do a
- 4:45separate CBO if you're going to combine
- 4:47images and videos. For me, I just keep
- 4:49it simple. Just pick one, only videos or
- 4:51only images, and run with that. Now,
- 4:53let's talk about the budget. How do you
- 4:55calculate the budget for a CBO? Well, I
- 4:57have a golden rule that I would have at
- 4:58least $10 allocated out of my total
- 5:01budget per ad in the CBO. So like I
- 5:04said, how many ads do we have in the CBO
- 5:05right now? Well, 1 2 3 4 5 6 7 8 9. If
- 5:08we take nine and multiply it by $10,
- 5:11that is a total of $90 a day, a day,
- 5:15yes, that you want to run this CBO ad.
- 5:18So, you take the number of ads across
- 5:19all of the ad sets total, multiply that
- 5:22by 10, and that is going to be the
- 5:24minimum daily budget that you run with.
- 5:26The one exception to this rule is I
- 5:28would never recommend that you run a CBO
- 5:30at less than $50 a day. I don't care
- 5:32what your budget is. If you don't have
- 5:34enough cash set aside to run a $50 a day
- 5:37CBO, you should not be running paid ads.
- 5:39It's my personal opinion. If you don't
- 5:40like it, I do not care. So now that
- 5:42we've gone over the structure, let's
- 5:43actually build this inside of the ad
- 5:45account. So, I've got almost a brand new
- 5:47ad account here. What we're going to do
- 5:48is go select create new campaign. We're
- 5:50going to hit sales and hit continue. I'm
- 5:52not going to go over naming conventions
- 5:54in this particular video, but name it
- 5:56something that makes sense for you at
- 5:57the campaign, ad set, and ad level.
- 5:58Here, we're going to set the campaign
- 6:00budget. Let's use our example to $90 a
- 6:02day and make sure that you have campaign
- 6:04budget selected. Everything else here,
- 6:06we're not going to touch. We're going to
- 6:07keep it exactly as is and move on to the
- 6:09ad set level. Here we are at the ad set
- 6:11level. We're going to have website
- 6:12conversion location. Performance goal is
- 6:14going to be max conversions. Data set is
- 6:16going to be the pixel that you're
- 6:18currently running your traffic to.
- 6:19Conversion event, we are going to be
- 6:20optimizing for purchases. I almost
- 6:22exclusively optimize for purchases and
- 6:25almost exclusively run sales campaigns
- 6:27because anything else is mostly mostly,
- 6:30with very rare exceptions, a complete
- 6:32waste of your time and money. Now, we're
- 6:33going to keep scrolling down. Under the
- 6:36attribution window, I typically do 7-day
- 6:38click and 1-day view and I keep the view
- 6:40through completely turned off. You can
- 6:41do testing on this once you get a little
- 6:44bit more advanced, but initially this is
- 6:46a super standard setup, 7-day click
- 6:481-day view. Now, onto the budget and
- 6:49schedule. I always schedule my ads for
- 6:51the next day at 6:00 a.m. EST. This
- 6:55account is actually in PSD, but imagine
- 6:57it was in EST. I typically run my ads
- 6:59right around that time to start getting
- 7:00active. Next, what we're going to do is
- 7:02edit the location. So, I do the majority
- 7:04of my testing in the United States.
- 7:06We're just going to search United States
- 7:07here, select it, and remove Singapore,
- 7:09which is there for some reason, and
- 7:11continue to scroll down. Minimum age,
- 7:13you can keep. Languages, I typically
- 7:15keep broad as well in the US because
- 7:17it's an English majority speaking
- 7:18country, at least that's what my
- 7:20customers are going to be buying in. And
- 7:21then finally down here, as I said, we
- 7:23keep it broad. We're going to have the
- 7:25Advantage Plus placements on. So, there
- 7:27is nothing for you to edit within here.
- 7:30You can keep it completely broad, don't
- 7:31touch anything. All right, now we are at
- 7:32the ad level. Once again, apply your
- 7:34naming conventions here. Keep it nice
- 7:35and neat and organized. You're going to
- 7:37have create ad selected, single image or
- 7:39video. You can keep multi-advertiser ad
- 7:41selected. Now, let me tell you this one
- 7:43time. I'm going to upload an ad here
- 7:45just as an example and I'm going to show
- 7:47you how much nonsense Facebook tries to
- 7:49push on you with their updates. I am
- 7:52utterly amazed because I still spend
- 7:54time in my ad accounts every single day,
- 7:55but I'm amazed at how many [ __ ]
- 7:57settings, and that's me putting it
- 8:11very lightly, there are for Facebook to
- 8:13try to take your money. Majority of them
- 8:15are a complete waste of time and money.
- 8:17So, just follow what I do here. This is
- 8:19basically the only setting that it's
- 8:20actually okay to keep it on. And even
- 8:22then, I haven't split tested
- 8:23extensively, but keeping it on
- 8:25personally doesn't hurt. Under
- 8:26destination, you're going to put the
- 8:27website URL that you're driving traffic
- 8:29to. Make sure that this matches the ad
- 8:31the uh if you put your URL in the ad
- 8:33copy as well, make sure that it matches
- 8:34one to one. And also make sure you
- 8:36deselect optimize website destination
- 8:38because if that's on, it will send to
- 8:40traffic outside of that direct domain,
- 8:42which we don't want. Display link, you
- 8:44can keep empty. Browser add-ons, you can
- 8:46keep at none. You don't want to have any
- 8:48add-ons. Now, let's actually upload the
- 8:50creative. Let's say we're running image
- 8:51ads in this case, we go ahead and select
- 8:53image ad. So, as we're going through the
- 8:55creative setup, again, make sure that
- 8:57just about every setting here is turned
- 8:58off. If you want to split test it later
- 8:59down the line, you can. But for me
- 9:01personally, I keep everything off here.
- 9:03So, promo codes is unchecked, branding
- 9:05off, site links off, products off. We're
- 9:08going to hit next and actually upload
- 9:09the creative. And just like that, we got
- 9:11our phenomenal creative there. If you
- 9:12want to take a look, boom, beautiful 10x
- 9:15row ads coming very soon. Now, I don't
- 9:18adjust any of the aspect ratios here.
- 9:19We're going to hit next. Now, you're
- 9:20going to fill in your copy, primary
- 9:22text, the headline, and the description
- 9:24is all there and it's a fantastic
- 9:26message, run more ads. Going to hit
- 9:27next. Image generation, deselect all. It
- 9:30won't even let me deselect it. Now,
- 9:32under enhancements, you want to make
- 9:33sure all of these are turned off. And be
- 9:35careful cuz there's a lot of them. So,
- 9:36you got to scroll down. For example,
- 9:38overlays, turn off, visual touch-ups,
- 9:40music, text improvements, all turned
- 9:41off, animation off as well. All right,
- 9:43there we go. There's our lovely
- 9:44beautiful ad. Let's keep cooking. Keep
- 9:46scrolling down. Call to action, I
- 9:48usually have learn more. I actually have
- 9:50seen a slight lift in learn more versus
- 9:53shop now, but I think most of the time
- 9:55it's pretty trivial. So, depends also
- 9:56what kind of ad you're running. So, I
- 9:58keep it a learn more. If you are going I
- 10:00set this up for one ad, I'll just
- 10:02duplicate it, you know, two times, and
- 10:03I'll go in and edit the creative to
- 10:05change it to a different variation. And
- 10:07after that is created, I'll duplicate
- 10:08this ad set two times, so I have three
- 10:11ad sets in here. The CBO budget is
- 10:13currently set to $90 a day, three ad
- 10:15sets, three concepts, three ads per
- 10:17concept, for a total of nine total ads,
- 10:20as you can see here. And they're all
- 10:21going to be beginning to run tomorrow
- 10:23directly at 6:00 a.m. Eastern. I
- 10:25actually made a separate video to this
- 10:26that is unlisted on my YouTube channel
- 10:28right now, where there's so many
- 10:30questions that come up after you launch
- 10:32the ads. It's like, how do I know when
- 10:33to kill an ad? How do I know when to
- 10:34scale an ad? What if an ad doesn't get
- 10:36spend? I briefly covered that, but it's
- 10:37a little bit more nuanced. What about
- 10:39testing landers? Do you ever use ABOs?
- 10:41There's like a million questions that
- 10:42occur once you click publish ads, and
- 10:45you actually start getting data. So, I
- 10:47put together a separate video. If you
- 10:48want to get access to that, just go to
- 10:49my Instagram, it's at markbuildsbrands,
- 10:51and DM me the word law, like l a w,
- 10:55because these are laws of scaling. Now,
- 10:56the last thing I want to touch on here
- 10:57is why does this strategy actually work?
- 11:00Well, we have seen, or at least I have
- 11:02seen over the years, Facebook push many,
- 11:04many updates to the point where I
- 11:05remember in 2019, they said that CBOs
- 11:08were going to be mandatory, and they
- 11:09were going to just delete ABOs off the
- 11:11face of the earth. So, every time that
- 11:12Facebook will launch an update, you have
- 11:14to kind of take with a grain of salt,
- 11:16and most importantly, trust the data.
- 11:17So, for me, I have a lot of freaking
- 11:19data. Not only do I have my own ad
- 11:20accounts that I run, I have friends' ad
- 11:22accounts that I'm looking at, but I also
- 11:23have all of our students inside of BBA,
- 11:25where we are instructing them the best
- 11:27testing structure. Which So, we've been
- 11:29doing a lot of testing internally to
- 11:30find out what that best structure is,
- 11:32and this is what we found works really,
- 11:34really well. So, the number one reason
- 11:35why it works is just straight data. You
- 11:37know, there's a lot of people that will
- 11:38say, "Oh, this is like not optimal, or
- 11:40you should do, you know, straight up ASC
- 11:42campaign." Look, to be honest, I really
- 11:44don't give a [ __ ] I care about data
- 11:46over anything. A lot of times, Facebook
- 11:48says, "This is how the Facebook machine
- 11:49operates." And all the [ __ ] I don't
- 11:51care. I listen to what the data says,
- 11:52and I am willing to test just about
- 11:54anything in pursuing the right kind of
- 11:55data. But more importantly than that,
- 11:57the real reason why it works is because
- 11:59of what I was mentioning earlier, where,
- 12:01you know, I have shifted much more of my
- 12:02ads to being more unique, more creative
- 12:05diversity, and increasing my ratio of
- 12:07variations to new concepts. But I don't
- 12:09necessarily want you to think about it
- 12:11in terms of like ratios, like 80/20. I
- 12:13think the right way to actually think
- 12:15about how you approach your testing is
- 12:17you basically just change the order in
- 12:19which you're testing variations and new
- 12:20concepts. So, again, like I said
- 12:21earlier, the way that I used to do
- 12:23testing is let's say one concept, like
- 12:26one new concept, and then variation,
- 12:28variation, variation, variation. Just
- 12:31like this, right? And the only
- 12:33difference now is that instead of just
- 12:36testing with one new concept, I'm
- 12:37testing with many. I showed you in this
- 12:40video an example of three, but typically
- 12:41I'm testing with even more than that.
- 12:43Sometimes as many as 10 new concepts for
- 12:46a single product test. So, if I were to
- 12:48visualize it, it's more like this. Like
- 12:50if this were kind of the the five ad
- 12:51sets that I'm running, it's new
- 12:52concepts, and then all the variations in
- 12:55here, right? So, yes, I'm actually
- 12:57testing more ads typically when I'm
- 12:58launching, and that's not a hard rule.
- 13:01The point is,
- 13:03if it hasn't been made abundantly clear
- 13:04by this point, is that you need to
- 13:06prioritize more creative diversity and
- 13:08new concepts in your testing, and less
- 13:10variation testing. Because one is
- 13:12clearly going to move the needle way
- 13:13more than the other. And I know, like
- 13:15the number one objection here,
- 13:16especially when you run CBOs, is, "Well,
- 13:18what if like this ad set gets no spend?"
- 13:21Or like, "What if this ad or these few
- 13:23ads get no spend or being cannibalized
- 13:25by the other?" Well, good. You don't
- 13:26necessarily need to test every single
- 13:28individual ad at the same level of ad
- 13:31spend across, you know, every single ad
- 13:33concept that you're running. Sometimes,
- 13:35sometimes, if Facebook doesn't spend on
- 13:37that particular creative at all, or
- 13:39very, very little, then it means that
- 13:40that creative wasn't that that good to
- 13:42begin with in the first place. So,
- 13:43that's not a hard rule. Sometimes,
- 13:45Facebook is [ __ ] I'll be the first
- 13:46[ __ ] one to admit sometimes Facebook
- 13:49will not spend on a creative, even
- 13:51though it has a like the lowest cost per
- 13:53conversion. So, it's a tough balance.
- 13:55You know, all I can say is that this is
- 13:56supported by data, and I can say one
- 13:58thing for 100% sure in my heart of
- 14:01hearts, which is that you need to be
- 14:02prioritizing creative diversity
- 14:04post-Andromeda. This variation spamming
- 14:06game is just not nearly as effective as
- 14:08it was previously. I kind of alluded to
- 14:10it earlier, but this strategy is exactly
- 14:12what we teach inside of Brand Builds
- 14:13Academy. So, if you're interested in
- 14:15like having somebody actually like help
- 14:17you out, you know, set this up, and more
- 14:19importantly, like the feedback loop that
- 14:20comes after it, you can always click the
- 14:21first link in the description and fill
- 14:22out an application. We'll see if you're
- 14:24a good fit. Other than that, guys, thank
- 14:25you so much for watching. I'll see you
- 14:26legends in the next one. Peace.
About this transcript
This page contains the full transcript of how to test facebook ads post andromeda update by Mark Builds Brands, generated from the public captions YouTube serves with the video. The transcript has 3,320 words across 454 segments, with the original timestamps preserved so you can click any line to jump to that moment in the embedded player.
What you can do with it
Use the transcript to take notes, quote the speaker, build a study guide, generate a summary with ChatGPT or Claude via the YouTube Summary tool, or export it as a timed subtitle file with YouTube to SRT. You can also re-open it in the transcriber to choose a caption track available on the video.
Free YouTube transcript tool
YouTube2Text is a free YouTube transcript generator — no signup, no daily limit. Paste any YouTube link and get the full transcript instantly, with timestamps, click-to-jump, caption language selection, AI prompts for ChatGPT, Claude, and Gemini, and exports to TXT, SRT, VTT, or Markdown.