Thomas Petit on ASO, conversion optimization, and why VO and tROAS don’t really work for subscription apps
What do you learn if you advise more than 20 apps on growth strategies? Apparently, quite a lot, judging by Thomas Petit.
Thomas Petit is a mobile growth consultant for Deezer, Lingokids, Playspace, AppAgent, and many more companies. He’s spoken at App Growth Summit, App Promotion Summit, Mobile Growth Europe, ASO Conference, and that’s just the first part of the list. And he’s a frequent contributor to MobileDevMemo.
Somehow, he also finds time to enjoy life in Barcelona, Spain.
You can listen to the podcast here, or keep scrolling for a full transcript.
You can subscribe on Anchor right now, but Growth Masterminds is currently propagating to pretty much every other podcast platform on the planet. It’s also already available on Spotify, Google Podcasts, and Apple Podcasts.
But if you prefer to read your information rather than hear it, here are the highlights and a full transcript, lightly edited for readability.
ASO is now conversation optimization
“And this is actually super critical because while the other changes only touch organic, this will apply to everything, because every single click that you pay on Facebook, Google, any other network, would actually have to pass through this page.
And there’s a pretty big drop of anywhere between 50 and 90% so this is actually really critical for making UA budget profitable.”
Getting featured: no longer such a big deal?
I don’t know if people are not acting from the featuring tab or it’s just there’s more featuring, but this was quite concerning today.
I actually remember a conversation we had before the store was revamped by Apple and in the industry we all thought that featuring would get actually bigger in impact. And what we’ve seen is that it actually got smaller over time, so that’s definitely a reality.
Why mobile marketing data is getting more complex
Obviously the more you go down the funnel the more the data becomes complex because you’re looking at specific segments and you’re looking at more and more events that are only specific to your app and are not really something that is a benchmark. And especially when you try to start modelizing a Lifetime Value of your users. Maybe you’re using a combination of ads and IAP. Maybe you’re using a very complex scheme of subscription with different lengths, different value, different … also you need to factor the currencies and so on.
So it’s really like the more you go down the more there is a huge quantity of data, but also a more tricky way to interpret the data.
Subscribe to the podcast … and check out the full transcript
And … here’s the full transcript, lightly edited for length and legibility.
John Koetsier: Hello and welcome to Growth Masterminds, the podcast where we give you insights from experts to help smart mobile marketers get even smarter. This is just the third episode. For the first episode we went to the United States. For the second, we were in Australia. And the third, we’re in Europe, specifically Spain. My next guest is a mobile growth consultant for Deezer, Lingokids, Playspace, AppAgent, and many more companies. He’s spoken at App Growth Summit, App Promotion Summit, Mobile Growth Europe, ASO Conference and more.
(If you can sense a thread that’s running through those conferences you’re on the right podcast.)
He’s also a frequent contributor at Mobile Dev Memo and a prolific tweeter at @Thomasbcn. He’s one of the smartest people I know in user acquisition, mobile growth, and digital marketing. Thomas Petit … please say hello!
Thomas Petit: Hi everyone. Thank you John for this very nice intro. Happy to participate in the mobile masterminds today and share a few stories.
John Koetsier: Wonderful, wonderful. So you’re joining from Spain but you’re not Spanish by the sound of your voice. Tell us a little bit about yourself.
Thomas Petit: Yeah so I’m French. This is sometimes really hard to hide obviously, especially on voice only.
So I was born in France. I built a business there that didn’t work. And just after I moved to Spain, that was about 10 years ago, I started on web marketing, mostly eCommerce and similar before I moved to apps about five years ago. Which doesn’t seem like that much, but in terms of mobile timeline that’s like I’m almost a veteran with just five years of experience in the field as it’s always changing.
I spent a few years dedicating my time to one app in the health and fitness space. And I moved a bit more than a year ago to independence, where I wanted to see different verticals, different stages… so I’m collaborating with about 20 apps at the moment and this is very exciting and I’m learning things every day.
John Koetsier: Wonderful, wonderful. You’re living the dream, living in Spain, beautiful weather, probably still working on all the favorite apps and everything. It has to be super interesting.
So you’re working on 20 different apps at once. That’s really neat because you get a very broad sense of the industry. It’s got to be challenging because you get to dive into different apps with different models and different user acquisition needs and everything like that all the time, but you must see a wide range of marketing challenges, marketing innovations, user retention and acquisition strategies.
Thomas Petit: Yeah that’s actually the main reason why I moved independent, because I wanted to broaden my experience. I mean some particular channel or audience wouldn’t work for some apps, and I wanted to broaden my skill a little bit.
It definitely is challenging in terms of mental load like from switching to being focused on one and then to another, but it’s actually very interesting. In my particular case I think what eases a little bit is I work only non-gaming and very often for 3-4 particular verticals.
So I do see interesting patterns between them.
Benchmark always have limits, but it gives you a pretty good sense when something is particularly good or particularly bad, to have this first party benchmark in my mind and to help developers focus on the right priorities. Which is very often the problem, it’s not capability but what do we prioritize, especially as I work with many very early stage apps where internal resources are very limited.
So it’s all about prioritizing.
App Store Optimization
John Koetsier: Excellent, excellent. Must be a great way to live and work.
So let’s get into it … we’re going to talk about a bunch of different things on this podcast. We’re going to talk about App Store Optimization, we’re going to talk about the increasing complexity of mobile data, we’ll talk a little bit about creative optimization, maybe we’ll hit the duopoly — which I noticed is a favorite subject of yours recently on Twitter — and some growing obstacles for indies and early stage app makers.
So I’m going to jump right in on App Store Optimization. That has been something that has been of tremendous importance for many years obviously, and organic discovery is still really really critical for a lot of companies, especially if you have limited funding. Talk to us about what’s happening there, what’s kind of new there, how pay for play is coming in, and maybe as well some of the benefits of getting featured.
Thomas Petit: So the thing here is, I think the main change, if you take a big picture perspective and not like small change and all, is that this field moved in perception a lot.
If you asked somebody three, four, five years ago the first term that would come in mind to them would be ‘keywords.’ And it’s true that back then it was relatively easy to actually make a difference by just tweaking keywords a little bit, because the competition was low.
But I think that was from the very beginning a mistake, and that mistake has been corrected in the sense that I see App Store Optimization as a very broad discipline that basically touches everything that happens in the store. And here the thing is we sometimes forget that what’s very specific about the app industry is that we’ve got this bottleneck between any communication we might do and users experimenting with our product, which is these users have to go through the stores.
And so it’s all about this particular bottleneck, and I actually think that today keywords are relatively not important in the sense that Google has a very complex semantic way of guessing what your app does, regardless of what you input.
But also because on the App Store, on the iOS App Store, it’s become a lot more competitive to make a difference by the keywords.
And I think that’s great because it means this field has moved to the other topics, and I think the big one here is conversion optimization … how you change the visuals of your app page to trigger better conversion among the viewers. And this is actually super critical because while the other changes only touch organic, this will apply to everything, because every single click that you pay on Facebook, Google, any other network, would actually have to pass through this page.
And there’s a pretty big drop of anywhere between 50 and 90% so this is actually really critical for making UA budget profitable.
But also to maximize your outcome without changing much the quantity of traffic. And I think a lot of attention has moved in ASO to conversion optimization and visuals in the last say two years, and I think it’s great and to me it’s a lot more interesting than keywords, but that’s my personal preference.
I think this is the big elephant.
And then you’ve got a lot of small things you need to factor: one is that there are ads in the store which wasn’t the case like three years ago, focusing about how to get featured is also in my opinion part of ASO, but then it goes also to smaller topics like monitoring the weight of your app because that impacts the conversion optimization, monitoring, obviously, ratings and reviews which have always been very important.
So it’s sort of the ASO discipline is a very broad field, and I’m glad that finally people see this way and not as a keyword discipline because this is just a small part of it.
John Koetsier: I really like what you said there about ASO being not just about organic discovery but also being something that optimizes your efforts for paid acquisition, because they’ve got to get through that gate right?
And you pay $2, $5, $3, or whatever it might be for somebody to go view that app on Google Play or the iOS app store. If they don’t like it, if it doesn’t look good, if it doesn’t match what you were advertising, if it just doesn’t seem right for them … they won’t convert and those dollars get flushed down the drain.
Super, super interesting.
Thomas Petit: In my personal case it’s almost like an accident that I put the focus there because when I was working at this fitness app called 8fit, I started there about five years ago, so I guess it was still quite in its infancy back then.
I arrived and I had a fairly good knowledge about how keywords work there, so I tried to tweak them differently: different duration, many geographies. And the truth is the uplift that I was getting were actually really really small because the low hanging fruits were already taken. And very soon my attention shifted because I noticed how a higher conversion rate was actually improving my position in the paid options.
Basically in any UA network you’re only as good as the conversion from CPM to CPI, or CPA now that there is event optimization. And people often forget that it’s not only about how many times people will click on your creative and how good is your CTR, but actually that if your conversion rate is low in the store this is like making your position in the auction unsustainable. You would have to bid a lot higher to get the same inventory.
And so we already did a lot there, maybe because we’re lucky to start a little bit before other players, this really gave us an edge for about two years of being able to drive cheaper quality traffic thanks to that.
I would say one of the other changes in ASO is that nobody’s done, so when there’s something that is making the difference people will notice very fast. And as the focus has changed there, obviously it’s a lot harder to make the difference, but I would say you still have to invest quite an important effort there. The sense that before you could win over the competition if you were better in this field, now it’s rather the opposite, it is if you don’t do it you’re going to lose. So it’s sort of you’re forced to do it, but regardless why you’re doing it you have to do it anyway.
Getting featured in the App Store and Google Play
John Koetsier: Super, super interesting.
Talk to me a little bit about getting featured. I mean with exposure to such a wide range of apps you must have seen that multiple times. I talked to somebody on our second podcast just last week who got featured in Australia and other nations — they’re Aussie so that’s why they focus there — and they saw hundreds of thousands of new installs because of that.
And that was on the iOS app store as well as Google Play.
What have you seen about getting featured? And we know that’s changed significantly with the app store as it’s changed in the last year, what that looks like and how that works. What have you seen recently for the impact of getting featured?
Thomas Petit: So my overall experience, and I’m talking here about across many apps so I think it’s partly representative of what’s happening in the store, at least for non-games because games are fairly different, is that the impact is a lot less than what it used to be.
I don’t know if people are not acting from the featuring tab or it’s just there’s more featuring, but this was quite concerning today.
I actually remember a conversation we had before the store was revamped by Apple and in the industry we all thought that featuring would get actually bigger in impact. And what we’ve seen is that it actually got smaller over time, so that’s definitely a reality.
That doesn’t mean that it’s not interesting to get, they’re still free installs!
There might be different people than you can target elsewhere, so I still think it’s an extremely high valuable thing to try to pursue, but because the impact is less it’s also had a healthy effect which is I see developers being a bit less obsessed about it. And while it’s great to be featured, it’s also not a sustainable strategy because you’re not going to have it all the time, and I don’t think this is something you should rely on but that should come as a bonus.
If it comes great, that’s awesome. If it doesn’t, well keep working.
There are some tricks that help gain feature, but there’s no perfect recipe. There’s one thing that in my own experience was disappointing. When I was fairly small, we were a very small team, it was really hard to get noticed and we barely got featured at all, but that’s when we were really small. So even a smaller feature would have been a huge win for us.
And then fast forward a few years later when we had a more organized team and a lot more professional about every part of the app, which obviously had better design, better UX and so on. We started getting featured and eventually got featured quite heavily at some period which really was nice and we had more interaction with Apple team, so that was extremely valuable.
But in terms of impact, it’s sort of too bad that the impact came when we needed it less because we would be bringing a million installs on our own and then we’d get 10k installs. We could almost not see it sometimes.
So I really wish that Apple would put a lot of focus on small and indie apps because that’s the ones that are really helped by it. It also brings more variety to the store. It’s true that when you’re just starting it’s hard to get noticed and very often your app may not be on the quality standards that Apple requires to feature apps … it’s just how the game is played and that’s normal.
Marketing data and increasing complexity
John Koetsier: Yes, yes, unfortunate the rich get richer.
It is super interesting something you brought up, the major change to the app store came in last year. I thought, and many others thought ‘wow they’re doing significant in-depth features, you’re going to get featured for a whole day, you’re taking up the whole screen or a big chunk of the screen’ and we thought that would have a huge impact.
And it turns out, I mean in my experience of using the iOS app store, it turns out that I skip that tab entirely and I don’t look at that tab. I jump right into ‘search’ or some of the other tab, but generally into ‘search’ because I’m searching for a specific app. And maybe that’s just me, but I avoid that tab mostly for games and for apps, and therefore I think because it’s out of the regular flow of people’s use of the app store that’s decreased its effectiveness.
Interesting, interesting things. It was not how Apple intended it. They intended to give more impact there but it seems like that’s how it’s worked out.
OK … let’s talk about marketing, mobile marketing, user acquisition and this increasing complexity of data.
And so we’re turning away from the ASO and the Google store optimization conversation and we’re turning to this data flow that’s coming in to user acquisition managers from all the partners that they’ve got the media sources, ad partners that they’ve got, and they’re bringing in all this data. Talk to me about what’s coming in and how it’s more complex than it used to be.
Thomas Petit: So I guess there are a few reasons why it is more complex.
One is if you look back a few years ago, it was always a breakthrough when you were saying that you were looking beyond the install, five-six years ago it was like everybody was focused on CPI.
Obviously the more you go down the funnel the more the data becomes complex because you’re looking at specific segments and you’re looking at more and more events that are only specific to your app and are not really something that is a benchmark. And especially when you try to start modelizing a lifetime value of your users. Maybe you’re using a combination of ads and IAP. Maybe you’re using a very complex scheme of subscription with different lengths, different value, different … also you need to factor the currencies and so on.
So it’s really like the more you go down the more there is a huge quantity of data, but also a more tricky way to interpret the data … which very often you will need to project early on to be usable for user acquisition.
In this sense I don’t really think that it is more complex than before, it’s just before we weren’t even looking at it. It’s sort of we were not smart enough to understand that this was where the economics are really happening. Maybe it was a bit easier, you would just throw in a lot of traffic and it was profitable, so we would care less.
It’s become a lot more competitive so because the profitability is lower you would focus on getting smarter and getting more and more … like the fruits in the tree are higher and higher so you need to develop tools to go and grab them. I think that was an evolution that happened in parallel and I don’t remember specifically when I started with subscription apps, there was basically no tool at all that would help us understand the exact value of our user. It was only all averaged and we had to rely on the information from Apple console which is really hard to use to iterate on your product, because you can’t really segment the way you want.
And I think this was a very radical change, like here mentioning subscription in particular but it doesn’t really matter, I think there are more tools in all aspects of data management because the whole flow has become more complex. Profitability is less and you need to get smarter and smarter. There are also more needs for this tool.
I remember that we wanted to buy tooling a few years ago and we couldn’t because those tools didn’t exist. I’m very glad that there’s lots of very professional tools that came out to the market, a lot more consolidated sub-markets within it with very serious competitors, also challengers so it’s really really interesting to monitor. I just think it’s because marketing got smarter, like it’s not because before we didn’t need that data, it’s basically we had no tools to use it and we had less need to look at very complex things.
The thing is there’s no way you’re gonna… I mean unless you’re really really lucky, or your concept is so innovative that you’re way ahead of everybody … which never happens … you will have to dig down very deep in the data. And it’s also part of the reflection that the marketer has become a lot more of a data analyst than what he used to be, more of a media buyer say before. And I think it’s just a reflection of how the field is maturing and getting smarter and going deeper into what really matters.
So that would be my interpretation, basically we’re less dumb than before.
Rise of the subscription model for apps
John Koetsier: It’s pretty interesting that a couple years ago, or even longer than that you were in the subscription space. I mean that’s obviously the hottest thing right now. It’s super super hot right now. A lot of apps are moving to subscription revenue. There’s a lot of revenue to be made there, and of course Apple’s jumping in there as well with Apple Arcade, so you were a way ahead of the curve.
Thomas Petit: I just got lucky. I mean I wasn’t aiming at a subscription model initially, but it’s true that I find it myself very interesting in the way you interpret the numbers. And also because it was so different from where the industry was at the time.
It was very early on driven by premium ads that would pay in the app store, and then a mix of IAP and ads would be the norm, especially in games, and the subscription not only because they were new and different, but also because the problems that come with it are actually extremely interesting in terms of renewal and all.
It’s true that I got a little bit specialized in that. I think it was accidental but yeah I got lucky because I did a bit after things got turned into the hot topic, so it’s obviously very useful for me today as I see many app developers asking me about specifics of the subscription based monetization.
But you never know … things turn pretty fast.
I think they’re here to stay for a number of reasons. They make sense for many businesses because Google and Apple want them, but you never know, maybe in two, three years people will be tired of having 25 subscriptions and we’ll see something new happening.
So you can never take too much rest and always monitor what’s becoming hot in the future. So that’s my challenge.
Why VO and tROAS don’t really work for subscription apps
John Koetsier: That’s great advice. I know that there will be subscription fatigue. I mean you’re certainly going to see that in the streaming media space especially for video right? Disney+ is coming out, Hulu has something, Netflix has been there for a while, Apple TV+ is out and there’s so many …
How many little holes are you going to drill in the bottom of the bucket of your bank account for money to drain out on a regular basis?
And we’re going to have that in apps as well, but I’ll ask you one question here because you’ve been in the space and you’re successful in the space. How does app marketing … how does mobile marketing change when you are focused on a subscription monetization model versus let’s say an in-app purchase model?
Thomas Petit: Well there’s definitely differences in terms of how you interpret your LTV, but very often this work is not necessarily going to be executed by a pure marketer and media buyer, so this data will be fed back to them.
So that would be one difference, but not necessarily the core difference. I think that there are two very big changes today, one being that a lot of the platforms, and specifically Facebook and Google, have moved from optimizing for install to optimizing for events and a value-based value optimization on Facebook and tROAS (target return on ad span) on Google.
And the thing is that the value optimization model doesn’t really work for subscription. They work really well when you have a huge variance in the revenue.
So typically gaming where a lot of the revenue is driven by only a few users has huge variance, e-commerce has also fairly high variance, and then this optimization model makes a lot of sense. But if you’re selling the same subscription to everyone and even though you would have a monthly and a yearly subscription the LTV of those different users is actually fairly similar, which renders the whole model of value optimization a bit useless.
So most of the subscription businesses I know operate on event optimization.
Basically: have you subscribed yes-no, which triggers another problem, which is the completion of this event and specifically in the case of long-term subscription, most of the health and fitness space would operate on a yearly subscription.
Also fairly common in the education space, but then your conversion rates mean that it’s really hard for the machine learning from Google and Facebook to optimize on if the conversion is like 2-3%. And obviously if your price points at entry is $1-$2 it’s a lot easier to get conversion rates of 10-20% that makes this machine very efficient.
When your conversion rate is like 2-3% it is really hard and so a lot of the industry in the subscription world is optimizing to free trials which is both great because then you manage to make those systems work on event optimization based on the free trial. But it’s also very problematic in the sense that starting a free trial does not fully correlate to LTV. And especially there are huge differences in constellation rates among different ages, and I’ve got a specific anecdote about that.
Maybe I’m talking a bit too long on the topic but I’ve got an anecdote …
How machine learning can fail marketers
John Koetsier: No it’s great, this is awesome stuff, keep going.
Thomas Petit: I remember when Google introduced UAC a few years back and I was like ‘oh sure I’m going to train this machine to bring me more free trials’ and the truth is Google did its job in the sense that I was like ‘okay I want to pay let’s say $10 for every trial that is starting’ because on average I was realizing that I was making my money back this way, and eventually Google delivered.
And I realized there was a problem in my data which is I had my customer trial exactly where I needed it.
But then when I looked at this cohort mature I realized that my return on ad spend was a lot lower than what I would expect. And the thing is because I only fed a very early event to Google, what happened after Google brought to me a lot of the users who know exactly how to complete a subscription, and specifically users under 20 years old who are a lot more savvy using the phone.
And what happened is the conversion between free trial and paying subscription was completely off the chart, but by the bottom. Like it was less than half of what we usually had.
And that’s where I started to think that I need to get a lot smarter about the event I’m optimizing for, and to adjust to this new reality of event optimization. We found a couple of tricks that help, but it’s a lot less intuitive than it might look at first glance.
And the lesson here is really look at your cohorts: don’t assume that they’re going to behave the way your previous cohorts do because as soon as you’ve got a little bit of variance we’re actually talking about big money differences.
Google will give you exactly what you ask for …
John Koetsier: That is super interesting. I mean the lesson there is be very careful what you’re optimizing for because Google will give it to you and if that does not result in actual revenue .. you are screwed.
That is a great insight and you are stuck between a rock and a hard place because you couldn’t optimize on the variable that you really wanted to, on the KPI that you really wanted to, which was the paid conversion, because the volume wasn’t enough there to train the algorithm, train the AI to do what you wanted it to do.
Really really challenging.
Thomas Petit: It was challenging.
We actually found a few tricks here and there to battle this. Maybe one thing we did not do right is if we’d put a lot of money into the machine then even a low conversion can be trained to do … like I mean it doesn’t really matter what’s your conversion rate if you have the budget too low to send a 100 or 500 events every day back to this algorithm, then it will work again.
So maybe we were too small so we had to find a few smart work-arounds around it. I have to admit that it took me many months to find those that work, with many failures in the meantime, but we found a couple of recipes that definitely helped with it.
John Koetsier: Nice, nice.
And that was of course one of the main criticisms of UAC when it first came out is that it costs an awful lot to train that algorithm, and I think they’ve gotten a lot better at that now but there were many marketers who spent hundreds of thousands of dollars training that algorithm, and once it’s going it’s going, but there can be a steep cliff to come over there.
The challenge of being an indie developer in a winner-takes-all market
Thomas Petit: Yeah I agree, and I don’t think it’s a bad thing by itself, but it leads us back to a topic that I think we were going to mention a little bit later, which is it’s making it really hard for the small guys, for the indie, for the early stage, in the sense that it splits even further …
I mean the App Store has always been a winner-takes-all model, has always been very unfair and like .001% of apps would collect more than half of revenue and it’s always been like this, but it seems like all these trends combined to make it even more in the sense that on the UA side you need to have a minimum volume and minimum budgets to make this machine work at their peak efficiency.
We mentioned that tooling became a lot more complex and all these tools, I mean there’s a lot of tools out there that are really brilliant and didn’t exist before, but if you have to pay like $1k a month here and $3k there and $5K here, and 2% of our spend on this one … I mean if you’re really big it makes a lot of sense because you’re making a better return on ad spend and this tool improves you.
But if you’re small you just can’t afford to pay 5 or 10 grand a month for tooling and you’re doing it the old school way, and obviously you are at a disadvantage. So I mean this is something I’m observing and I don’t have a recipe against it, I’m just saying it even reinforces the initial nature of a winner-takes-all of the store.
It’s something that’s becoming even more true today even if it was already the case before.
John Koetsier: Yeah, yeah I totally see that, I mean in terms of Singular obviously we focus on companies, app publishers who spend $4 million and up annually in advertising right?
And there are many customers who spend in the tens of millions annually, and some who spend over a hundred million dollars annually. And that is not where you enter the market if you’re an indie, if you’re an independent.
Thomas Petit: Yeah absolutely, and very often the added question I’m trying to educate people on what they should prepare, and when you’re going to grow you’re probably going to need tools like Singular and other types of tooling.
And I say okay I understand very well where this is going later when I will be spending a few million, but what do I do now? And I say you do it manually, or you build it but it’s going to cost even more in engineering time and so on. You’d rather focus on your product and just try to get there as soon as you can to save all this time.
I don’t see a lot of other recipes.
John Koetsier: Yes and it’s really really challenging, I mean you probably saw the thing… I think Eric Seufert tweeted it out and I retweeted it, it was probably last week and it was on Reddit.
There was this app developer, an indie app developer and he decided that he needed to not only be an app developer, he needed to understand user acquisition, he needed to understand mobile marketing, and he documented what he did and there were eight methodologies that he tried and different things that he did, and he documented the results of each of them.
But if memory serves, I think he invested a few hundred dollars or maybe under $500, certainly under a thousand in each of those eight different methodologies, and all of them kind of failed.
And it was it was a little bit sad to read. I was super happy first of all that that he was totally public and open about what he was doing, but it was a little bit sad to see the budget figures that he was able to put out there were just insufficient to do what he needed to do and to help him learn at the rate that he needed to learn in order to become one of those high growth apps.
Thomas Petit: Yeah, yeah absolutely, I read this article. It was really nice to see the transparency but also fresh tone of somebody discovering a field that I’m in, and it seemed the guy was actually pretty smart and was learning on each but it’s what you say, there’s a minimum amount that you need to actually trigger the optimization.
It’s really really hard to move with tiny budgets.
It’s also really hard to get help because most agencies, consultants, whatever help you can find are actually going to need minimum amounts to help you to make it worth it for their own resources. So it’s really hard to be in this position. I do try to dedicate some of my time to help people in this case and try to avoid some of the mistakes.
I think this guy took a right approach but sometimes going with just a little budget is really making it impossible, like entirely. It’s just a sad reality that is good to know before you start spending money but it’s not uncommon. I’ve seen it elsewhere, but it was nice to to see it in the open, publicly, so definitely a recommended read.
I’ve read it on Gamasutra, but I don’t know, maybe it was a repost.
Creative optimization: the only differentiation in user acquisition?
John Koetsier: Right, I think you’re probably right. I think it was Gammasutra and not Reddit, but thank you for that.
Let’s jump back then and talk a little bit about creative optimization. You tweeted out something, I think it was a week ago, something like that, about Miri growth, and they did a study: 3000 ads for 30 apps, and your conclusion from reading what they had done and what they looked … your conclusion was creative was the main differentiator.
That’s super interesting to me because I’ve done some research on that as well, and it seems like creative is a huge differentiator and sometimes more important than any other factor. In fact every time that I’ve seen, more important than any other factor.
Can you talk about that a little bit and what you saw in that study, what made you say that?
Thomas Petit: Sur e… this article was really interesting and also the transparency of an agency sharing their success and failure, and so thanks to their team for that, and to Adam who sent me the article.
The thing is it’s even a broader trend and not just an observation that I made based on this, where it’s like I had this slide that I showed public USA, we’re back to the Madmen. It’s sort of like the old school way of looking at marketing that would be all about the concept of the campaign and not so much about its execution.
And we went the complete other way a few years back where it was all about the numbers and well, whatever the creative we don’t really care, and it’s sort of we reached a middle point understanding that the work of a marketer is basically uniting the data and a brilliant creative concept. So it’s sort of like there’s a train behind it.
I think my tweet was a little bit extreme because I said creative is the only differential in UA and obviously it’s not the only one, I want it to be a bit provocative so that people would react on it. That worked really well because then I’ve got a couple of friends from agency that were like ‘no it’s not true, there’s a lot of knowledge I know’ but it is true. You need a solid foundation before you get there, but then you’ll reach a point where the creative is really where the big wins are going to be made because the other ones … either they vanished with the change in network, and specifically with event optimization and value based optimization, where targeting is a little bit less … either it’s less important because you need bigger samples to train the machine or you can’t access it at all.
Like in the case of UAC which in our case was was doomed to happen because of the change in privacy that users are now demanding.
But it’s been a challenge for marketers to completely switch their mindset from having hundreds and hundreds of micro-targeting into ‘okay I’m going to give the keys to the machine and let it work just because it’s better than myself.’
Which basically leaves you with two major levers on the marketing side: which data and which event am I feeding back to the machine, and which creative am I actually showing to users … which is the the place where the big differences are made. And here by the creative I want to repeat something we said before so it’s obviously a lot on the ad creative but don’t forget about the store page which is actually an extremely hard creative to tweak because the ad you’re probably going to be able to show relevant ads to people, maybe you’re going to split them by gender or by type of audience and then you can chatter to each audience slightly different USP (unique selling proposition) but all of these people are going to go through the same landing page in the store which makes it really hard.
You really need to nail the creative in the store so that all these different audiences would find relevance and beauty. So it’s the creative is like two fold in this case, sometimes even threefold if you have a landing page and meet in between which I’m seeing growing.
So definitely that’s where the work should be focused on in a marketer’s life today, and I think it’s really really hard because we as marketers to be so data driven and excellent at number crunching, dealing with huge number of metrics, complex funnels and so on, that we tend to hire very analytical people and then we try to have these analytical people be also creative. And in my experience these two things tend to be rare to find in one person, like the most creative people are a bit less analytical and the most analytical are less creative.
So you really need to build a team that has these two sides and particularly transforming data into insights that are understandable and actionable for the designer and creative mind is a talent that is rare to find in the markets, and very very sought for and looked for.
John Koetsier: And the pendulum swings, and it swings, and it swings, and it keeps swinging, and it keeps changing.
Thomas Petit: But that’s what makes it interesting too you know, we all are in this space also because it’s moving fast so we can’t really complain when there are big changes.
There’s ‘ah it’s not like before!’ yeah that’s the nature of this work so you accept it or do you change the field. Yesterday’s reality is very rarely next year’s reality in this field.
So you just have to keep moving.
John Koetsier: Yes indeed, we live in the future.
Excellent, let’s move to our last topic which is the duopoly. This has been a very interesting past year for both Facebook and Google, and frankly for big tech in general. I think Europe has been a bit ahead of the United States specifically in this, but I think with the new election coming up in the United States and certainly where the democratic candidates for presidency are concerned, big tech is definitely on the horizon as something that they’re looking at very very carefully in terms of antitrust, in terms of power to move public opinion, in terms of using their power and abusing their power in some ways.
So we look at the adtech space and we look at ad networks in general and we see Facebook and Google are huge. We see a bunch of other sort of little bit lower tier challengers, maybe the Apples, maybe Snapchat, maybe Twitter, maybe others like that, Pinterest and others like that.
And then we see thousands of other ad networks out there.
Talk to me a little bit about some of the current challenges and problems with the duopoly and how you see advertising and marketing evolving over the next couple of years.
How Google’s and Facebook’s uber-growth accelerated the app revolution
Thomas Petit: So there’s a very interesting perspective about this which is it really depends through which lens you’re looking at it. And it’s true that one side of the story is looking at how a privacy and user data are so obviously users are like ‘Oh no I don’t want my data to be sold’ but at the same time they also want to see relevant ads and they also keep using this product.
And I think here that’s sort of the wrong discussion in terms of marketing in the sense that if you look at the scandals, like the Cambridge Analytica scandal or all recent discussions about the flows in YouTube algorithms, that’s actually not a very relevant conversation to have with marketers. And I don’t think it’s really something that is super critical in the ad tech conversation that there is today.
And there are completely different points of view on this, so maybe US commissions will look at “oh but how are the news filtered between Republican and Democratic coverage in Google news” which is also not particularly relevant for advertisers.
And then you’ve got a whole completely different side of the story when you look at it from a marketer perspective, which I wish I would not say that myself, and you know my position on this, but the fact that Google and Facebook have grown so big is actually a marvel for marketers, because it means less headache, fewer partners, easier management and it’s actually been great.
I remember also reading Eric right on this that without Google and Facebook growing so big, and in particular growing so big in terms of market share in the online advertising market, it’s very likely that apps would not have skyrocketed the way they did. They were an enabler to the industry in the sense that they made everything a lot easier to scale which is less true in the webspace, but the concentration in the app space is so extreme.
It’s basically if you can’t make Google and Facebook work for you it’s going to be extremely hard to get to the very top, but at the same time those tools … they’re the same for everybody. So it also here gives a little bit of space for independents to actually have the same capability to advertise.
And so it’s super interesting to see how users, politics and marketers are looking at this topic through completely different lenses.
And I like to look at it from a political side, but also from a user and from a marketing perspective, and I have this contradiction inside me as a marketer: I want the most data that are possible, as a user I don’t necessarily want this data to be spread out everywhere. So it’s sort of like this insight problem that most marketers should have. And in a way there’s a lot of controversy but there’s a lot of contradiction in this debate.
At the end of the day just like we saw that GDPR ended up benefiting Google’s and Facebook’s dominance because they were the ones who could really easily deploy the resources to adjust and so on, but also because it made the walled garden even more defendable. You go out there today and Google and Facebook’s stance on privacy would be ‘Oh yeah we know everything about you but we’re going to keep it for ourselves, we don’t tell anyone, actually we don’t even know ourselves because it’s some super AI behind that we don’t even know how it’s working, like nobody is looking at your data.’
Which in a way is true, this is not analyzed by people to exploit your weaknesses and stuff, but at the same time it means that it reinforced the modes that are already preexisting and that a new challenger coming to the market today to compete against Google and Facebook will have it really hard, because it’s a lot more complex to gather the scale of data and the depths of data today than they did in their time when things were a little bit more open.
So it’s sort of an interesting internal contradiction here in the sense that yeah, maybe my data is actually better protected inside the walled garden than outside of it, but then is it also good for society as a whole and for advertisers in general to have two actors that gather let’s say 2/3 or 3/4 of the market.
I actually like a bit of of diversity myself, but it’s very hard to mix the privacy demands with some freedom in the market, and I don’t really know how this is going to evolve. I can’t see Facebook and Google really losing a lot of steam in the near future, but the midterm future is I think quite interesting to look at because there’s still a bit of uncertainty there.
John Koetsier: That’s a super interesting perspective, I mean at one and the same time the duopoly is an aggregation factor for supply. It’s an aggregation factor for demand. It’s an aggregation factor for advertising obviously and the data that’s required for that.
But it’s also an aggregation factor for regulation, and you can apply regulation on two players much easier than on a thousand, and they’re more regulatable in that sense, although getting it through legal challenges and everything can be difficult, but they’re more regulatable in that sense than 10,000 companies out there.
Thomas Petit: Absolutely.
I agree with that and I’ll make a parallel that I don’t know how relevant it is, but if you look at how Facebook specifically is dealing with attribution data, what they’ve always done is actually enable only a handful of companies to access the treasure and to pass it by back to developers. And they maintain a number of MMPs who have access to this marketing API, anywhere between five and ten at all times. If it was more than ten and if any developer would be able to tap into this API I think it would be a nightmare for them to monitor and weed out the bad player who is abusing the data and so on.
By focusing on just a few players it’s a lot easier for them to provide support, to provide improvement and so on, but at the same time they’ve been smart enough to not make this number too small. And I think if there was only two actors in the MMP space then there would be a real problem of abusive pricing and abusive dominance. Since Facebook has been really smart in always letting like five to ten, and if two would merge they would give access to another one just to maintain some balance.
And I wish that on the longterm it’s what’s going to happen in other fields, like I’m all in to have not a thousand but maybe three, four, five providers for videos and a couple of YouTube competitors. Same for social, you don’t want to have a million apps there but I think it’s healthy to have a few different players to keep innovation level high and to keep price down and to keep things a little bit more balanced.
But man … maybe I’m starting to be political myself now.
John Koetsier: Well Thomas, I just have to thank you for spending this hour, almost an hour now with me. This has been insightful. This has been interesting. This has been eye-opening and it’s been a real pleasure.
And you’ve done it all on a Friday night in Spain. It’s Friday morning for me, it’s Friday night for you, I know this is Europe you’re supposed to be off early and having a great weekend. Thank you so much for taking the time … I’ve got a full day of work ahead of me still if that makes you feel any better, but I want to thank you for being on the podcast.
Thomas Petit: Yeah no worries, that was my pleasure. As an independent I don’t have strict a time to work and it’s not unusual to have partners in the States, so I’m very used to it, and it was a complete pleasure to participate with you today and have this little chat.
I hope the listeners will enjoy it and I invite them to react on my provocation and show me facts to counter it, and the debate would probably be beneficial to all of us.
John Koetsier: Wonderful.