Podcast

Learn from Fast Growing 7-8 Figure Online Retailers and eCommerce Experts

EPISODE 463 52 mins

Scaling Meta Ads: An Advanced Playbook for eCommerce Growth → Marin Ištvanić



About the guests

Marin Istvanic

Kunle Campbell

Marin Istvanic is a partner at Inspire Brand Group, specializing in media buying for eCommerce brands. With over seven years of experience, Marin has grown from a junior media buyer to scaling brands to $30 million in annual revenue. He is also a co-founder of an internal eCommerce brand and an advocate for creative testing and performance optimization.



In this episode  Marin Istvanic shares his journey from studying computer science and playing semi-professional football to becoming a partner at Inspire Brand Group, a media buying agency. He discusses the importance of product market fit, the role of creatives in media buying, and how to craft irresistible offers. Marin also explains the structure of his agency, the significance of testing and scaling ad campaigns, and strategies for optimizing ad performance. In this conversation, Marin Istvanic shares insights on optimizing ad performance through creative strategies, the balance between static images and video, and the importance of creative diversification. He discusses metrics for scaling ads, the significance of cost caps, and best practices for managing underperforming ads. Additionally, Marin provides valuable tips for preparing for Q4, particularly for Black Friday and Cyber Monday, emphasizing the need for proper inventory management and targeted offers.

Timestamps:

  • 02:25 – Marin’s Background
    Marin shares his journey from computer science student and semi-pro footballer in Croatia to becoming a media buyer. His passion for analytics led him to a digital marketing apprenticeship and eventually to Inspire Brand Group.
  • 04:53 – From Junior to Partner
    Marin describes his rise from junior media buyer to head of performance and eventually partner at Inspire Brand Group. He attributes his success to analytical skills, curiosity, and a hunger for learning.
  • 06:28 – Importance of Product-Market Fit
    “Ads are just an amplifier.” Marin stresses the need to have a dialed-in product and offer before launching ad campaigns.
  • 08:34 – Crafting Irresistible Offers
    Marin provides examples of optimizing product bundles and pricing to improve conversion rates and average order value (AOV).
  • 11:05 – Why Marin’s Agency Avoids Creative Production
    Marin explains why Inspire Brand Group focuses exclusively on media buying rather than producing creatives, keeping the agency boutique and specialized.
  • 14:00 – Working with Clients
    The agency’s 20-client cap ensures high-quality service. Marin outlines the geographic spread of their clients and minimum ad spend requirements.
  • 15:47 – Creative Testing Process
    Marin discusses how his team tests ad concepts, focusing on quality over quantity, with a structured approach to scaling winning ads.
  • 19:24 – Evaluating Testing Results
    Metrics like hook rate and hold rate are used to iterate on creatives. Marin emphasizes the importance of scaling only proven concepts.
  • 29:59 – Scaling Campaigns
    Marin explains how his team scales ads using Advantage Shopping Plus and cost caps, ensuring minimal risk while maximizing performance.
  • 41:26 – Full-Funnel Optimization
    Before scaling, Marin evaluates the entire sales funnel to identify potential drop-off points and fix them for better ROAS.
  • 42:03 – Q4 Strategy Insights
    Marin shares tips for Black Friday and Cyber Monday, including inventory planning, aligning targets, and crafting offers like bundles to boost AOV.

Takeaways:

  • Marin transitioned from computer science to digital marketing.
  • He emphasizes the importance of product market fit before running ads.
  • Creatives play a significant role, but they are not the only factor.
  • Nailing the pricing is crucial for making offers appealing.
  • His agency focuses on quality over quantity in creatives.
  • Testing is essential to find winning variations of ads.
  • Facebook ads are based on assumptions that need to be validated.
  • Learning from data is key to improving ad performance.
  • Marin prefers a boutique agency model to maintain quality.
  • Effective ad copy is as important as the visuals. Creative has a significant impact on ad performance, with 60% influence.
  • The effectiveness of static images versus video ads varies by product niche.
  • Creative diversification is crucial for reaching different audience segments.
  • When scaling ads, focus on CPA and ROAS metrics.
  • Consistent results over three days are necessary before increasing budgets.
  • Cost caps help manage ad spending effectively on Facebook.
  • Do not kill ads that are performing well, even in testing phases.
  • Ads should complement a well-optimized landing page for better results.
  • Prepare for Q4 by aligning targets and managing inventory effectively.
  • Be aware of changes in the calendar, such as Cyber Monday falling in December.

 

—-

🔔 Book Announcement:

📈 ‘E-Commerce Growth Strategy’ by Kunle Campbell

 

Exciting news for our listeners! Kunle Campbell, your host and e-commerce expert, has just released his new book: ‘E-Commercer Growth Strategy.’ This essential guide is packed with strategies for attracting shoppers, building community, and retaining customers in the e-comerce space. Drawing on insights from the 2X eCommerce podcast and Kunle’s extensive experience, this book is a must-read for anyone in the e-commerce industry.


SPONSORS:

This episode is brought to you by:

Drive Sales & Reduce Marketing Spending with REVIEWS.io

1️⃣ REVIEWS.io
Drive Sales & Reduce Marketing Spending with REVIEWS.io

REVIEWS.io offers an affordable, all-in-one platform for collecting and displaying customer reviews. It’s easy for you to set up and simple for your customers to leave reviews, helping you boost both trust and conversions from anyone discovering your brand on Google.

With REVIEWS.io, you can increase your Google conversions by up to 26%, improve click-through rates, and lower your ad costs—all with a platform built for busy eCommerce operators like you.

To find out more, head over to Reviews.io
Get 10% off your first year—use the could ‘2x’ when you sign up or click here.
New customers only. Terms and conditions apply.

 2️⃣ Address Validator
Reduce Failed Deliveries & Boost Customer Satisfaction with Address Validator
Address Validator is the go-to tool for ensuring accurate deliveries every time. By validating customer addresses in real time, you can avoid costly failed deliveries, reduce shipping expenses, and provide a seamless checkout experience for your customers.
With Address Validator, you can:

  • Prevent delivery errors and save on reshipping costs.
  • Enhance customer trust with faster, hassle-free deliveries.
  • Automate address validation, freeing up time for your team to focus on growth.

To learn more, visit AddressValidator.com.
Try Address Validator Today. Your first 100 orders are free, with no credit card required—start optimizing your deliveries now.

 

Transcript

Kunle Campbell (00:00.898)
Hey Marin, welcome to the 2x Ecommerce Podcast.

Marin Istvanic (00:04.763)
Thanks for having me. It’s my pleasure to be here after listening to so many episodes of yours.

Kunle Campbell (00:09.696)
Incredible. I like to speak to to listeners of this podcast. So thank you for your support all through the years. Marin, you have a quite an interesting background. Do you want to give us your journey into just media buying and in general, how did you get to doing what you are? You’re now a partner at at an agency, you know,

and it’s called Inspire Brand Group. And it’s been a phenomenal journey. Do you want to share that journey with us, please?

Marin Istvanic (00:47.685)
Yeah, definitely. It is definitely an interesting journey. I was on my senior year of college studying computer science, but alongside I was playing semi-professional football. So basically I was playing in second Croatian division and I was trying to figure out like what to do with my life. Should I go with the computer science? That is like probably the best option at that time in Croatia to become a developer. Should I like go full in on football?

But then somehow I stumbled upon digital marketing. saw an ad for digital marketing apprentice. I started learning everything there is about it. I didn’t know what that was. I went through the whole Google literature to understand like what is an auction, how the ad works. Like it sparked my attention just because I’m very analytical, good with numbers. And that was like way more interesting than what I was studying.

But obviously I didn’t want to just like start running ads for the sake of running ads to learn them. I did not have money at that time. So luckily I had one guy that was doing Facebook ads. He was probably the only guy in Croatia at that time doing Facebook ads. He was a friend of a friend. So I started working with him, you know, like he was giving me those small tasks. Can you find the audience? back, back in the day when the audience was a thing in, in Facebook, can you upload these creatives?

It was great for him because he got someone who could do those like basic tasks. It was great for me because I got the experience to actually see how it works when you’re working with the client. Soon enough, I started freelancing and soon after that, I connected with Matt, founder of the Inspire agency. So I started working as a junior media buyer. Soon enough, I advanced to mid to senior to head of performance. And currently now after seven years,

I’m a partner at the same agency and also partner at one of the internal brands because we realized through the years that we can scale other people’s businesses through Facebook ads. So we thought like, okay, why not scale our own? So currently I’m juggling between the roles of a partner in the agency and internal brand. On the internal brand side, we hit 15 mil in our second year and this year we are on a path to do 30.

Kunle Campbell (02:52.878)
Hmm.

Marin Istvanic (03:07.525)
So this is doing way better than what we expected. So my attention is currently divided between those two.

Kunle Campbell (03:10.637)
Hmm.

Kunle Campbell (03:15.387)
Do you see your brand activity eclipsing what you’re what you’re doing in the media buying or do you think they can coexist sustainably?

Marin Istvanic (03:26.833)
So whoever I talked to, said like, you’ll get to the point where you have to decide which route do you want to go. But my partner is kind of like more leading the brand. I’m more leading the agency. I’m just doing the media buying on the brand side and helping with the growth. So at this point, we are in a good spot that we are not necessarily overlapping, but we are assisting one another. So at this point, I think we are in a good position.

Kunle Campbell (03:53.58)
Yeah, one of the businesses is very high margin and the other is is a low margin business, but big numbers. So it’s it’s a it’s a big decision or a good problem to have. So best of luck with that. You said that. Yeah, so you’re talking about initially your when you were looking for roles, where did you find the initial role?

Marin Istvanic (04:04.731)
Thank you. You explained that perfectly.

Marin Istvanic (04:18.949)
you mean where I started as a freelancer.

Kunle Campbell (04:20.84)
Yes, yes. You said you saw an advert somewhere. I’m just curious to know where you found the advert.

Marin Istvanic (04:24.625)
Yeah, so it was an ad for one of the digital marketing agency here in Croatia in my hometown. It was just like a digital marketing apprentice, someone who would fill the spreadsheets and do those basic tasks. But I never before heard about that role, so that’s what got my attention. But then actually my first job as a freelancer, I got on Upwork.

Kunle Campbell (04:53.888)
Okay. It’s it’s phenomenal that you you’re able to move from junior to to head of eventually head of performance and then a partner. What is it in you? Do you think Matt saw and continue to reward you know, in terms of responsibility and value over the years over the seven years?

Marin Istvanic (05:15.257)
Yes, that’s great question. So I would say that like I’m born for this job. Like if I go through, you know, like some personality test, I would get results that like that are suitable for this job. Very responsible, detail oriented, great with numbers, curious, trying to find some hacks, you know, to beat the system. So basically everything that I’m…

that is describing my character is like great personality trait for this job. So I think that’s one thing. The other thing I saw this opportunity and I saw that I’m enjoying this job. don’t see it as a work. It’s a pleasure to me. And anytime that I have a free time, I do some extra. I’m listening to many podcasts. I went through the whole Google literature support. I’m engaging.

with other people, I’m just like constantly learning. So that was probably a combination of my character and hunger to learn more because I actually find something that doesn’t feel like a work to me.

Kunle Campbell (06:28.438)
interesting, super interesting. Okay, I want you to either disagree or agree with this statement. 80 % of media buying is in creatives in for matter.

Marin Istvanic (06:43.556)
I would disagree because Facebook told us that 56 % is the impact of the creative to the performance. It definitely has a huge part, but I would not say 80%.

Kunle Campbell (06:55.342)
Let’s jump into it. So why? Why do you think? So what are the other bits outside of creatives that really matter? 56 is a big number. We’re talking 44 % now. What’s in that 44 %?

Marin Istvanic (07:12.059)
So let’s go a step back. So for me, it doesn’t make sense to run the ads until you have a product market fit and until you have a proven offer. Because although I’m a media buyer, for me, ads are just an amplifier. They’re like a fuel on a fire when you have something that’s already working, something that’s dialed in. So I would say that structure is product, offer, then the creative, and then the…

stuff in the ad account, just by technical stuff, you can improve the results maybe just by 20, 30%, depending on how bad the ad account in what shape it is. But then actually I had so many ad accounts started performing better just when we like switch the offer, when we find a performing good performing angle. So I would say that

In between the creative and the technical part, you need to sort your offer and you need to sort your angle.

Kunle Campbell (08:13.036)
your offer on your angle. Should we speak to this offer? And what is a breakdown? What do you see brands that are really killing it with with you working with you do to really nail that offer and get the product market fit?

Marin Istvanic (08:34.353)
So, once you have a product market fit, you know have a product that people want. It’s now how to make it appealing enough that they would pay the price that you put. So, first thing is, okay, you need to nail the pricing or you need to justify your pricing in a way that it is appealing to the users. So, for example, I can speak about one of the tests that we did. We had one month bundle, two month bundle, and three month bundle.

And basically the price anchored that 99 % of people bought three month bundle because the difference between two month and three month was just $10. So nobody in the right mind would buy two month bundle. And then we thought like, okay, this is crushing. So then what we did then is that we, increased the prices and then we said, if you buy two month bundle, you will get additional two month for free. So basically we did a BOGO, but for two month.

And then everybody switched to buying two month bundle. Although our margins stay the same, our AOV stayed the same because we increased the prices, but our conversion rate went up. So that means that we were actually earning more. So just by two tweaks, my raw Asindhiad account for the same ads in the same ad account structure increased 30 % just because we had an offer change that was converting way more. So for people that was that like, you want to find that something that

looks like an irresistible offer that people when they see that like they say like hey this is a great deal I don’t want to miss that.

Kunle Campbell (10:09.518)
It makes a lot of sense. So in this experiment that you’re talking about, it sounds to me like this offer bonding was, these were like on-site offers. So these are landing page tweaks as against like meta account tweaks. Is that correct? Okay. Okay. The reason I asked the initial question around the 80 % is because

Marin Istvanic (10:23.664)
Yeah.

Marin Istvanic (10:29.261)
Exactly, exactly.

Kunle Campbell (10:38.102)
I know for a fact that your agency, you focus exclusively on the media buying aspect of things and you essentially work with your supplied creatives and then you give feedback. Do you want to give us a synopsis of how you do this and why you choose to not serve, you know, play around with creatives from an agency perspective?

Marin Istvanic (11:05.135)
Yes, I thought about this extensively and multiple times and chatted with my partners and with some other agencies. They all said, when you start doing creatives, it’s basically like you have another business. So you need to like at least four more people, need to create the strategies, copywriter, video editor, graphic designer, maybe project manager, someone who’s going to source the creatives. And like, that’s another, that’s totally another business. And then like we are starting from scratch.

We need to build our processes. We need to find proper people. We need to dial in everything that other creative agencies, are doing for the couple of years already that they have proven track of records. That’s one thing. The other thing is I want my agency to stay boutique. I don’t want to add additional services. I don’t want to add additional people because all of that is…

like additional headache, I’m currently in that spot where we have limited number of clients. I know that I can pick who I want to work with. I know that we can deliver great results with our media buying. So at the moment, it’s like a perfect mix between the, how do I say it, results and the number of clients and not to be overwhelmed and not to stress too much and still

earning well enough.

Kunle Campbell (12:35.852)
makes sense. you have how many clients and where are they located? I know for sure that all your clients, the reason why I really wanted to catch up with you on this podcast is because all your clients are ecommerce. So how many clients do you have? And how are they geographically dispersed?

Marin Istvanic (12:55.825)
Yeah, so we have a limited spot of 20 clients. Obviously, let’s say in Q4, we can onboard one or two extra. Also, sometimes it depends what the size of the client is, but usually, as I said, three media buyers plus me, and there’s 20 clients. So basically, each of media buyers can handle between four and let’s say six or seven ad accounts. Most of the clients are based US, Canada.

and then I would say some are based in Australia and UK. So basically, big four English speaking.

Kunle Campbell (13:30.542)
Absolutely. It’s the same kind of breakdown for the listener base of this podcast. We’ve reached 150 countries. But yeah, over 50 % of our audience is United States and then it trickles to all of the other Jews who talk to us in English speaking. Okay. Now, from a client perspective, how much do clients, your 20 clients on average typically sort of spend with you guys?

Marin Istvanic (14:00.753)
So basically the minimum they need to be spending is 50K a month. That’s kind of like the minimum minimum when we onboard something, but we have some clients that are spending 1 mil, 1.5 mil a month. So average is, I would say around like 10 to 15K a day, which turns out to be like 300 to 500K a month.

Kunle Campbell (14:29.006)
Okay, so if I’m spending 300k a month on Google on Meta ads, how many creatives must I send your agency in order for you to sort of deliver my target ROAS or CPA I have in mind?

Marin Istvanic (14:48.273)
So I’m always the guy that goes for quality above quantity. So I would say like minimum is, let’s say, four concepts a week, each concept with at least three variation. Because when you build a video, just like having a different hook or a different length or a different thumbnail in a video or different layout or a text message in the image,

It maximizes our chances of finding a meaning creative. So when I say four concept, that’s like a unique concept, but we need each concept with at least three variation to maximize our chances of finding a meaning creative.

Kunle Campbell (15:28.622)
Got it, got it, got it, got it. So four creators, four concepts, and then three variations each to do it. And then you can give feedback if one variation starts to work really well, so you could further iterate on it. Okay, so that would mean about four times four, so 16, 20 creatives a month. Okay, that makes sense. And then what is the process from handing over this creative?

Marin Istvanic (15:42.851)
Exactly.

Marin Istvanic (15:47.569)
16.

Yeah.

Kunle Campbell (15:56.63)
to yourselves? What do you then sort of do? Where are you tweaking? What are you doing in the ad account to deliver my target CPA or my target ROAS?

Marin Istvanic (16:11.441)
Yeah, so basically client sends us a link, say, Hey guys, this is ready. We download it, we upload it, and we have our sandbox testing campaign. Basically we are doing all of our testing in an ABO, which means that each new ad set is a new test and each new test is a new concept. So basically if I get four concepts that week, I would have four different ad sets. In each ad set, I would have only variations of that concept.

because my goal is not to find one ultimate winning ad. My goal is to find a winning variation within the concept. And then whatever works in my testing, I can keep scaling. So although my testing campaign is labeled as testing, I have multiple ad accounts where my testing campaign is actually my highest spending campaign, just because I increased the budget there. And then, okay, whatever works, I would take the post idea of a winning creative and I would put it into additional campaign, which is…

scaling campaign, which is Advantage Shopping Plus with a cost cap. So I’m not a big fan of using cost cap for testing, but I’m a huge fan of using it for scaling. This way we are taking only proven winners and we are setting a cap with a high inflated budget, which we are okay Facebook spending. And there’s a very, very low chance Facebook delivering bad results because first of all, I pick up a target CPA that I’m comfortable with.

And the second of all, even if Facebook like overspends, it cannot overspend to the extent it would be terrible just because I graduated only winning ads there.

Kunle Campbell (17:49.46)
There’s a lot to unpack from there. Let’s start out from the first, which is AVO. Do you want to give listeners just to break down that acronym, AVO? A-B-O, B-O, A-B-O.

Marin Istvanic (18:01.337)
Yeah, so basically that’s budget optimization. That means for each of the asset, we are dedicating a specific budget that we want to spend. So basically if I launch four tests and I put $100 on each of the tests, I would spend $100 on a daily basis for each of the tests. That’s a bit different compared to the CBO, which is campaign budget optimization, where you put the budget

And when you put like, let’s say those $400 on a campaign level, and then Facebook decide on which concept, on which ad set it would spend. So it could be happening that concept number one gets $350, concept number two gets $50 and concept number three and four don’t get any spend. I’m not a fan of that because Facebook can often push the budget towards the ad that has the highest click rate, ad that has the, maybe let’s say click bait hook.

ad that has the longest watch time, ad that has the biggest, that’s targeting the biggest country. So all of those signs, indicators are not necessarily correlated to the performance. And in ABO, I can control my budget over the performance and I can adjust my budget accordingly. That actually is maximizing efficiency of my ad account.

Kunle Campbell (19:24.406)
Makes sense, makes sense, makes sense. So back to the testing, the sandbox testing phase, how much of reach or impressions do you want to reach in order to make a statistically significant decision as to whether to go for an ad concept or to just can it?

Marin Istvanic (19:46.737)
Yeah. So it’s very hard to wait for a statistical significant, especially if we have multiple tests running. have some clients where we are testing, you know, 15 concepts a week. And if I start with $100 and I want to reach statistical significant, like, and most of the tests are failing, I will get into position that my ad account is struggling. So there’s no fixed level of spend or level of impression that we need to reach.

It depends on the AOE, on the number of tests that I’m running and the overall ad account budget on a daily basis. So if I’m spending, you know, like 2K and I’m launching $400 on a daily basis, that’s 20 % of the budget that goes to for testing. I could be overwhelming the ad account. But on the other side, if I’m spending 10K a day and then I launch 10 tests, that’s just 10%. So basically I can allow myself to be more aggressive.

and I can allow some of the tests to run longer. Also, some of the tests might look bad on the ad set level, but when you enter the ad level, you could see that there’s like one ad that got most of the spend, but bad results. So we would kill that one and let other running. And then after, let’s say, I don’t touch like anything for three days, but after three days I evaluate, okay, is there any variation that is performing bad? And then like, let’s say after five and seven days, I can actually…

make a decision whether it makes sense for me to kill the concept completely. If it’s performing, then I can scale it. So I get to the point that, let’s say out of 10 tests, I turn off four, I keep two at the same budget, and then I increase budget on the remaining ones. So it’s always a mix. And the budget and the performance is evaluated always based on the ROS or CPA.

Kunle Campbell (21:41.292)
I like that. You know, there’s a broad concept of heuristic that says like, to optimize, you need to see what you can cut off, you know, take away rather than add. And from your concepts in this respect, you’re seeing what can be taken off initially, what’s really weighing everything down, and then you’re putting more concentration into what has prospects to be potentially graduated to the skating campaigns.

Marin Istvanic (22:09.891)
Exactly. Cause like for me, Facebook ads are just a set of assumptions. You have an assumption, you create the ad around that assumption. You have multiple variations. Then you get the data. That’s the beauty of digital marketing. You get the data and then you double down on what’s working and you discard on what’s not working.

And then if you’re constantly applying learnings, what angle is working? If the older creator is working better than the younger creator. If when we mentioning the product is working better than we like opening up with a problem when we are showcasing the product compared to when we are having a long intro about the problem. Like you get so many learnings when you stack one on another, you get to the point that you actually are getting better with every ad that you’re running and you’re actually, which is

actually allowing you to spend more with the better results. That’s how you actually scale.

Kunle Campbell (23:08.504)
What’s your take away on ad copy? So we have the creatives, which could be image or video. Correct. What’s your agency’s approach to headline text, know, text in the description and then, you know, ad copy. Do you, I’ll leave it to you.

Marin Istvanic (23:30.351)
Yeah, so we are pretty much not testing that at all because Facebook told us that creative has almost 60 % impact on the performance of the ad. So we want to have the most focus on that. We usually just have an ad copy that’s aligned to the angle of the video. So let’s say if you have a video that’s talking about anti-aging, we would have an ad copy and the headline that matches that angle. If you have a video that’s talking about anti-acne,

or like video that’s talking about like not wearing a makeup, we would have the ad copy that’s aligned to that angle. From time to time, we would challenge that ad copy with existing winning videos. But in most cases is like, I would rather test hook. I would rather test thumbnail. I would rather test length. I would rather test creator in a video compared to the ad copy.

Kunle Campbell (24:19.754)
Okay, okay. And then what’s your take on static images versus video? What’s what’s a what’s a good ratio to aim for?

Marin Istvanic (24:29.275)
So again, there’s no clear answer. It depends what your product niche is. For example, images work amazing for clothing, shoes, jewelry, basically where you can understand what the product is just with an image. So for clothing, some of the best performing ads are carousel ads, catalog ads, you know, like you just see an image because I know like my girlfriend, she likes to swipe through those like…

like a lot of the clothing combination on that carousel ad. But for the other products where you actually need product explanation, product demonstration, identifying the problem, it’s very hard to sell it with an image. That’s one thing. The other thing is no matter how good image ad you create, most likely it would be doing retargeting.

So Facebook is pretty good at like allocating type of the ad to the funnel, to the part of the funnel where the user is. So like I can see by percentage of new visits in Orbeam and in Triple Whale that no matter how good image ad I create, it does retargeting. Also, for example, if you create an image ad that says like 50 % off and you launch that with a broad audience and with all the possible exclusions of your buyers, of your…

video watchers, of your Facebook, Instagram, Engagers, of your website visitors, basically you’re instructing Facebook to go after like completely new people, but you have an ad that says 50 % off, that ad would like 100 % do retargeting. And I can see that my percentage of new visits, no matter the full exclusion, is just 20%. So sometimes no matter like how good image ad you have, cannot reach cold audience. What I would say, creative diversification is extremely important.

So you want to have videos, want to have UGC, you want to have unboxing, you want to have reviews, you want to have image offer ads, you want to have us versus them. So you want to like cast a wide net of your creatives because that’s how you’re reaching different pockets of the audience. And then obviously you double down on what’s working. So I would say it’s definitely a thing to have a mix of video and images, but don’t rely too much on the images because you’ll get to the point that…

Marin Istvanic (26:51.427)
all of your ad account is doing just retargeting.

Kunle Campbell (26:54.37)
Yeah, and I guess this creative diversification is something you sort of instruct or guide your clients through, particularly if they are initially skewed towards a certain creative type.

Marin Istvanic (27:07.085)
Yeah, exactly. We do analysis to understand what’s the split, what kind of formats are working. And we have a document called Ads Bible with the styles that we know are performing. Let’s say founder ads, unboxing, us versus them review, mini VSL and stuff like that. Basically our formats that we know have the highest chance of getting best results.

Kunle Campbell (27:37.454)
Okay, that makes a lot of sense. Okay, so let’s talk about the next step. So an ad has clearly shown prospects. What are you looking at? Are you looking strictly at conversions in the testing sandbox stage?

Marin Istvanic (27:54.245)
So if I’m making decision whether I will scale an ad set or not, yes, I’m always looking at the CPA or ROS. If I’m looking how to potentially iterate on a video, then I’m looking at soft metrics, which are like early indicators. So if I see an ad is crushing, okay, let’s increase the budget, but I want to replicate it potentially. So, okay, let me check for the opportunities on which metrics I should iterate. For example,

Let’s say my hook rate is not great, but my hold rate is amazing. So I would check that video. I would try to understand, okay, why this video is keeping the attention of the user, why they are interested in watching more. And then my hook rate is below the average. So that means I can make even better winner, winning ad if I improve the hook. So I want to understand where’s my opportunity. And I’m iterating only when I have discrepancy. So if I have bad hook rate, bad hold rate,

Probably that video would not be a winner and there’s no point for me to iterating on that. If I have a winning ad, I would analyze this metric to try to improve it. Also, if I have an outliner, let’s say a video with an amazing thumb stop ratio, but pretty bad hold rate, I know there’s potential opportunity to improve that video and become and create additional winner from that video, then I would iterate on that part where it’s lacking. So to summarize, to make a decision based on

optimization and whether I will kill something or increase the budget, I’m looking at the ROAS. When I’m looking at the iterated creatives, would look at thumbstop ratio, hold rate, watch time and click-through rate.

Kunle Campbell (29:35.992)
Got it, got it, got it. And the graduation phase, do you want to, what’s your approach? Do you just scale with the specific number in mind from a budget perspective? What are the metrics? What are the things need to be made? Tweaks need to be made in order to, to scale, you know, creative.

Marin Istvanic (29:59.739)
So I’m not a fan of making a lot of daily tweaks in the budget because that turns out to be just a day trading. So in order to scale, I need the consistent results, at least three days of a good results at my target CPA. If those both conditions are met, then I would increase the budget. How I would increase the budget depending on what budget level we are.

So for example, if I’m on $50, I can bump it to 100. But if I’m on $500, I cannot bump it to 1K because that would be overwhelming. So then I need to go more gradual, let’s say from 500 to 700, from 700 to 950 or stuff like that. That’s one thing scaling in the testing campaign. And the other thing, which I call graduating, is actually taking the ad account, taking the ad post ID and putting it to different campaign that advantage shopping plus.

That ensures that all the comments and likes that social proof is spread. So when I graduate the winner, it’s not starting from scratch. It’s already a proven ad with a lot of comments, which Facebook is favorizing, which users are favorizing. So that ad has like say a jumpstart in terms of the performance. And when I graduate five ads, which are all winning ads, there’s a very low chance that that campaign would not perform just because whichever the ad Facebook picks.

That’s a good ad. That’s a proven winner.

Kunle Campbell (31:27.01)
And is it CBO you’re using now at this stage of scale?

Marin Istvanic (31:30.701)
No, it’s Advantage Shopping Plus. That’s not really new, but the last campaign time that Facebook introduced two years ago, which is kind of more simplified version. cannot pick gender, you cannot pick age range. You just have one ad set and you pick exclusions of your audience and you pick country and you put ads there and that’s it.

Kunle Campbell (31:59.299)
all your graduation campaigns on Advantage Plus now.

Marin Istvanic (32:04.347)
pretty much, I would say, ABO testing and then scaling Advantage Shopping Plus with Cost Caps.

Kunle Campbell (32:10.186)
or cost caps. Okay, do you want to explain cost caps to the audience? I know it’s been explained in a few other episodes, but it’d be great to just explain to recap what’s, you know, cost caps up to the audience, please.

Marin Istvanic (32:22.545)
Yeah, so basically cost cap or in other words is cost per results. So you’re defining Facebook a number that you don’t want it to potentially spend if it cannot hit that number. So let’s say we are selling something for $200 and our target price is two. So then we tell Facebook, Hey, Facebook, I’m going to put the cost cap of $100. And if you cannot hit cost per purchase for $100, do not spend.

And then interior, if Facebook would not spend, if you cannot hit, but sometimes it doesn’t work that way. It works like on a seven day period. some one days it would spend 120. It would get you a CPA 120. The other day it would get you 90. The third day it would get 105, but it would average out on 100. For example, we see that during the weekends conversion rate is always higher. All of my clients have better performance on weekends.

So if you leave the cost cap same, let’s say, and you put 10K budget, on Thursday it would spend five, on Friday it would spend five, but let’s say on Saturday it would spend nine, on Sunday it would spend 10. So Facebook kind of like adjusts the budget and adjusts the pacing to the estimated action rate. And then basically on an average you get that, you get the…

CPA of that target that you put as a cost cap.

Kunle Campbell (33:53.516)
Makes sense, makes sense, makes sense. Is there any other thing we’ve not covered in this scaling phase? That’s very important.

Marin Istvanic (34:03.933)
Not really. That’s pretty much it. I would just want to highlight once again that I’m scaling all of my tests, no matter if the campaign is labeled as testing. And I’m not turning off any ad that I graduated into scaling campaign. So a lot of times I would see someone like has a good ad with that set in testing and they graduate the winner, but then they killed the ad set in testing. It just doesn’t make sense just because you don’t want to compromise whatever is working in testing. That’s one thing.

And the other thing is you still want to scale the winners in testing, no matter if it’s labels as testing. So those are like two main rules. And third rule is like, don’t be too aggressive on testing. Some guys would like get, you know, two sales after like one day and they’re like, they would immediately increase the budget. Like it could be just a fluke. So you want like at least three days of data before you make any adjustments.

Kunle Campbell (35:02.808)
So you’re saying that in the test campaigns, if something’s working, just leave it.

Marin Istvanic (35:10.211)
Yeah, yeah, don’t ever kill whatever has good results.

Kunle Campbell (35:14.648)
Okay, so, but what’s the benefit of that if you’re moving it, you’re graduating it to a scaling campaign and you’re throwing more advertising dollars to it or at it?

Marin Istvanic (35:24.625)
So actually Facebook told us that ABO and Advantage Shopping Plus are slightly different algorithms and that the overlap should not be more than 30%. So basically, even though I’m scaling the same ad in two different campaigns, they’re not necessarily hitting the same people. And also even if they are, I care more about the performance than about the overlap.

So I would rather like scale from 5K to 15K and stay and have good results and then have, I don’t know, 60 % overlap, then like be on 5K and then have 20 % overlap. Just because all of my clients care about the performance and that’s how I’m making my decisions.

Kunle Campbell (36:08.3)
Makes sense because I was just have the impression that you’d spend in the in the in the in the sandbox test is typically a few hundred dollars, I guess, a month a day rather.

Marin Istvanic (36:21.381)
it depends how many tests I have. I have one ad account where I have currently like we’re testing 10 concepts a week and I have probably 20 ad sets active and some of the tests I launched all of them with 100. I have some ad sets that are all like 2K that are like on 700. So depending how good the test is in testing campaign, I’m scaling it there. So in that ad account, yeah.

Kunle Campbell (36:47.338)
in the testing.

Marin Istvanic (36:49.475)
So basically in some ad accounts, my testing campaign is actually my highest spending campaign just because I scale whatever is working there.

Kunle Campbell (36:57.262)
Okay, okay, okay. Okay, that makes sense. That makes sense. Makes sense. And with with all of that, when do you start to sort of pick up on the signs to start to cut down? Sometimes you do not necessarily need do you do you cut down a campaign over time? Or do you just stop it? Do you cut down budget rather in a campaign? Or do you just

Marin Istvanic (37:23.833)
Yeah, I cut budgets if the performance again is not good at the selected period. So if the performance is bad at one day, I would not act on it. But if the performance is bad for two, three days, I would, okay, see what’s happened, like what changed, because there are only four metrics that are impacting your ROS. CPM, click-through rate, AOV, and conversion rate. So when performance is down, one of those changed.

So for example, if I see that click rate started dropping, okay, there’s something with the creative. So, okay, let me check if I can actually like resurrect the ad. I call that resurrection and the handbrake. Basically, I would pull up the budget down because maybe I scaled it too aggressively. And at this current level, Facebook cannot get me the target CPA or target roles that I need. That’s one thing. And the other thing is I would duplicate that ad.

give it a fresh chance that they call a resurrection. Because if something worked for a month and a half, and one ad and the audience combination worked for a month and a half, and I managed to scale it from 100 to 2K, that single ad set. And now magically five days are bad. Okay, it tells me there’s something outside of my control that happened. So I would kill that ad, I would duplicate it to give a fresh chance. And in like…

probably 70 to 80 % of the cases, that ad just starts picking up with the performance again. And I get to the point that like this is working good. So I would not necessarily turn off something when it’s bad right away, if I manage to scale it.

Kunle Campbell (39:03.854)
Okay, okay, that makes sense. you give it another chance, you duplicate it, see where it goes. And do you if it goes the right way, that’s fine. If it doesn’t, would you just gradually start to cut it? Would you kill it completely? Or would you say, okay, let me cut this 2k to 1k or to 500 and see where

Marin Istvanic (39:07.203)
Exactly.

Marin Istvanic (39:14.799)
then I would probably kill it. Yeah.

Marin Istvanic (39:22.415)
Yeah, so basically that handbrake is when I first pull up the budget. So let’s say we were on 2K and then like, okay, it’s not working as intended. Okay, let’s pull up the budget to let’s say 1.5K and let’s duplicate the ad. If none of those help, then I would probably kill it.

Kunle Campbell (39:38.254)
You just kill it. Okay. Makes a lot of sense. And speaking about things outside of your control and circling back to your first point around offer tweaking.

Kunle Campbell (39:53.624)
When, how much agency do you have as a media buyer to say, okay, this landing page needs to be adjusted in order for my target ROAS to be realized or my target CPA to be realized?

Marin Istvanic (40:12.665)
Yeah, so I have a funnel calculator that I built in Google Sheets. It basically tells me where in the funnel brand has drops. So, we have a number of clicks, then number of landing page views, which tells me, okay, whether the page is loading fast enough. Then we have number of landing page views that proceed to view content. So basically, to what extent people are interested in learning more about the product.

And we have percentage that tells us what percentage of people that landed on a product page actually added to cart and what percentage of people that added to cart initiated checkout. So before I start with every client, I do this analysis to tell the clients, hey, like you have gap here, like people are not even adding to product. So you need to fix that to order to get that better Ross. Or for example, average of people that proceed from checkout to purchases.

is 50%, but on your side it is 25. So, okay, let’s check what’s happening there. Maybe you’re charging shipping that people are not expecting. Maybe they have some objections. So, okay, we want to sort that before we actually start by our media and before we start scaling because otherwise it would be very hard if you don’t have your funnel dialed in. That’s why I said ads are just a cherry on the top.

Kunle Campbell (41:26.307)
like.

I agree. I agree. I agree. That analysis is amazing. If you can share it with members of the audience, please do. I’ll add it to the show notes in this episode. What’s your take on… So we’re recording this just for some context at the end of October 2024. What’s outlook for Q4 of…

Marin Istvanic (41:38.139)
for sure.

Kunle Campbell (41:58.328)
we’re in Q4 already but for Black Friday Cyber Monday of 2024.

Marin Istvanic (42:03.451)
We could do a whole episode just on that, but I’ll try to make it short and meaningful. So basically first Q4 is a period where brands could have up to 60 % of their revenue just in those three months, especially if you’re a gifting product. So I would say align your targets properly. For example, for our internal brands, we decided that we want to lower our targets for October.

because we want to fill up the funnel for November because from the last year we have data that a lot of people converted during that period. So we intentionally lowered our targets and we were running like just on a break even because we know those people will convert. That’s first thing, understand your number. The second thing is make sure that you have inventory that you can actually support the scale during the Black Friday. Thing number three, make sure your ad account is not capped.

So you don’t get into position that you’re scaling, but you just magically cannot spend more than 10K a day or like 5K a day, which because Facebook put the cap. So that’s going to hurt big time. The fourth thing is define your offer and define it based on your margins. Okay. Do you want to have a discount, which is the easiest to implement, but it your margin the most. Do you want to have a free gift with purchase that has a high perceived value, but it’s not going to necessarily convert?

as much as discount. Do you want to have some aggressive offers like buy two, get two, buy three, get three? Do you want to make bundles, which are my favorite way for Q4 because they allow you to increase your AOV. So even if you give them a discount, you’re still not hurting your margins too much. And people have like a perceived value that this discount of 20 % on $300, that’s $60 compared to let’s say 20 % on

$50, which is just $10. So basically that perceived value for discount is way bigger on the bundle and it protects your margins. And the fourth thing is make sure to like align your start date and end date when you’re going to launch the offer. And for this year, just a small tip, be mindful that this year, Cyber Monday is in December. So if you’re making forecasted revenue for your Q4 for your November,

Marin Istvanic (44:30.563)
just make sure that you do not take last year data because last year Cyber Monday was in November and this year it is in December.

Kunle Campbell (44:41.211)
I think Shopify does a really good job now at that, but not all analytics or grinding, know, analysis rather is, is done on the Shopify dashboard. want to do that, use other platforms to do that. Marin, it’s been an absolute pleasure having you. Thank you for the final set, set of tips for people who want to find out more, you know, about Inspire. It’s Inspire, Inspired.

Digital group dot com. That’s where to get you are linked to it in the show notes. Are you active on any social media platforms?

Marin Istvanic (45:18.415)
Yeah, pretty much I’m most active on Twitter and I’m sharing this like tips and tricks on a daily basis. And in the past year, like there’s not a single day that I did not tweet it, something that’s going to be useful for you to run the ad account. And recently I started posting on YouTube. So those are probably two places where you can find me and learn more about me, how I work and actually…

get some actionable insight that you can implement in your ad account right away.

Kunle Campbell (45:52.564)
It’s been an absolute pleasure having you on the podcast Marin. I’m going to link to your ex as well as your YouTube. All I can say is I have extreme gratitude for this hour. Thank you very much.

Marin Istvanic (46:07.537)
Thank you for having me once again.

Kunle Campbell (46:11.726)
Okay, let’s get, sorry, I just went to pause it and then we’ll take it from there.

About the host:

Kunle Campbell

An ecommerce advisor to ambitious, agile online retailers and funded ecommerce startups seeking exponentially sales growth through scalable customer acquisition, retention, conversion optimisation, product/market fit optimisation and customer referrals.

Learn from eCommerce Entrepreneurs & Marketing Experts


Get Free Email Updates by Signing Up Below:

Podcasts you might like