Level Up Your App Marketing: Measurement, Brand-Building, and Taking Risks — Sherina Khalidi, Deezer

David Barnard:

Welcome to the Sub Club Podcast, a show dedicated to the best practices for building and growing app businesses. We sit down with the entrepreneurs, investors, and builders behind the most successful apps in the world to learn from their successes and failures. Sub Club is brought to you by Revenue Cat. Thousands of the world's best apps trust Revenue Cat to power in-app purchases, manage customers, and grow revenue across iOS, Android, and the web. You can learn more at RevenueCat.com. Let's get into the show.

Hello, I'm your host, David Barnard, and my guest today is Sherina Khalidi, VP of Performance Marketing at Deezer. Sherina has been at Deezer for almost a decade working across multiple marketing functions. On the podcast, I talk with Sherina about why performance marketing always hits the ceiling, how to think about and measure brand marketing, and why ignoring A/B test results can sometimes lead to better outcomes. Hey, Sherina, thanks so much for joining me on the podcast today.

Sherina Khalidi:

Hello, David. Thank you so much for receiving me. I'm really, really glad to be here.

David Barnard:

Yeah, really looking forward to chatting. I feel like I've been saying this more and more recently on the podcast. And we actually have a few episodes that'll be released soon where I say a lot of people in the subscription app industry haven't heard of Deezer and don't realize what a big app it is, how long it's been around.

There's a lot of sleeper hits in the subscription app industry that aren't top of mind to folks. So I actually wanted to kick things off there. For our audience who might not be as familiar with Deezer, what is Deezer? Give me a brief overview of what it does and maybe even a little bit of history of the app.

Sherina Khalidi:

Yes. Deezer, it's a music streaming service. It was founded in 2007 by Daniel Marhely, a French developer and businessman in Paris in his mother's kitchen. So it's really a classic startup story. At first, it was really a website, of course, in 2007 and then we built an app for iOS and then for Android. So now we're really cross-device. We have a very strong website and also very performing apps.

Today it's available in hundred and more countries actually, but our strongest market is really France. It's our home market. Deezer has an interesting history because it started with the downfall of CD sales in music industry. It was launched at a time where everybody downloaded MP3 basically. It slowly became an opportunity for the music industry to monetize through monthly subscription, as you know.

So it's really the core of our monetization model, direct monetization model. It's really monthly subscription. But we are not owner of our content, so it means that a lot of this subscription money goes to royalties to the labels and the recording companies. We also monetize through ads, but it's really very marginal compared to the revenue we make from subscription.

And we also developed a lot indirect monetization through bundles with partners like telcos, hardware, retailers. So we are present in the US actually through a deal with Sonos. We also present in Latin America through a bundle with Mercado Libre, which is huge, absolutely huge there, and we are partnering a lot with telcos. That's really how we actually started.

David Barnard:

Yeah, it's fascinating. And Spotify had just started in 2006, so Deezer is really one of the early pioneers in moving music to subscriptions and moving music online. As you said, maybe a lot of folks listening to this podcast don't even remember back in 2007, 2006 and even before that of pirating MP3s and buying songs on iTunes and stuff like that. So yeah, Deezer has been around a long time and been a part of this wave. And while it's not as well-known as Spotify, it's just an incredible business.

Sherina Khalidi:

Yeah, indeed, indeed. It was at the heart of a big industry shift actually, and Deezer learned monthly subscription models really on the go. I remember when I joined Deezer, it was in 2015. And I remember that back then, there was major discussions within the product team and, of course, with the marketing team as well as to how far we should go into frustrating users, free users, to get them to pay.

Because really at first the model, it was a freemium. It was really based on the idea that ads could also support our business model. But the truth is that with ads alone, we'll never be profitable. So once we realized that, there was a big question of, okay, what do we do with all these free users who are enjoying our product but not paying, not profitable?

And so we went very far into shutting down all the features and trying to frustrate a lot users, it was a while ago, and settled down to, oh, but in the end, degrading the user experience, the free user experience actually doesn't provide any good outcome because all what you do is showing your free users that, well, your product is absolutely not satisfying on a free basis and they will just leave and use another service.

David Barnard:

Freemium is still somewhat underrated, especially now a lot of subscription apps, my own included, I started using a hard paywall and forcing people to subscribe to my weather app, because similar to Deezer, I have high expenses with the weather data. There's been a lot of talk the last few years of subscription apps not offering freemium or like you said, degrading the experience. But to become a really big business, alienating those free users does limit your pool and like you said, push people to other apps.

So it really depends on what kind of business you're trying to build and how you're trying to build it. But it's interesting to hear that y'all did push harder into subscriptions and then had to dial it back. I think that's a lesson for a lot of folks as they do think about more strongly monetizing is making sure that you do still have a valuable enough freemium experience to keep those users around.

Sherina Khalidi:

Exactly. That's a very subtle exercise where you have to show your value for money, but that doesn't mean that your free for the [inaudible 00:06:27] basically.

David Barnard:

I mean, even to this day, how do you and Deezer consider that freemium experience as far as are ads working? I mean, do you monetize those freemium users enough? Or is it kind of a these subscriptions are subsidizing those free users with the hope that over time more and more of those free users will upgrade?

Sherina Khalidi:

No, we actually reworked a lot of the free experience and the way we work with advertisers to actually be profitable on the free users. We keep really focusing on subscription to drive the majority of our revenue. And the way we see the freemium is more as a conversion tool, even though we did not crack it yet, to be honest.

David Barnard:

It's tough.

Sherina Khalidi:

Yeah, it is. It is. We're actually doing the other way around now is that we are opening more and more our product and our user experience to non-Deezer users. We want to encourage our users to share more and really open the doors to also users from other streaming service.

David Barnard:

And then how does the partnerships, you mentioned partnering with telcos and partnering with other brands and stuff, how does that end up playing into that freemium experience? Do those partnerships actually help subsidize the subscription to make it more affordable or even free via those telcos and others where the user themselves aren't having to pay for that more premium experience?

Or how do you work... Again, I think a lot of apps who aren't at Deezer's scale are maybe starting to think, how do we grow and how do we better monetize for users? How do we make it more affordable or more accessible when we know just broadly across the economy, not everybody is going to pay for everything? How do these partnerships play into that?

Sherina Khalidi:

Well, it's basically working on a bundle principle. We have actually several types of deals. What we're doing when we work with a new partner is that we do things ad hoc. We tailor it to their business objectives, and we look for the place where it match. So it's really about finding a win-win relationship. And it's true that for our partners, mainly bringing music to their clients, to their customers, it's bringing emotional connection. It brings a lot to their brand values.

It also helps them connect better with the younger target audience. So that's really what we're working on. So for instance, one of our big partners is Mercado Libre in Latin America, as I mentioned. So Mercado Libre is a huge retailer in Latin America who built a bundle with Disney+ for the video streaming and Deezer on the audio streaming part. So on the user perspective, yes, of course, the idea of a bundle is indeed to have a better price as a package. And for us, it's really about retention.

David Barnard:

Again, another thing, a lot of subscription apps I think over time will find things like bundles to be very effective. Because as you said, it's really fascinating looking broadly across the subscription app industry in something like Amazon Prime. Amazon under Prime now has video streaming and music streaming. They've added more and more and more to these bundles in order to provide that value.

And while it does lower maybe our pool in some cases, you're also expanding the pool of people who are then able to afford the product. So it lowers the ceiling, but raises the floor, I think. What I meant by that with lowering the ceiling and raising the floor is that while your average revenue per user may go down in a bundle, you're raising the floor and that you're making it more affordable to more people.

So even though you have a lower revenue per user, you're going to have more users actually paying. So it raises that floor of getting more people into the app and paying, even if you're not making quite as much on an individual basis as you would if they subscribe directly.

Sherina Khalidi:

That's really about the price elasticity. Totally. And for us, partnering means an extended reach, obviously. So the mathematics are positive for us.

David Barnard:

So the next part of that I wanted to dive into is that you do compete with some very large players in the market. I already brought up Amazon Prime Bundle where people can get music. Now, Amazon Music is not... Not to disparage Amazon, but it is not something they seem to focus a lot on. I don't think a lot of people use that as their primary music streaming. But for folks who maybe can't afford or don't want to pay for music, but it seems like that's a smaller part of their bundle.

But then you have Apple Music. You have YouTube music. You have Spotify Music. Deezer is competing globally against some very big players. And so what I think would be interesting to dive into next is how you think as a performance marketer about the balance of brand marketing when you're going up against such strong brands, but then also the performance marketing, which you're more focused on. How do you think about that, especially in light of competing with these massive brands, established brands in the space?

Sherina Khalidi:

That's a daily challenge. I don't think there are many app developers that are competing against so big competitors when you think that Spotify is actually the smallest of all. And the way we see things obviously is that just as a philosophy to start with, we are the challenger. We can do what others can't. We don't have everyone focused at us, and we're small and agile.

To some extent, it's very challenging indeed because we are maintaining the same product as a Spotify, an Apple, or an Amazon with maybe 20 times less developers than them. But we still keep up the pace. So of that we are super, super proud. And we really think that being small is actually, it can be totally an asset, an advantage. And when it comes with partners, actually about partners, that's a big advantage for us because we can really allocate a lot of resources to our partners because partners is the weight a lot in our revenue.

There would be a drop in the ocean of Spotify and actually usually partners are looking to compete also against the big GAFA, the giant of the web. So we know that it's an advantage. On a day-to-day basis, it is indeed very challenging and not always on the very positive side, especially when we know that Amazon, Apple, to name them, they don't really... Profitability is not an objective for them.

So when we have Amazon buying our branded keywords on Google, they set the bidding so high, it's really hard for us to keep up to pace. But we still do. And I do believe that having to be profitable and having strong constraints actually make us more efficient and more smarter basically. We are looking to get profitable in 2025. That's really something that we are working on as a company effort for quite some time now. Our ambition is really to become one of the first to be profitable in the music industry.

David Barnard:

I can imagine that's pretty frustrating when you have a massive company like Amazon buying against your keywords, which then makes user acquisition even more challenging. But on a day-to-day basis, how do you think about allocating resources to build the Deezer brand, but then also do performance marketing in a way? And as you said, as the competitor, as a more nimble company, you have to be more efficient and effective. So how do you think about that balancing the performance marketing side versus the brand and how do you build both?

Sherina Khalidi:

For me, performance marketing doesn't go without the brand. Yes, our competitors are so big that we can see it straight away, but I think that's at some point, something that all apps encounter. When you focus purely and only on performance marketing, you hit very quickly a ceiling, a ceiling beyond which you cannot scale again. So yes, you will catch some low-hanging fruits, you activate two to three acquisition channels, and you will see indeed user coming in, and it will be instantly profitable.

And that's the beauty of performance marketing. But you won't go very far with that. And I really believe that the days where this was working are really over for the app industry because it only worked in the app industry maybe for the past 10 to 15 years and period. On any other industry, you have to work on your brand awareness because that's the beginning of the funnel.

And really the conversion campaigns, all the performance marketing that you do is at the bottom of the funnel. So for me, you cannot really perform in marketing if you lose track of the funnel. So brand marketing goes with performance marketing. It's just that so you will work your brand on the upper funnel, and yes, it will be more challenging to measure. That's for sure. And it will be more challenging for you to convince your top management that you need to do it.

So this, of course, I get. I mean, it's a daily issue that we all have to fix in performance marketing. But if you don't do that, your scaling possibilities will be very quickly, very limited. And I'm not even talking about retention here because obviously also to retain your users, you do need to have a strong brand awareness and a strong brand preference, because you know that at any point your competitors will come to catch your users.

So for me, it's indeed very important to have a very good balance between the two. And that's totally something that... At Deezer, I'm one of the best advocates for brand marketing because I can... We are the first line, on the forefront to see when our action lose impact. We cannot go further than what we did because we focused too much on the very lower part of the funnel.

David Barnard:

How do you define the difference between performance marketing and brand marketing? What would be some examples of brand marketing that you do that you consider brand versus performance?

Sherina Khalidi:

So the main difference is really the objective and the KPI behind it. So an example of brand marketing activations that we have is currently we have this massive TV and out of home campaign live in France. All the campaign and all the messaging is really about enforcing the positioning of our brand. It's not talking about the price, about a trial period. It's not even encouraging you to download the app. It's more really about what it means to be a user of Deezer basically.

David Barnard:

The vibe. Deezer is cool.

Sherina Khalidi:

Yeah, no, so beyond Deezer being cool, which is obviously one of our brand value, as you can imagine. It's really about making very clear to our user and non-users that for Deezer, everything is about music and everything is about living the music. We are not positioned on all the audio. We are very centered around the music, and all our branding, marketing is really around this core idea. We also want to put back the user at the core of our brand positioning.

So it's also all about the user. It's all about you. At these activations, we don't expect to have immediate return on investment. Besides the fact that we are not able to measure, that's not even the point. It's really what we are. We are using brand trackers. We are using post-tests. So we are really more on the quantitative studies, qualitative studies to make sure that we do have a positive impact on our brand image, but we know that we won't get immediate revenue out of it.

What we do though is on performance marketing, we are very focused on amplifying everything that the brand marketing team is doing to make sure that we do the final little push, the little pinch for users to make sure that users who are ready to convert will actually download our app and convert into paying sub.

David Barnard:

Yeah, and I think that's what a lot of people maybe underestimate about brand marketing is that when people do have some level of affinity or understanding of the brand, then when you do hit them with that performance marketing, that performance marketing can be way more efficient and effective because you've warmed them up as a lead. And to backtrack a little bit, I do think it's interesting, Deezer's scale, you were talking about performance marketing not working anymore.

I think a lot of people listening to the podcast would be like, wait a minute, our performance marketing is working great. But what they'll find, and you alluded to, is that as they scale, they're going to find that performance marketing to become less and less and less effective. And then as that happens, that's where brand marketing does start to come into play. And you do have to be at a certain size and have the finances to be able to spend in a way that isn't going to affect the bottom line immediately and measurably, but will then make your performance marketing more effective.

And so this discussion is maybe for the smaller apps to be thinking in the future, this is how we should be thinking about it, and then for the bigger apps that are already hitting these boundaries is that they really do need to figure out both sides of the equation.

Sherina Khalidi:

Exactly. At first, it will work. At small scale, it will be super profitable and you will be super happy, but then it's very short term. So as soon as you have a bigger competitor, like we do have at Deezer, when you start fighting against an Amazon, for instance, and Amazon is a very good example, what do you have left if you don't have a brand? What do you have left to actually retain your user and to keep acquiring new ones?

All you have is emotional connections that you managed to build with your brand. And this is not something that... It's an actual expertise to know how to position your brand well, target your audience well, understand very well your users. It's really about art and science. It's about how you do business while being more understanding your users.

David Barnard:

Let's talk about the science part of that a little bit though. How do you measure? You mentioned a little bit qualitative measurement and doing user panels and things like that. Tell me a little bit more about that. You mentioned you do have KPIs around brand marketing.

So then why don't we talk through both the measurement of the performance marketing side of things and the measurement of the brand marketing side and how you maintain both at the same time to even on the brand side, making sure that you're not overspending or being wasteful in that spending? And then how do you measure the performance marketing piece and then account for the brand marketing in that?

Sherina Khalidi:

That's one of my most favorite topic because there's no black and white answer to this question. For me, everything became even more complicated will all the privacy rules. We cannot track everything we want anymore. On the app side, you cannot even get anything before the last touch point. So when you think about that, it leaves you with pretty much no direct tools to measure the impact of your branding campaigns.

For me now, measurement is really about multiplying the source of truth. You have to use different methodologies, different techniques to measure different KPIs and make sure that they all align and they all point to the same conclusions, which you want to be. You're going in the right direction basically. So the way we do it at Deezer is we do use many different ways to measure.

On the brand marketing side, we are very much measuring the trend of our paid consideration towards our users through quantitative studies. So it means that every quarter, every semester we ask a panel of our audience how they feel about does Deezer knows music? Does Deezer understand me? Do I feel close to Deezer? And we especially focus on the paid consideration part, which is basically, am I considering paying for Deezer?

What's actually reliable on that is that we follow the trend. Maybe it does not measure super precisely the actual level of this consideration, but what we want to see is an upward trend. That's also what we are doing in terms of performance marketing. We follow the same pattern. So we do have last-click attribution. We are working on a multi-touch attribution model, but we know that it will somehow be biased because of the different ways you track on web versus app.

And as you know, we are cross-device, so we do have lot more data on the web journey than on an app journey. We also use incrementality testing. So we do as many incrementality tests as possible to validate our assumptions. And as part of the incrementality tests, we also try to figure out what is the right level of investment on the digital upper funnel campaign, so campaigns that will drive installs or campaigns that will be much more about really rich in frequency, how do these impact our performance campaign?

So these are things that you can actually do very easily with a conversion lead study through Meta among other partners. And we have the MMM, which is more of a statistic model. So it's not fully functional at Deezer, not fully satisfying, but this is still something that we are working on to keep the pace of what's really being done in the measurement part at the moment.

So my main goal actually, my main focus is to make sure that we're always one step ahead in terms of measurement. For me, measurement is everything. If you cannot measure what you do, all your efforts are a waste. You have no idea what you're doing. So yeah, my next steps really on the measurement side is to use the results of incrementality tests to apply some scores into our MMM. Once you validated your assumptions, it's not super scientific, but you can work with these assumptions and end up with a golden boot where you know exactly precisely what worked, what doesn't work.

David Barnard:

I think some of our audience... And when you said MM, I was expecting you to then say MMP, mobile measurement partner, which I know a lot of folks... I think fewer folks will be familiar with the idea of media mix modeling. Can you explain that? And then is that all done in-house or is there a tool that you use for media mix modeling?

Sherina Khalidi:

Yeah. So media mix modeling, it goes beyond tracking and attribution. So MMPs for us, they are vital. We could not work without MMP. But as I mentioned with privacy, it's become very limited. So you can understand your last touch points, but you cannot understand anything else. Medium mix modeling, the objective is really to help you understand the impact of any parameters to your business. So it can also be competitor's activations, competitor's offers, but it's mainly used indeed as a...

To make it obvious, it's really about your media mix. So it helps you measure the impact of offline media such as TV, out-of-home, print, and it can also help you understand the impact of your digital brand campaigns. You know those that you do on YouTube that nobody can really measure, but we all know that it works. So to go beyond attribution, media mix modeling is really about building a statistic model, comparing your spend versus your daily or weekly incoming traffic to outline actually the variations, and also to define a baseline.

But it's not perfect because ideally for me, in my dream world, an MMM would be able to tell me what will happen if I stop investing on anything. We all know that yes, we do have a baseline, but if we stop investing on it, the baseline will decrease. And yet weirdly on MMM, your baseline is always flat. It's not perfect. You cannot base all your decisions on an MMM, unfortunately. That's why for me, having multiple source of truth is key. You validate your assumptions again and again to most inform decisions that you can.

David Barnard:

And is your media mix model all done internally? Do you have a whole team dedicated to that? Or is it a few people in the... I imagine you have a marketing engineering or a growth engineering team. Is that mostly done internally, or are there any specific tools that you use to help with that?

Sherina Khalidi:

Yes, the answer is we do everything internally. We have an amazing data science team that spends time on building those MMMs and improving them. We work definitely super well with the data team to improve it. To me, I would not say that doing it in-house is the best setup for two reasons. The first one is that, okay, maybe those data scientists that you have internally, maybe they would be better working on your product algorithm, for instance, on some maybe more product related missions.

And the second aspect is that what I find maybe lacking a little bit is that when you work with an external company, they see so many different company problematics and challenges and they're 100% open to what's happening outside. So they will be able to feed you with the best solutions for your own challenge. So that's still an open question. We do it in-house, but it's something that I work on for 2025. Maybe about that, the best setup would be to have a hybrid of internal and external.

Because indeed, internally, you will have all the essence or the knowledge of what's happening in the company and in your industry. If you outsource a little bit of that, you take an external consultant to audit what you're doing and make some recommendation coming from outside, that might be a good win.

David Barnard:

You mentioned that part of the challenge of media mix modeling is that you don't necessarily get precision in understanding what changes will directly impact the overall profitability of spend, but then you had mentioned before that you layer on incrementality testing. So how do you use incrementality testing in combination with media mix modeling to better be able to understand and predict which parts of your marketing spend are working and which parts are maybe less efficient?

Sherina Khalidi:

So the idea behind it, and that's really something that's done by companies like L'Oreal, it's my top of mind company, but when I heard that they did that, I was like, okay, I absolutely need to do that too. So the concept behind it is that through incrementality testing, you are able to calculate a score between what you actually see as direct tracked data and what was more hidden. It's usually... Well, it can be on both ways.

Maybe you do attribute more directly to certain channels than what they actually contribute, like for instance, search, to say that's a very basic example, and maybe you do under attribute to other channels that come earlier in the funnel like social ads. So you can with incremental testing estimate a score and use that score to modify your statistic model, your model mix modeling.

David Barnard:

So in your incrementality testing, are you completely shutting off specific channels or how do you actually do the incrementality testing?

Sherina Khalidi:

I think really incrementality testing as level where you need to be well... Either you have an excellent agency to get you on track. You do lots and lots of trainings. But basically on incrementality testing, there are different methods depending on what you want to check and with which partner. So one of them is GeoLift, and GeoLift is really about shutting down some regions and comparing them to not shut down regions, but the regions have to be similar.

And there are tools on Google that help you do that actually on Google Ads. And when you shut down, you can calculate the increment. But there are many other ways of doing incrementality testing like, as I mentioned, conversion lead study, brand lead study, where you have a controlled group and an exposed group and you can measure the uplift between the two.

David Barnard:

We're getting super in the weeds, which is a ton of fun because we don't always get to do this on the podcast, but I think this can be really helpful for folks to be thinking about this. And again, there aren't that many companies at Deezer's scale doing this, but the companies that are on their way toward Deezer's scale should be thinking about these things. And so I think it's super fascinating.

And so just a couple more questions on that front. Would you also potentially switch off specific types? So maybe you don't actually cut all YouTube Ads, but you withhold a certain type of brand marketing and keep the other brand marketing. What other ways do you use incrementality testing to understand the performance?

Sherina Khalidi:

So on brand marketing, you have to keep one thing in mind, brand marketing is not a short-term impact. So switching it on, switching it off, if you expect to see a difference between the two, I don't believe in it, because compared to... For instance, you work with an ad network and you're not really sure that the ad network actually provides on the conversion side.

Then switching it off and on, you will see the impact. You will see it straight away because it's a short-term channel. But whenever it's long term, you have to be a bit smarter than just switching off and see what happens, because the actual impact can happen maybe next year or in two years. You don't know that. It can be very slow decrease, but still also very consistent and hard to reverse.

David Barnard:

My last question there then is have there been any especially interesting results from the incrementality testing? Have you done a test where you removed a channel and wow, things just drop like a rock? Or and the opposite, have there been channels where you drop them off and you just don't see a difference at all? Any surprising results to the positive or negative in some of your incrementality testing.

Sherina Khalidi:

Actually it validates more assumptions that we already had and the surprise was to which extent it was. I would say that all the campaigns that we do that are more upper funnel campaign that will optimize based on installs and not necessarily paying subs. We knew that it had an impact that was not so easy to measure on our performance campaign, but we did not imagine how much. It really confirmed the absolute necessity to keep investing on upper funnel campaigns, even though they won't have a short-term impact.

What's really interesting with incrementality testing is that it actually validates the fact that all the more conversion campaigns that you have, they only amplify what exists already. They don't create anything as such. So that's, again, the brand marketing versus performance marketing debate. But yeah, performance marketing does only amplify what has already started way earlier in the funnel.

David Barnard:

Again, just so fascinating and fun to have somebody like you on the podcast who works day in, day out having to make this work. And like you said, Deezer pushing toward profitability in 2025, these are the steps that you take to make that happen. And so it's really great to get this level of insight from you on how that all works at the scale of Deezer. I did want to shift gears and talk a little bit about Deezer's approach to experimentation and testing because I think you have a lot of really interesting ideas there.

Sherina Khalidi:

One thing that I'm absolutely fundamentally convinced of is that nothing can be done in performance marketing if you don't innovate, if you don't test everything that you can test. And that's also why I really love this job. You have to be curious. You have to be daring, and you have to not be afraid of failing. You only invest only in the most secure campaigns with the most secure audience and the most secure channels, the whole relevance of what you do with fall, either in terms of volume, in terms of cost, or in terms of quality.

You will once again hit very quickly a ceiling if you don't get out of your comfort zone and go test some stuff. So that's why I also insist a lot towards my team. In France, culturally, we are not raised to test and fail. We have this belief that we need to succeed straight ahead, at the first trial. And if we are not sure that we'll succeed, well, better not try. So it's a lot of education to do, I would say, of company values to infuse of same speed to repeat to say, no, actually I will never reproach anyone in my team to have failed at something.

What would be a problem for me is people who don't try. So what I really expect from my team is that they do test a lot. They have to be very honest of the result, of course, and not be afraid to say, "Okay, I tested that and it failed." And that's fine. I mean, the main question is what are the learnings? Learnings are so valuable that yes, it requires sometimes to sacrifice a little bit of budget, take a little bit of risk.

Because in the end, what happens when we do the math at the end of the year and we look at everything we tried, all the biggest fails that we had versus all the biggest wins, what we see is that wins always compensate for fails. I think that's also why we've been so successful in the past years in performance marketing. I don't like to brag, but we did improve our LTV/CAC. We more than doubled it in two years time while scaling up on volumes at the same time.

So pretty incredible. And for me, there are, of course, many reasons and it's also a company effort. It was also because we collaborate very well with other teams and we are all looking the same direction. But the main factor that helped us really scale while improving our profitability is all the tests that we run. I'm really, really very proud of my team, the test stuff that I would have never dared to test when I was younger.

David Barnard:

That's such a fascinating topic from a leadership perspective that as a team you're spending a lot of money in a failed test. Well, not completely failed. Like if you spent $50,000 in a new ad campaign and it performed half as well as another ad campaign, you're not necessarily losing hundreds of thousands of dollars because you're still getting some level of performance, but you're losing money in that you could have potentially put that budget behind another ad that was performing or another channel that was performing.

So how do you instill that in a practical level inside a team that you may inefficiently spend a considerable amount of money, which I mean, just as a human, it's hard to reconcile that and go to work the next day and be like, "Oh man, I just wasted money." And then how do you put guardrails against that as well to where you aren't going to lose hundreds and hundreds of thousands of dollars in these experiments?

Sherina Khalidi:

First of all, we do frame our test very well. The risk-taking is actually quite limited if you look at the big picture, if you look at the overall. And that's really what I encourage the team to do is to look at the overall TV stack. As long as overall we are profitable, it's okay to test stuff and indeed lose money. So if you look at it in an isolated way, yes, it's money. We would've loved to spend it better. But rather than looking behind, it's more about looking ahead and get the valuable learnings that would make the next test actually overperform and compensate for the loss.

David Barnard:

And then how do you incorporate gut feeling versus raw data into making some of these decisions where maybe sometimes something appears to be underperforming, but you just feel like it actually is the right thing to do.

Sherina Khalidi:

It's being data-driven. As you can understand now, it's really key for me. So data is at the heart of everything. Measurement is at the heart of everything. But at some point, there are stuff as a marketer that did that you know. And when the data is not showing it, you should definitely to me still listen to your gut feeling and do it anyway and see what happens. And I do have an excellent example on that.

We had this homepage last year, our Unlocked homepage, that in terms of design, in terms of clarity, nobody was really satisfied with. We developed two alternative versions in which I think all of us really believed that one of these two will overperform. So we would say more of a branding Unlocked homepage with key messages around our brand and our product.

And there was more of a player looking like Unlocked homepage, a player, but Unlocked and obviously with features that are very limited and encourage you to subscribe whenever you wanted to stream a song basically. So we ABC tested these pages against each other and none of the alternative versions won against the control version. And so yes, that's the moment I think that we all agreed to listen to our gut feeling, which was okay, in our case, the control version was not satisfying to us.

It was not conveying the messages that we wanted to convey, and it was not user-friendly at all. So based on that, we made the decision to go live with the branded version. We implemented it. We rolled it out. Now that we did, a few weeks after, we see positive results compared to the old version. So that was one of the example where we are actually right to listen to our gut feeling.

David Barnard:

I think the key there is user centricity is that you're doing what is best for the user. And I think a lot of marketers, and we see this broadly throughout the industry, a really annoying example is the gaming ads where the gaming ad is completely fake. The game itself doesn't actually do anything that was shown in the ad, but those ads perform, but then people get into the game and they're confused.

And that's just a really blatant example of not being user-centric where you can do things in marketing that "work" or that the data shows works, but you're not, to the point of this whole conversation, you're not building a brand. You're not doing something that makes the users feel good about playing that game, even if it did actually convert better than an ad that did show the actual game.

So how do you think about that user-centricity and how that permeates marketing and product and throughout the entire user journey? And that was a perfect example of it is a homepage that as a marketing team, you felt put the user at the core. Even if it didn't immediately A/B test better, over the long haul, putting the user at the center did make a difference.

And it showed over time, which, again, that's the hard part about being data-focused as well, is that sometimes you're fooled by your A/B test. Because in whatever specific timing or external factors or other things that you can't even account for in an A/B test led you one direction, but it was actually the wrong direction in the long term. So how do you think about that user-centricity and how to apply that and those gut perspective to making those decisions?

Sherina Khalidi:

User-centricity, I think, is really a game changer when you shift your mindset as a company towards listening to your users and stop think that you are part of the users and that your personal opinion does matter a lot because you use your product. And B, that everything become from product features, the product led growth, which does not take into account really the needs of the user.

For me, the shift really means that you equip yourself to listen to your user, ask them the right question, and listen to them all the time, and actually delegate some of the decision-making. Good example is when we changed our logo. So before 2023, we had an equalizer as a logo. It was our logo ever since our creation, so from 2007 until 2023, we had the equalizer as a logo.

We knew that we needed to change for different kind of reasons, but we knew that we had this big problem of brand attribution. Somehow our logo didn't seem to be recognized very well, which is surprising when you know that it was there since forever. And so we decided to change it, but we did not decided to change it just because we were convinced that it was needed.

We decided to change it because we asked our users. We asked them to draw our current logo to see if they were able to draw it. We also tested our existing logo against two alternate versions, and we were like, this is not a personal opinion. Now our logo is a purple heart, so a big, big change. Really, really, really internally it was a big change to go from a very tech equalizer to a purple heart, which is much more into the emotion, connecting emotions with our users.

So nobody really believed that the purple heart would be the winning version because it was so disruptive and maybe looking more like a dating app, stuff like that. But actually what came out of the studies was that it was a winning version, and it was a winning version by far, and it was a winning version on all target audience and through all KPIs. And for me, that's the example of user-centricity.

We listened to the users and we made decision based on it. And what we see really all our signals is that it was the right decision to make ever since. I also believe strongly that it really helps my performance campaign to be, again, more profitable while scaling up. So yeah, user-centricity should be at the heart of any marketing actually.

David Barnard:

Well, I think that's a fantastic place to wrap up this conversation is that... And again, I think in a world of A/B test this and that and use every dark pattern you possibly can to encourage people to subscribe in the first couple of minutes, a lot of that works. And sometimes you use some of those tactics to get to the profitability level you need to get to. But in the long run, what wins is providing a really good experience and a good feeling to the users.

So I think it's a fantastic place to wrap up. And then even especially throughout this entire conversation about brand and performance and the combination of the two, the fact that listening to users and being user-centric, even so far as changing your logo, then actually increased performance on performance marketing is just fascinating. So thank you so much for all these insights.

Anything you wanted to share as we're wrapping up? We're going to include links to your LinkedIn. And I know Deezer is hiring, so we're going to link to the job page as well. But anything else you wanted to share as we're wrapping up?

Sherina Khalidi:

Yeah, we covered everything. Thank you very much for inviting me to your podcast. I mean, these are questions that I'm always happy to have a follow-up on, so I would be super happy to connect with anyone interested in measurement especially so we can share experiences.

David Barnard:

Don't sleep on the job links. Paris is an incredible city. And Deezer is not remote, right? So it's pretty much everybody, all the job listings I saw were Paris-specific. So if you work in mobile marketing and want a cool job in a cool city, go check out the Deezer jobs page. Just talking to you for the last hour, it sounds like it's a fantastic place to work and really doing some cutting-edge things in marketing and being the underdog and getting the fight against the bigger players in the space. So I think it's a really cool company and a really cool opportunity for folks.

Sherina Khalidi:

Yes, definitely. So if you want to move in Paris, Deezer should definitely be on the list of companies you apply to. We do have several jobs opening at the moment, and obviously there will be many opportunities in the future. At the moment, we're looking for a senior SEO manager, by the way. So if you're interested, you can apply on our Deezer jobs board and also reach out directly to me. Voila! Voila! La vie Paris is so wonderful.

David Barnard:

All right, thank you so much, Sherina, for joining me. This is a fascinating conversation.

Sherina Khalidi:

Thank you, David.

David Barnard:

Thanks so much for listening. If you have a minute, please leave a review in your favorite podcast player. You can also stop by chat.subclub.com to join our private community.