How to: Structure Meta Testing & Scaling (Andy + Jason 04 11 25)

Summary

Highlights

• Segment audiences and ad creatives to methodically identify high-performing elements over time. 🌟
• Implement diversified testing strategies using pivot tables and aggregated data analysis to ensure accuracy. 📊
• Focus on specific performance metrics to optimize ads progressively, aligning with set budget constraints. 💡

Key Takeaways

• Identify and test multiple angles and creatives in Facebook ads to find winners progressively. 🍏
• Use CBOs strategically for scaling successful ad sets, focusing on data-driven decision-making. 📈
• Emphasize on creating systematic, step-by-step processes in ad structuring to manage and evaluate campaigns effectively. 🛠️

Overview

Chapters

• 00:00 - 00:30: Introduction to Meta Testing and Structuring Strategies The chapter begins by discussing the overarching aim of the strategy, which is to identify and test 'winning' ad sets. The process involves various structuring methods, which will be explored throughout the chapter. Key to success is identifying multiple winning angles, compiling them, and deploying them into a Campaign Budget Optimization (CBO) strategy. The chapter emphasizes that a greater number of successful ad sets will sustain the CBO strategy for a longer period.
• 00:30 - 01:00: Using Angles in Advertising The chapter 'Using Angles in Advertising' discusses strategies for promoting different products within a business, particularly a fruit basket business. The focus is on diversifying advertising angles for each product. For example, under apples, different angles are suggested, such as apple slices, apple pies, or apple sauce. Similarly, for bananas, there are angles such as banana cream pie, banana ice cream, banana bread, etc. The chapter emphasizes the importance of testing these angles to find the most effective advertising strategy.
• 01:00 - 01:30: Testing Strategies and Audience Targeting The chapter discusses testing strategies and audience targeting in advertising campaigns. It starts by illustrating a scenario involving different advertising tests using 'applesauce', 'banana bread', and 'orange juice' as metaphorical examples in a Controlled Business Launch (CBL). The narrator outlines various strategies for structuring these tests efficiently. Additionally, it emphasizes a previous misconception about ad sets and audience, clarifying that all 'apple winners' (successful ad variations or target audiences metaphorically represented by apples) should indeed be grouped within the same ad set.
• 01:30 - 02:00: Optimizing Ad Set Budgets and Retargeting The chapter titled 'Optimizing Ad Set Budgets and Retargeting' discusses a strategy for improving the performance of ad sets by mixing audience segments from different fruit categories, metaphorically speaking. Andre, the partner of the speaker, suggested the notion of retargeting audiences with this mixed approach. The strategy involves taking winners from different audience segments (fruits like apple, banana, and orange) and mixing them to see if the performance improves. This approach proved to be significantly more effective than previous methods.
• 02:00 - 02:30: Challenges with Facebook Ads and Creative Testing The chapter discusses strategies for managing Facebook Ads and creative testing, focusing on how to allocate budgets effectively. It mentions running five winning ads within an adset, which allows for larger budgets in successful adsets. The chapter also highlights the challenges of handling individual ads per adset, noting that it can be difficult to control and requires Facebook to allocate spending across multiple ads, necessitating more budget and time.
• 02:30 - 03:00: Understanding Campaign Angles and Ad Text The chapter discusses the dynamics of campaign angles and the importance of ad text in advertising. It highlights a recent realization about an underperforming ad that previously succeeded. Despite its poor performance, the ad continued to receive traffic due to its high engagement metrics. This was likely because of Facebook's algorithm rewarding engagement metrics, leading to disproportionate traffic distribution. The team had to intervene by removing the ad to redirect traffic to better-performing ads. This leads to broader questions about optimizing ad performance and ensuring that there are measures in place to accurately assess the effectiveness of different campaign angles and ad texts.
• 03:00 - 03:30: Impact of Headlines and Copy Hooks in Ads This chapter discusses the various strategies for testing advertising strategies, specifically focusing on the impact of headlines and copy hooks. The speaker outlines their personal involvement in leading a project centered around copy and AI-generated images. A new initiative called 'pruning campaign' is introduced, which appears to be an approach to refining and optimizing ad content.
• 03:30 - 04:00: Performance Metrics and Optimization The chapter discusses the process of optimizing campaign budgets. It begins with using low budgets for initial tests to identify potential winners. Once initial tests indicate success, the budget is increased for further validation, referred to as a 'mid-form' stage. The process is likened to going on a date, suggesting a gradual and exploratory approach to budget allocation.
• 04:00 - 04:30: Scaling Campaigns with Creative Variations The chapter discusses strategies for scaling campaigns by introducing creative variations. It uses an analogy to demonstrate a low commitment approach, likening it to meeting for coffee before deciding on dinner and a movie. The focus is on setting budgets to distribute copy effectively, and the importance of determining the amount of copy and images needed based on budget constraints. The chapter suggests creative testing by mixing different headlines with varying ad texts to optimize the campaign output.
• 04:30 - 05:00: Importance of Account Structure and Testing Approaches The chapter "Importance of Account Structure and Testing Approaches" highlights the significance of proper campaign structuring and testing methodologies in advertising. It emphasizes the necessity to separate different campaign angles, such as "apples and oranges," for clearer performance analysis. The discussion also touches on the ideal setup of advertisements, using two headlines and ad texts, and iterating with the same set of images, to maintain consistency in testing and evaluation. This allows for a focused analysis of each campaign's performance, facilitating better decision-making.
• 05:00 - 05:30: Case Studies and Real-World Examples This chapter discusses the use of faces in front-end testing for advertisements. The premise is that faces in images attract attention and improve engagement with the accompanying copy. By testing different combinations of images, headlines, and ad text, one can determine which elements capture the audience's attention most effectively. It describes a scenario where three different faces are used across multiple ad sets, each paired with two different headlines and ad texts, to identify the most successful combination.
• 05:30 - 06:00: Evaluating and Selecting Winning Ads The chapter focuses on methods to evaluate and select the most effective advertisements. It discusses analyzing aggregated data from various ad sets, such as multiple hits of each headline and ad text, and suggests different testing frameworks like 2x2 or 4x4 setups. Emphasizing the importance of the advertising angle, the chapter underlines the significant impact of the lead generation on the success of ads.
• 06:00 - 06:30: Testing Various Angles for Veterans Campaign The chapter titled 'Testing Various Angles for Veterans Campaign' focuses on the challenges and strategies involved in advertising, particularly in crafting the ad text. The discussion highlights that while long-form content allows for storytelling, short-form ad text, which is often more challenging, requires precision and impact due to limited word count. The irony is noted that copywriters, accustomed to verbose storytelling, find it difficult to condense their message into the brief, impactful statements necessary for effective advertising, as they only have a brief space to make a strong impression.
• 06:30 - 07:00: Exploring Professions and Segment Testing In this chapter, the focus is on exploring different professions and the process of segment testing. The narrator humorously recounts an unexpected observation about assumptions being challenged, particularly about the effectiveness of certain methods. The discussion revolves around running headline tests and searching for leading indicators in a lead generation funnel scenario. They set up a situation where a funnel runs to a pre-sale, addressing multiple possible flows but concentrating on the pre-sale path. The search for key metrics and insights is central to understanding the funnel's performance.
• 07:00 - 07:30: Determining Effects of Budget and Creative Testing The chapter discusses strategies for optimizing the effectiveness of marketing campaigns by examining cost metrics such as CPM (Cost Per Mille) and CPC (Cost Per Click), along with key performance indicators like CTR (Click-Through Rate). The speaker highlights the necessity for achieving a CTR above 5% for successful campaigns. The testing involves using various Facebook settings including mobile newsfeed only, Facebook only, and Wi-Fi only, targeting users at home. Additionally, the chapter mentions the use of a 2% lookalike audience for reaching potential customers effectively.
• 07:30 - 08:00: Developing Ad Copy and Testing Methodologies The chapter titled 'Developing Ad Copy and Testing Methodologies' discusses strategies for improving advertising outcomes through both ad copy creation and thorough testing methodologies. The focus is on metrics that drive decisions, rather than immediate profitability. One example given is the importance of knowing specific funnel metrics, such as the percentage of people who begin checkout in an e-commerce context and how many eventually make a purchase. By acquiring three times the data, marketers can better optimize and predict outcomes. The chapter emphasizes the necessity of understanding how many leads convert from initial interest to a completed action, like purchasing, and how critical it is to surpass expected performance benchmarks in these metrics.
• 08:00 - 08:30: Concluding Thoughts and Advice on Media Buying This chapter discusses key strategies and metrics for media buying, emphasizing the importance of analyzing data metrics in marketing funnels. It suggests that the initial data collected from lead generation funnels or quiz funnels can offer valuable insights. The author notes the significance of certain time markers in video sales letters (VSSLs), such as the importance of five-minute engagement, and contrasts it with the common focus on two-minute drop-offs. The chapter advises that understanding such critical data points can enhance decision-making in media buying.
• 08:30 - 09:00: Final Reflections on Strategy and Testing The chapter discusses the importance of evaluating user interaction data for strategic decision-making, particularly in e-commerce. It highlights the need to track user progress through various steps, such as reaching the social security page, and understanding the drop-off rates at each stage. This data-driven approach is emphasized as it provides invaluable insights that can triple the information available for making informed decisions.

How to: Structure Meta Testing & Scaling (Andy + Jason 04 11 25) Transcription

• 00:00 - 00:30 okay cool so let me let me just start high level with sort of what the goal is the goal is really to find to test and find those winners in adset budgets bunch of different ways to structure it we'll go through that and then you want to grab those winners and you want multiple angles and then you want to grab those winners and you put them in a CBO wide open and you let them go over time the more winners that you get in that CBO the longer the CBO will last and the
• 00:30 - 01:00 more budget it can take so you have a fruit basket business right so you're selling fruit baskets and you find and you have apple banana orange let's say so you're like under apple you're like apple slices apple pie apple sauce a regular apple whatever banana banana cream pie banana ice cream banana bread banana regular and so on and so forth right and then what and then after you test each one of those and you find those
• 01:00 - 01:30 winners you're going to drop in applesauce banana bread and orange juice let's say in a CBL um and then we'll talk about that structure in a second there's a couple different ways to do it um but let me get this out here in case I forget we did a test so I believed incorrectly that the ad set is the audience so we're going to put all the Apple winners it could be two right but all the apple winners in an ads set and then we're going to put all the banana
• 01:30 - 02:00 winners in another adset and whatever so it's going to go apple people banana people and orange people so Andre my partner had this idea and he was like well like what if they just need retargeting for that audience so we don't know we do know retargeting i'm guessing you guys don't either so like right so we mix the fruit so one apple winner one banana winner one orange winner another apple winner another banana another orange and that performed better um a lot better actually we also
• 02:00 - 02:30 found that um the winning I think we do five now so we'll do five winners in an adset kind of the max and then basically run them by adsets so then those adsets can get larger budgets we have also tried um individual ads per adset but the of that is is like it's just kind of harder to control and then Facebook has to get spend around all of them which given enough budget and time it
• 02:30 - 03:00 will but like what we found recently is we had an ad that used to be a winner that's no longer the winner and Facebook just kept it was super short form kind of clickbay and Facebook just kept giving it all the traffic we actually had to uh pull it out i was like Facebook's just rewarding this because it has high engagement and it's the worst performer anyway so that's where we're going so the question then is is like how do we get there um so let's go back
• 03:00 - 03:30 to the testing strategy there's a million testing strategies this is ours it's not the holy grail but the thesis is basically the same for everybody so let's start with the huge number of copy and AI images that is a project that I'm personally leading within the organization and basically what I call those campaigns this is a brand new thing we're working on so we have what I call a pruning campaign and that's like a ton of
• 03:30 - 04:00 stuff it's all adset budgets all low budgets um one ad per adset and then we run low budgets try to basically do the initial test on those winners and then we take them to like I don't have a name for this one but kind of the mid form and you're like all right let's put some more budget on this and let's see if these are actually winners so I'm looking for leading indicators and then I'm looking to actually test so the way that I think about it is you go out on a date with somebody you're like "We're
• 04:00 - 04:30 just going to meet for coffee if I can if I have a good time with coffee then we can go to dinner in a movie." Right so you're just trying to have low commitment all set budgets to distribute all of that copy right historically we've set up this way and it depends how much copy you want to create and how many images you want to create and what your budget constraints are that's really what it comes down to but we used to run headline one add text one headline one ad text two headline
• 04:30 - 05:00 two ad text one headline two ad text two right and then the same three images in all three and when you think about the campaign structures and what you're testing the campaign is always the angle this is obvious but I'm just saying it just in terms of how my head thinks so you don't want to run apples and oranges in the same campaign you should be able to look at the general performance of a campaign and know how that rough angle is doing right and then because then when you dive in you're really focused on the
• 05:00 - 05:30 copy let's say and for us we always almost always anyway test faces on the front end and the reason we do faces is because an image a face just gets people to stop and read the copy that's all I'm trying to do i'm like what's the winning copy so let's assume for a moment we're running these you know I got three faces and same three faces in each adset i got two headlines two ad texts and then I run it for a little bit i kind of see what's working
• 05:30 - 06:00 what's not but then I look at the data in the aggregate because you have multiple hits of each headline and multiple hits of each ad text right and so you could also do a 4x4 right so you have 16 adsets so like there's different and what yeah there's like there's a bunch of different ways to do it but the general framework is is the same and I usually just like to start with the 2 by two test the angle is a good what we do know is the lead matters a ton um
• 06:00 - 06:30 on your on your um on your on your ad text so like what what I'll do or what we will do is we're going to come up with maybe it's mid form or long form i'm not talking about short form here uh ironically the thing that our copywriters struggle the most with is short form because they're used to being verbose and telling this like dramatic story and I'm like "Dude you got three lines of text you need to just rock somebody in the jaw really hard really fast you have less words right?"
• 06:30 - 07:00 So it was kind of funny actually i didn't expect that i was like "Oh they'll kill this the same." And they're they're just long-winded and used to that so anyway so you run these headline ones headline two and you're looking for basically leading indicators so let's assume for a moment you're running just a basic lead gen funnel and let's assume for a moment um that it's running to a pre-sale and I know there's a bunch of different flows that you can run where but let's just say it's going to a pre-sale now what I'm looking for is I'm looking
• 07:00 - 07:30 for the event like yes I'm looking at CPMs and CPCs and all of these things right i'm looking for a CTR above a certain level right typically for us we need to be north of 5% adr um oh we are also testing mobile newsfeed only Facebook only Wi-Fi only okay yep that's a big one because I want people who are at home i want people who are on Facebook and I want people who uh 2% lookalike audience which is roughly
• 07:30 - 08:00 going to perform the best again I'm not looking for profitability i'm looking for the metrics for example if you know in your lead genen funnel like if we use just ecom metrics 30% of the people that get to the initiate checkout should buy so I'm going to get three times the data so you may know that 30% of the people that fill out a form on a lead hit the ping train and get bought i know it's probably better than that but it just
• 08:00 - 08:30 gives you more data up front right or let's just depending on what your quiz funnel is but there's probably a or not quizunnel um lead genen funnel there's probably one metric in there that is also really really telling so on VSSLs for example five minutes is super important so a lot of these guys look at the two-minute drop off right but if you have a hook that's 15 seconds maybe you run a video let's say you're running 15
• 08:30 - 09:00 seconds 30 seconds like you're with that data so really I'm like all right I just want to know how many people got to 5 minutes i need it a little bit further down so let's assume this is refi and you require social security you can see how many people are getting to the social security page maybe what's the drop off there or that step right after which just gives you more data to make decisions on ecom it gives us three times the amount of data okay and then when I take a look at
• 09:00 - 09:30 the budgets so let's assume I have a $100 CPA offer that's what we're targeting for me i'm like all right that means the initiated checkout needs to be like um I should have done a $90 CPA let's call it a $90 CPA my initiate checkout needs to be 30 bucks and so if I'm coming in at 45 maybe it's day one I let it run two days right it's got to stabilize whatever no big deal but really what I'm doing is I'm looking
• 09:30 - 10:00 at the funnel metrics because if your funnel metrics look fire but your cost per click is $2 and you know you're never going to be able to scale it it'll work at two bucks by the way I've had this but like you just can't drive that scale because your costs are too high then you have the right angle you just need more engaging more rewarding imagered driven whatever creatives but that angle works and so the biggest you know I think the biggest mistake that affiliates make or
• 10:00 - 10:30 that affiliate mindset today and we used to have it add purchase end all be all and it's like yeah you can you can do that but and and lead genen's a little bit easier than a $100 CPA let's say but it still comes down to the fundamentals so what are those just couple metrics that you're looking at to get more data to feel more confident in the campaign and angle that you're running what we have found and you
• 10:30 - 11:00 mentioned this uh earlier today is that there's like a winner that winner like just works right and so that ad copy is going to last a long time a long long time and so you're going to have to refresh the creatives and stuff like that so and feel free to interrupt and stop with questions at any time so now I went through and I found that this angle generally like kind of works i think that headline one and add text two work and these images are okay
• 11:00 - 11:30 by the way I'm losing and I'm losing because my uh creatives are faces now let's assume for a moment that you're running um refi so we did do this also um on refi we may come up with our three winning home images and those are just the three houses we basically use over and over and over and the reason that you want to use the same images is because you're
• 11:30 - 12:00 just removing one variable right so again you're not focused on profit so if I came to you and I was like Andy you work for me your campaign is getting killed this week you lost like five grand like what the hell is going on you're losing $1,000 a day which we have to run those budgets because we have to drive scale whatever and you're like "Yeah but look what I found." And you're like "I found that this ad text and these winning combinations really have amazing funnel
• 12:00 - 12:30 metrics the lead quality is really really high they get picked up off the ping tree in a second they all convert whatever it is i'm like "Okay that's super interesting." So now what we have to do is go back and rework that angle in those creatives to drive the CPC's down and that game is is where you're fighting the Facebook black box in terms of what do they reward you with does that make sense yeah yeah so after we figure so let's assume you add houses but after we figure out the winning ad
• 12:30 - 13:00 text then we restructure the campaign and maybe it's five adsets maybe it's 10 adsets and let's assume it's headline one ad text one and ad text two put all of those in one adset and then now I'm testing the creatives so you're just moving methodically through the process what that does is it pulls off a little bit of pressure from you and it allows you to think more clearly i I believe this because you're not just like I put
• 13:00 - 13:30 up this new campaign it's losing like what do I do well the goal wasn't to go from here to here the goal is to go from here to here to here to here to here to here and show progress because we know that most things aren't going to work so now you find this winner on this Apple one you got this Apple test campaign it's working just keep testing it like just just keep keep pushing it up as long as you can but then you you can take that and
• 13:30 - 14:00 you can drop that into a CBO so if you have one angle in a CBO we run two five 10% and open lookalike audiences we take the same ads put them in those four ad sets um push a you know day one,000 bucks on the CBO budget let it all placements all everything let it rip and let it go um and so now you have a block so now you go through and you do this again and you find the banana bread angle right and so you found the right copy you
• 14:00 - 14:30 found the right image and um and now all of a sudden you move to this other CBO now you got like two of these CBOS well for me I probably wouldn't take those two and put them together for me the number's three so now you get the orange juice angle is working really well so you have another CBO over here and then um and then we have those three CBO scaling campaigns typically we keep testing and scaling in separate ad accounts um I also do not I also do not
• 14:30 - 15:00 believe that media buyer should be running in the same ad account like how do you just easily go in and look at your performance sure you have a tracker but like you have to be checking your data against Facebook there's probably attribution loss on Facebook someplace maybe there's a delay in cost like um ad accounts matter so if some guy's with your ad account and throws up some Trump whatever you get some score on the back end that some other guy did like that guy that's insane to me and what also what that also
• 15:00 - 15:30 allows is the manager CFO whoever the hell it is the CEO is looking they can actually look at the performance by um ad accounts by manager and all of these things um you might get push back but push harder um tell me talk to me and that's what that's what they should do um yeah so I just don't I just don't believe in that same thing with pages don't share your pages we do that a lot
• 15:30 - 16:00 like all running on the same page that's a mistake um the number one mistake with that is what happens if your page gets flagged what happens if your ad account goes down what happens if a creative gets flagged what happens if copy gets flagged i've had all of these and so they may be like "Well we're going to get better performance." And I was like "Cool but what would you pay for insurance and diversification?" And so all of our media buyers run on um it kind of depends if you have two teams
• 16:00 - 16:30 on the same offer they run on the same pixel different pages different ad accounts so they share the pixel but the Yeah you can share the pixel yeah but ad accounts and pages are all independent for each media buyer yeah because like you're running different ang like unless everybody's doing the same thing and then I don't understand why you need a bunch of media buyers to be honest yeah yeah so you're testing different angles which is going to build different audiences off of different pages and
• 16:30 - 17:00 you're going to get different learnings and different ad accounts so like I don't know let's assume you're really good the guy who's not really good is probably dragging you down so there's that this will be affecting the performance of the page and because we don't really know what's going on with the page do we we can only assume by the way pages matter accounts matter how do they matter don't know i've been doing this a long time i know they're not all they don't all perform the same
• 17:00 - 17:30 we've run tests where we run CBOS and all we do is change the pages and not all pages perform the same right take that same campaign all post IDs different ad account different ad accounts do not perform the same now a seasoned ad account will perform better than a brand new ad account for sure but all seasoned accounts started as new ad accounts so you could have one testing account and one scaling account dude you can put something up and that
• 17:30 - 18:00 account goes down the AI just doesn't like it managed by Facebook don't care agency like none of that matters we had this book called the Backyard Bird Book it was a red the book looks like this it goes "My backyard bird book with an image of a red cardinal book crushed." For some reason Facebook did not like that image and it was one of the few books that we could actually run the cover it just like clicked hard i don't know it did we our ad accounts
• 18:00 - 18:30 kept going down we would always get them back 100% of the time but I can't you know it's like how do you get to 100 miles an hour you're like "Oh I get the car up to 60 mph and then it stops." Well now you got to start at zero again so you never get to that scale so a piece of it is is that you actually have to hedge anyway let's go back to the just in terms of the account structure so you want to start by testing the headlines and ad text and
• 18:30 - 19:00 you want to do it with three images um and you just and then in the aggregate you can run a pivot and sort of see how those are split because what you'll find is there'll be one instance of headline one ad text one image one and um you're going to let's assume you have a $100 CPA or 50 bucks a lead whatever it is and you're like "Dude this one's at like 12." And then you're going to find another in and so you're like "Dude this is fire." And then you're going to go
• 19:00 - 19:30 find headline two ad text two image two and like you're also at 12 well is like that a winner so you have to spread the data so when we were doing cover testing on books we would have like one ad set that's at like and we always wanted to be below three we would have one that's at like eight and one that's at like 93 cents i'm like okay but I'm looking for the data in the aggregate particularly over the testing if you had this happen on a scaling campaign kind of a different conversation because it has a bunch of budget but
• 19:30 - 20:00 whatever um yeah and so then we just work we just work through it that way so now you have this like winning thing it's working what else are you doing now that you found that what did I say won applesauce I think so now you have this applesauce campaign and that's doing well is that the best copy you're ever going to get for applesauce no absolutely not like you it is impossible to just have the ultimate winner so then you start thinking okay
• 20:00 - 20:30 what are the angles on applesauce right is it green apples is it red apples is it Fuji apples is it I don't know broadly Washington apples like you can start picking through all of these things and then you start going micro micro micro micro and now all of a sudden you may have a CBO Apple campaign based off of I don't know how many Apple types there are there's probably 50 but like all of a sudden you're running applesauce and you found that 20 work well 10 work well or five work well and so you're spending time diving into that
• 20:30 - 21:00 and you're just like something that matters a lot is guys call it the lead in the copyrightiting world but when they say lead they mean this i'm talking about the hook right so like um what's a what's a good line for apples and it was like "Moms don't give apples to your children be wary of giving apples to your children." Almost think of it as a second headline
• 21:00 - 21:30 because that's the first thing they're going to read that's like "What is this?" And the way that I visually represent it is if I walk up to you on the street and I'm like "Hey I want you to come into my dress shop." You're like "Dude I got to do like whatever we're on vacation." But I come up to you on the street and maybe this isn't a good example but this is still how I visualize it and I like push you not hard i'm be like "Hey man." Like "What?" You're like "Dude we got this dress shop but I got you to stop." Then I walk up and I punch you in the shoulder you're
• 21:30 - 22:00 like "What like do you want to fight like what's going on?" Right versus I walk up to you on the street and I don't punch you in the face i think that's too far but I just slap you across the face you stop the push and the punch the push you're just like what are you doing and you'll keep going goes like some crazy homeless guy right but the slap gets you to stop so that first uh I should come up with a name for it but it's because it's not a lead it's like it's a it's a hook i guess you're
• 22:00 - 22:30 trying to slap that person it actually doesn't have anything to do with the offer you're trying to call out the user quickly and get them to stop so let's let's come up with a controversial um you're like "Women belong in the kitchen." You actually don't have to have the rest of the story talking about that right you could have one that's like kids deserve to be spanked hard you could have um what's another one it was like
• 22:30 - 23:00 uh don't put water in your gas tank i just ran this on a sample Twitter post or whatever last week or this week you're saying that first line doesn't have to be congruent with the rest of the ad text no because you have Yeah so like it's just a stopper a thumb stopper yeah so like most people in marketing and in sales by the way they're trying to go from zero to 100 really quick and so one of my favorite stories is my father used to sell houses after he retired so he didn't lose his mind
• 23:00 - 23:30 going crazy right and I was like "What is your what's your goal?" And he's like "Oh my goal is to um sell a house." I'm like "Yeah I got it how are you going to get there?" He's like "Well I just you know I get them on the phone and then I get a le listing and then I sell the house." And I was like "Okay so three steps are those the only three steps you need?" Well the answer is no you have to have multiple
• 23:30 - 24:00 conversations you have to pull the market comps you have to drive them around let's say so they find a house that they like you have to verify that they actually have the funds so that you don't waste your goddamn time crazy story my dad's had a couple guys that are like literally would fly into Scotsdale and be like we're looking at like three to5 million houses he went under contract with a guy who didn't have the money and then the guy blew out of the contract he never had the money this was
• 24:00 - 24:30 just this game that he was playing he's a crazy person that's like I like being driven around and showing multi-million dollar houses like I'm rich you know um so if you break these things down into elements and then you start finding gains in each of these elements which takes more time but you have more certainty of execution and really now you're in budget preservation asterisk there are times when you're like dude we need traffic and like you just have to like
• 24:30 - 25:00 try to skip as many steps as you can right um anyway so then you just start testing these smaller variables right so like if you're changing the button color on a pre-sale does that matter it doesn't matter until you have a ton of traffic and 2% change in CTR matters right so like I really think through it in that in that testing framework any qu uh any questions on
• 25:00 - 25:30 this so far no I don't think so um it's fascinating honestly um we're talking about testing is done in the in an structure moving it up once you found those winners into CBO essentially in the layman's terms right yeah cbo's for scale and I there was a whole bunch of posts on Twitter um I'm going to mention Twitter a lot because I like took a super deep dive in the last like seven 10 days um but there was like this post that got a bunch of engagement that it's like "Oh you need to test in CBO." Well look I
• 25:30 - 26:00 literally just did it and the CBO found distribution in the in these like one ads and really at the end of the day when you think about what is Facebook's primary goal is it ad revenue it's user engagement because the more people engage on their platform in in theory it creates more ad units which means they have more to sell
• 26:00 - 26:30 right and somebody will buy it the market dynamics will dictate the pricing but if those users don't engage they could charge a million dollar CPM but they have no users nobody's going there right and so the CBO we just didn't really didn't have a lot of distribution in the testing now if I'm running a CBO for scale where all of those are winners well done i like I just decoupled risk and some are better winners than others and so on and so forth what I will say about
• 26:30 - 27:00 those campaigns and the most um um that this was crazy to me actually i didn't even know you could do this we downloaded all the data of a CBO i I did I downloaded all the data of a CBO and I put it in chat GPT and I was like create a graph that shows me the distribution by time um off of and then I would change it headline ad text if I had a bunch the like the y-axis got too long
• 27:00 - 27:30 and what it gave me was a heat map and so when you take a look at the intersection of different things it would create a different color or dark which first the first one was rainbow was terrible So I was like can you just make it all blue and so we could see these different levels of blue and then what I did is I was like great now do an overlay of another metric like CPC and what that allowed us to do was I actually saw this i will share this i just have to find the image or redo it I guess but it was like you actually saw
• 27:30 - 28:00 the distribution of traffic just freaking move across all the different ads but I already knew that they were all winners and that was fascinating to me but if your winner is over a 30-day time horizon like is really down here it takes it a long time to go find the place and that is a waste of budget it's way easier not your money don't care whatever but we don't operate that way and as affiliates at that time we
• 28:00 - 28:30 certainly don't and as an agency now we would just get fired all the time so I think that like the systematic approach of what you're doing and is like think of it systematically i think I just think about this real estate example all the time another example maybe this one works better for you is I was like you're single you want to go to a bar meet a girl and get
• 28:30 - 29:00 married right and you're like or or let's assume for a moment um you want to get laid you're like "I'm going to a bar to get laid." All right what are your what's your chance of success it's called 5% but if you go to a bar and then you got a bunch of phone numbers and then you moved to text message and then you moved to social media and then you moved to coffee and then you went to dinner what's your chance of success going up much higher even using the same
• 29:00 - 29:30 size sample set over here right now you may have again like I said earlier you may have a time constraint separate thing um where you're just like "Yeah we have to drive budget right now like for whatever reason." And so to fill cap or something like this right but but for the most part that systematic thing works really well and then as your big CBOS start stacking those will last a long time because you're constantly pulling
• 29:30 - 30:00 in all these different users at the top of the funnel and effectively letting Facebook remarket them with whatever else they want to see in whatever order Facebook decides they should see those ads um does that make sense yeah no this is great honestly a massive help really cool so when you when you think about this now if I was like "Hey man um go do it." Where are you stuck uh where am I stuck
• 30:00 - 30:30 um I mean probably what I'd do first is go and go back through the ads I've already run and look at what break those down into angles now cuz I've never really thought about them as angles um let's say I'm running which I am like a veterans um campaign so it's targeting veterans with this list of um so I've been running images grainy images of Vietnam War um uh choppers um rescuing soldiers
• 30:30 - 31:00 done well um y military always does well yeah um things like um you'll get like a closeup of someone holding like um it might be a US passport card with a with a photo of ID on it to imply that there's going to be some benefit um something like that um so trying to identify which of the ads that come in at the right price are could be used as angles and then then putting those into test ABOS i guess that would be where I
• 31:00 - 31:30 would start now yeah so like you you want to be able to to step back and let's assume for a moment you're doing veterans are you doing Vietnam veterans are you doing Afghanistan uh and uh and Iraq are you doing um I don't know guys who are 30 years old and got out five years ago right so like that's an example of going deeper right but then when you take which by the way you can take that campaign find that angle and then you have different testing campaigns so at
• 31:30 - 32:00 the campaign level it's very easy to see generally which one's working how about retired firemen how about retired policemen how about retired teachers so I think of veteran and I'm like dude really veteran means that you're basically yes there's the war component but in my head I started thinking of who's on a pension so what are the top 20 categories in the US where like people are on a pension teachers firemen um I
• 32:00 - 32:30 don't know nurses um like you just go down the thing and now you have all those campaigns and so just any aggregate because your testing stuff is generally going to be pretty close like I just don't think you're going to have this huge variant right or you're like I just want I came up with this basic ad copy nothing sexy not that amazing but now I'm going to go take a look at teachers firemen uh you know war vets and nurses who wins somebody will by the way for sure and so then you can just
• 32:30 - 33:00 look at the campaign structure very easily and figure out which one of those works and then you can start dialing in there and then you come up with like different um different uh images and hooks and all this and that right hey if you were if you were and then you could start breaking it down by uh years if you were a fireman between the year if you were a retired fireman between the years of X and Y you're not going to believe this yeah so
• 33:00 - 33:30 would you have a a test for each of those professions for example or would you Yeah so so I I would um but really what I would do is I would test one first and then just see how sort of your copy does right and so if you think about it like it's I tell my team this all the time i was like would you what would you do if it was your own money and so like how would you be efficient with the client's budget now let's assume we have an offer that's $1500 or $200 CPA target they're
• 33:30 - 34:00 like I probably wouldn't run that because like I don't have the budget to test that well you do so you come up with just a general structure and then you're like oh this is interesting so now I'm going to start moving it to professions it's just a staged gain approach very similar to like if if uh war veterans is working and you find that winning um ad text and headline and then you start testing a bunch of different images then you start testing off of those headlines and images then
• 34:00 - 34:30 you take the war veterans you move them to police and fire and nurses and then you run that because if you end up with a CBO and and structure is really up to you like you're at the end of the day you're not really going to solve performance with structure within reason but you may go dude I got a ton of war veteran angles right based on ages and whatever if you were marine if you were in the Navy if you were in the Air Force if you were in the Coast Guard right you could just keep slicing it into pieces now you
• 34:30 - 35:00 end up with a CBO with a huge number of assets around that right and then if you have that you're like "Yeah that's a pretty big CBO so like we're going to put the teachers in another one." Because at some point you're going to end up with something so big it's just a to manage does that make sense yeah yeah definitely um but yeah that's that's kind of how I that's kind of how I think through it okay so in terms of um actual
• 35:00 - 35:30 running testing creatives would you say start small or maybe like three images a day or something like that or what yeah so like let's assume you just started a new campaign okay you just started today you know nothing yeah you have some hunches let's say so I'm going to go with uh war veteran testing campaign whatever if you were a war veteran and you buy this product or fill
• 35:30 - 36:00 out this form then like XYZ might happen or you won't by the way so we I I like blind copy you're going to get a higher CTR you won't believe what rewards are out there for you or what government programs or what these supplements will do whatever it is right blind blind ad copy don't sell them on the copy don't sell a woman on the street in front of a dress a dress store get her into the dress store because now you control that whole environment
• 36:00 - 36:30 no so very bassy on the cockpit yeah it doesn't need to be wildly sexy so what's the thing that matters most um the angle so what's your angle i'm going after war veterans cool let's have a war veteran campaign um and then for me when I think about like headlines I could just come up with two i'll probably write like eight of them because like once I get in the once I get in the zone I'll take that headline and I just I force myself to write it
• 36:30 - 37:00 over and over different ways and then I'll go back and take a look and I was like I really like this one i like this one it's a guess whatever we're going to run these two um depending on if you're running short medium or long um if you're running medium or long I would say just change the lead you don't even need a hook at this point um you you can use one if you have like a tried and true which over time you will but you're just like I'm just going to run this lead and then I'm going to change the lead and the rest is the same so then you just got a bunch of angles up so then you're like well do I put up
• 37:00 - 37:30 20 of them like what if they don't all hit i'm trying to like preserve budget and whatever um you're like well a lot of time should be spent in looking at the data which is what most India buyers don't do they look at this data they don't look at this data and so if you start taking a look at those funnel metrics and figuring out like in a form like what's the one that's like a and it'll be towards the back by the way um but like where's your biggest drop off in the
• 37:30 - 38:00 form i don't know the answer probably depends on the offer maybe it's uh if you're running uh refi let's credit score or whatever right so that's a great one and probably even before social security i don't know i my guess is X percentage of people want to put in their social security number it's a hunch i have no data on this but you're like hey this just isn't converting because not enough people have good or excellent credit and refi but you're going to get way more data than the people who then have to
• 38:00 - 38:30 put in their social security and then go to a ping tree right and so you just get more decisions there and so if you get enough decisions right away you just you can go back and make a decision um and so I don't believe that a media buyer could or should unless you're running a ton of offers but you're just like doing you know three new campaigns every offer every it's too much because as much as I hate it
• 38:30 - 39:00 there's also a function of time otherwise I would just stand up a campaign put 10 grand on it and go "Yeah it hit no it didn't next." M but we know that if you put it at $1,000 and you go to 2,000 you kind of clip the you're more efficient with budget right and so it just comes down to looking at small pieces of that data that is a competitive edge most guys don't do it the vast majority of guys don't do it
• 39:00 - 39:30 because it's hard it's like work you're like who right it's like I go to the gym three times a you know three times a week oh do you like do cardio it's too no cardio who does cardio i'm like the guys that get ripped bro like those are the guys that do cardio right yeah totally so let's So let's go back with some a So I got 10 minutes here so let's go back with some action items so I think the first thing to do is you go back you take a look at your winners or you take
• 39:30 - 40:00 a look at sort of what the control is let's say right you have a control that's working and let's assume you're sitting at a 10% ROI and it's just not that great but like you're at 10% so then you think about what's going to move the needle the most and then how do you start testing against that right and by the way you can have two tests running at the same time you could have a test where you're testing five or 10 images all separate adsets the images are really going to affect your CPMs and CPCs if we have a CPM CPC issue typically it's images right and then you
• 40:00 - 40:30 can take a control image and set up another test where maybe you're just changing the hook or adding a hook and then you could have another one where you're like I'm going to pick one hook or no hook and I'm going to I'm going to test the lead portion of it of the copy the lead of the copy and so now you're just constantly evolving and progressing i hate losing i hate when my clients lose i hate not performing like even though we run on our clients money
• 40:30 - 41:00 this is this is me playing their video game right so you're like "Hey Jason come over to my house let's play uh um Call of Duty." And like I don't know it's your video game but I'm still losing i don't want to lose i don't know if it's the best analogy but you get my point um Okay where are you stuck where would you be blocked i'll ask again
• 41:00 - 41:30 um knowing when to stop I guess because at the moment um I I might be testing 10 to 15 different images a day at the moment so I'm producing them because it's with AI you can blitz through images suddenly you got 30 40 images to test like you really want to throw the the temptation is to throw them all up um and hope that one just stands out straight away um that's the kind of approach I've taken but if you're running each image on control copy at $50 a day like 30 images it's it's a
• 41:30 - 42:00 huge budget and if only if none of them crank or hit suddenly it's it's not looking like a a profitable day so totally so like when I looked at it and where we're going now is like I can literally come up with an angle have 10 versions 10 headlines 10 ad text 10 images in minutes m so I'm like well how am I going to be efficient with that well we're going to do low budgets day one and if I have a 10 by 10 by 10 let's
• 42:00 - 42:30 say even though the adsets are broken down each headline has 10 shots right headline one with ad text one through 10 so in the aggregate I'm getting a huge sample set of data to really figure out which ones are the winners and which ones aren't and so pivot tables are your friend to really just sort of take a look at that you know at that spend but if I ran 30 bucks a day on 10 images and 10 headlines that's three grand a day in testing mhm that got expensive real
• 42:30 - 43:00 right so then you're like do I need 10 this was a recent takeaway from us i was like ah we don't need 10 how many do we actually need to like test um an angle so we're doing five now so five headlines five ad text that gives me 25 i spent 30 bucks a day um what is that that's 900 roughly 900 bucks and so but now I have some confidence because maybe I let it run
• 43:00 - 43:30 two days and so you spend $1,800 but like now I'm really starting to get patterns because you know that winner can take 10 20 50,000 bucks a day so you're you're investing in the future of that what I would also say is make sure the headlines can be a little bit looser they don't because it's a headline they don't have to be consistent but I would say like try to make your um your ad text a little bit more congruent right because if one has a a hook and another one doesn't you
• 43:30 - 44:00 just added complexity like how do you know right and I also think as you test when you're up here and you start coming so you're like I'm testing veterans and then you're like okay cool veterans seem to work now I'm testing Vietnam um a Iraq I don't know Korean War like whatever right uh you just got out of the military or I'll do another one small budgets Navy Coast Guard Marines special forces air
• 44:00 - 44:30 force right and so then you start looking for which ones work and a couple of them will just work better like I don't know maybe mo more people graduate the Coast Guard and go to college so they make more money so their houses are worth more so the refi value is more the credit score is better i don't know um maybe special forces is that way right like I like I just I just don't know um but the the thing the thing to think
• 44:30 - 45:00 through is to be and there is no right way the only the only thing that is the right way is to think methodically and step by step that's the only thing that is the right way how you structure it like after we would train media bar so all of our media buyers will basically train from the ground up over the years and after they kind of got their their legs under them I could look at an ad account with and not pay attention to
• 45:00 - 45:30 the tag and I would know how that account is structured who's running it right cuz you teach your 16-year-old to drive a car and they get better over time and then they go into a race car like do they all drive the same no but I bet at a high school they all had the same driving instructor sure then they had different parents who extended that and you started creating separation of style right there is no right way so I think that's like that's also an important thing we were at this event
• 45:30 - 46:00 in Sun Valley big media buyers everybody runs differently big budgets by the way every guy there was seven figures a month right and so some more I think but um are they like are they wrong no they just do it their way yeah and for me that's since I started media buying maybe like I don't know what it is now three years ago near almost four years now um no one's really had the
• 46:00 - 46:30 same way of buying everyone's had a different way no one's been able to kind of teach me i've kind of picked it up from what other people have told me and then kind of figured it out almost i sort of um um knocked my own style of of launching and running and scaling up together and of course everyone's got the same way um but it's nice like you say to have and this is going to be a huge help to have some sort of methodical pattern to follow so that there's more control and there's more
• 46:30 - 47:00 I've got more of a grip on identifying winners and and testing and that's that's going to be massive for me um so yeah I mean this this call today has been a huge help thank you really appreciate it yeah I have a buddy he has this line he's like "You're one split test away from uh doubling your profit." Yeah or maybe he said quadrupling i don't know but it's totally true i mean one of my favorite stories is we had our first book that scaled to the moon same book Offer Never Change the hook changed
• 47:00 - 47:30 it was Let Your Kids Think About um what was it it was It's Okay to Be Different and we changed it to understanding your emotions and we literally changed that i think I've told you the story before and it just a call yeah and it just dude everything else was the same that was mental dude like literally dude it was nuts that's so crazy can you imagine the feeling so it just went through the roof of stuff that one small change just like changed everything that must been
• 47:30 - 48:00 absolutely crazy to contemplate yeah it um Yeah and then when you think about hooks I think about this all slap the slap the the reader so like I just was doing shoot from the hip some of these like long so so I had three people ask me they're like I sell this product like what kind of copy would you write and so one of them that just went nuts yesterday I wrote it in the car while I was early for a meeting and this guy was like I sell conversational cards like how do you sell conversational cards
• 48:00 - 48:30 like like a card game and um my hook was uh my wife is a drunk and then this the story by the way was actually pretty well written I thought but like um like what's that hook because as they move through the copy if they hit read more your engagement's higher right so you're like "Oh shit." So Facebook sees that that person raised their hand they're like "I'm interested in what they're saying." Right there's a we have a bunch of different copy
• 48:30 - 49:00 strategies around this and they're they're all correct and incorrect like they're all Some guys don't do it i think that's a mistake whatever um we got on this path of not using any emojis in ad text i was like when the did this happen like that works it'll hook the reader back in so anyway man I got to jump you have any questions on this or follow-ups just hit me on Telegram um and I'm I'm happy to I'm happy to help and um I think you're with an amazing team so tell them what I will do yeah
• 49:00 - 49:30 thanks Jason i really appreciate your time um yeah I'll let you know how it goes thanks man all right thanks bye-bye cheers bye