Prediction and automation are becoming two key components in the user acquisition journey as mobile marketers scramble for consumers in a competitive and budget-conscious market.
Most big companies are resorting to prediction tools to have an idea of approximate returns on their advertising spends across different channels and allocating resources accordingly while automation is used to optimize those campaigns and save time and the possibility of any human error.
So, we sat down with Enric Pedró, the VP of Growth at Tilting Point, for a detailed interview on how they implement automation, and prediction in their overall strategy.
Enric started his professional career as a sales coordinator at WildTangent Europe before moving on to Linking Mobile where he fulfilled the roles of Account Director (UK) and VP of business development (North America). Following this, Enric had stints at Fibonad and Zitro, before moving to Tilting Point, where he is currently the VP of Growth.
How does automation play a part in your overall UA strategy at Tilting Point?
We manage about 40 games at Tilting Point and we’re looking to double that in 2022. It’s almost humanly impossible to review so many games at the same time to ensure performance is expected, that we meet our targets, etc. So since late last year, we’ve been trying to automate as many things as possible – insights, campaign optimization, the results we achieved on an A/B test, and so on. Because one of our strategies is to assist as many developers as possible, which is only possible in one of two ways: hiring many many more people, or using automation.
Should people be afraid that automation will force them out of a job? Or does it free up UA managers’ time to do other human-led work?
That’s a good question. It’s going to take years of work for any tool to be capable of meeting all of Tilting Points UA requirements. To take a random example...are you worried about AI truck drivers taking over? Maybe in three decades. But even then you’ll need a human operator keeping an eye on truck performance.
Going back to the UA front, many tools can assist humans because of our limited attention span. Let’s say it’s work hours, eight hours a day. If you’re managing more than one or two games, it’s not possible to have enough time to manually do all your supposed optimizations and daily duties.
So it’s always useful to have an automated system that alerts you when your attention is required or suggests what you should pay more attention to. That said, you can’t take away the human component, the expertise of our growth team. Sure, Google Trends may tell me that Nikki Minaj’s songs are trending, but you still need a human being who can discern what’s relevant to your audience and provide creative direction. Again, you can use AI and automation, and third-party tools to help you optimize faster, freeing up your time to pay attention to what really matters and where to make a difference. That’s where we actually see the best results happening.
Great point. We basically use tools to take away the boring parts of a job, but you still have to interpret what that means and apply the learnings.
So I want to get your thoughts on predictive modeling and the role it plays at Tilting Point?
Tilting Point has been building our model on PLTV because once we manage to grow a game it tells us the predictive model, how much budget we have, as well as our predicted ROAS. So ultimately, being a good publisher is being able to go to a third-party developer and saying hey, your game is lacking X, Y, and Z, tell them how we’re going to mitigate this, then based on our predictive model tell them what they can expect to be hitting in terms of targets. This is all thanks to our data and analytics teams.But today we’re rolling out new predictive models. Because as you know, the market is constantly changing with iOS 14.5 and now with Google.
I’m curious to know how Tilting Point leveraged predictive analytics leading up to iOS 14.5 – because that was a very different world. What has been the evolution and where do you think it’s going to go in the future?
So the approach before IDFA deprecation – and I think across the industry – was to leverage as many data points as were available, knowing they had an expiration date.
How did Tilting Point leverage that? We did our homework. We knew when it was coming. And we knew there was a huge opportunity because the vast majority of marketers were scared about change. We knew for a fact that most people will avoid using iOS traffic because it was an unknown and use Android traffic instead. And therefore we knew that iOS traffic is going to be cheaper or maintain cost, whereas Android will become more expensive because everyone will be moving there.
So we took that as an opportunity and instead of following the market trend, did the opposite by investing more on iOS. And I’m not saying that all Tilting Point games did amazing, that’s not the case. But the vast majority have not seen an impact or have been doing better than what they used to do.
There are some games, though, that given the SKAdNetwork limitations didn’t perform as well as they used to. One of those is social casino games. They have to change the way they use tracking because now you have a limited period where they can track the user, after which you go blind. So it changes the way you approach user acquisition.
Looking to the future, with Google, changes are effectively already happening with Android 12, but they’re not going to push for it for at least a couple of years. It seems they acknowledge the struggle caused by IDFA’s deprecation and are working towards transparency - they just recently released a blog providing a developer preview of their Privacy Sandbox.
Tilting Point has excellent relationships with both Apple and Google, which we bring to our developer partners. We work actively with both parties to the benefit of our and our developer’s games portfolio.
Do you think Predictive Lifetime Value (PLTV) is going to be one of the main KPIs in the future? And how else do you think PA will change shortly?
PA needs to be able to adapt to the privacy constraint industry that we’re moving into. We have fewer signals, fewer KPIs, and fewer events to optimize the users we acquire. Topics API is something Google is working on which in theory suggests that it’s going to be easier for developers to piggyback on specific events that are being tracked by Google, which is already a lot better than what we had with Apple. But that’s in theory. Once we get Google’s privacy-compliant environment life then we’ll see.
Because being able to automate and optimize game performance and PLTV, it all comes down to assuming that these KPIs or leveraging events are available no matter the point in time. If you change that, then people start scrambling to figure out what’s happening similar to what happened with the SKAdNetwork.
There are too many things we don’t know until we see Google ID deprecation live. So if anything, I would use that as an opportunity for those developers that felt like they missed out on the IDFA application that we were discussing before because the same thing is going to apply to Google. Are people all going to move to iOS? I’m not sure, because iOS is more expensive. But I do know for a fact that most people will follow a ‘wait and see’ approach, which is not the Tilting Point way because we want to be leading in this space, not following.
Let’s say you have a new UA manager or someone new to the world of UA join your team. What are some tips you’d have for them to succeed in their role?
A manager needs to have an analytical mindset. But it is an interesting mixture because you need to balance the analytical with the creative side. So you need that rigorous, data-centric approach, but on the other hand, you need to have brainstorming sessions with the creative team, which is using the complete opposite side of your brain. So what I’m trying to say is that being a successful UA manager, it’s not easy at all.
How can you become a good one if you’re not analytical? Well if you’re part of a mid-large company then you will have people and teams that can help you. So at Tilting Point, if someone is unsure about how best to represent data, then we’ll be open about that so you’re absolutely clear. What if you’re learning on your own? There’s a lot of great content online. It’s just a matter of how much you’re willing to absorb as a sponge because ultimately that’s going to shape what you do next.
The most important thing is actually applying what you learn to setting campaigns, seeing why this creative performs better than others, or why these MMPs are not tracking specific events. Because if you don’t apply what you’re learning, you’re going to forget about it.
And you need to like to keep up because it’s a fast-moving industry. If you went into a cave and came out a year later you’d be like holy shit what happened! The point is simply that you need to keep up with what’s happening because a few months down the line it won’t be relevant anymore.
These were some important tips and insights from Enric on how they use prediction tools to their advantage at Tilting Point in their user acquisition journey and how they play a key role in the data privacy era that Apple and Google have slowly migrated towards.
Want more tips and insights on how you can improve user acquisition? Download our 2022 User Acquisition Guide for Gaming Apps here.