💎 Growth Gems #104 - Data, UA, and Creatives
Hey,
This week, I’m sharing gems on Data, UA, and Creatives.
These insights come from Warren Woodward, Mike Haze, Ava Savitsky, and Magda Zareba.
Enjoy!
🥇 TOP GEM OF THE WEEK
Cracking Games Growth
Gems from Warren Woodward (CGO at Upptic), Mike Haze (CTO at Upptic), Ava Savitsky (Creative Director at Upptic), Magda Zareba (UA Manager at Upptic) in Cracking Games Growth: Revealing the Upptic Playbook
I recently got some good reminders that I should not focus too much on insights about subscription apps specifically:
You’re not all working in B2C sub apps
Even for those who are (like me), we need to keep looking outside of our niche and the subscription model to find inspiration and elevate our game
I aim to stay curious (I even have it on a t-shirt, in case I forget)
So, this week, I’m sharing insights from the world of games growth.
Whether you're in gaming or not, many of those gems should resonate with you. If they don’t, you might be thinking too narrowly.
Upptic has launched about 30-40 games and has built tools to help developers grow their games, so I trust their games growth playbook.
The high level
💎 There are 3 different phases when launching a game, each with its scope of work:
Development - getting the game ready to launch
Soft launch - getting the initial batch of test data and building the full go-to-market plan
Scaled growth - transform the game from an idea to a scalable, profitable business
(02:49) by Warren
💎 You can break down the process of growth into 3 pillars and subteams:
UA - arms of the beast (operations and day-to-day)
Data - mind of the beast (tooling and analytics experts)
Creatives - heart of the beast (ads, app store assets, information, organic social)
(03:54) by Warren
This is the high level, and they shared many gems for each stage. Let’s dive in!
1. Development & pre-launch phase
Creatives
💎 In the development and pre-launch phase, you need to get the piping in place and start developing the go-to-market strategy. There are also some things you may want to do before you build the game, like live-play concepts to get data on what’s going to work or not in the market.
(05:40) by Warren
Warren later gave the example of large developers wanting to mitigate risks and understand what kind of game model to use, for example, when they have a specific IP (e.g., Sponge Bob).
He says you can test this by building mockups of app store assets for 3 different game models and test-market paid channels to see what works best and which markets work best.
💎 This development stage is a great time to build all the creatives you will need for a soft launch and any “coming soon” assets (hype trailer, pre-reg landing page, community building assets or announcements, etc.). It’s OK if you don’t have a lot of assets yet, you need to focus on the most exciting ones and the key moments in the gameplay.
(06:26) by Ava
Here are some examples of hype videos used before the games were ready:
Ava also said that if you’re really light on assets and you’ve launched projects in the past, you can always name-drop by using copy like “from the makers of”.
💎 Before you go to market, test any core gameplay loops and other features that are still in development or that you’re considering developing: A/B test their viability via creatives before you go to market to see what resonates with your audience. You can save a lot of time and development cycles.
(07:15) by Ava
💎 Store page assets (icon, screenshots) are one of the first things you need to think about on the creative side because you need to be able to submit your build.
(07:35) by Warren
Data
💎 In this pre-launch phase, you must get your infrastructure and data stack right. This is an area that tends to get minimized or de-prioritized.
(08:02) by Mike
When it comes to finding the right measurement partner, here are some questions Mike recommends asking yourself:
How are you going to distribute the app? App stores, direct downloads, etc.?
Where will you be promoting the title? Which communities, what infrastructure?
💎 As progress is being made on the game, the dev team integrates the SDKs and tracking into the app. This is a cross-functional effort because you have to understand how you’re going to track success in the early phases all the way into post-launch, whether it’s for monetization, retention, but also the engagement events that are key to measure for you to optimize the game.
(08:40) by Mike
💎 Pre-launch, you also need to start growing a community and build the hype about the game by having players interact. Discord/Twitter are often tools used for this, and you need to track and understand how you’re converting those players.
(09:35) by Mike
Here are three resources you might like around building communities: Building Hits with the Power of Community, Fandom Gone Viral (Building Community Partnered Games) on Naavik’s podcast and Building Games with Your Audience (BattleFly).
💎 You don’t start measuring stuff after the game is live. Performance marketing starts from the beginning.
(10:15) by Warren
UA
💎 At the pre-launch stage, the UA team takes care of the legal paperwork and SDK integration in collaboration with the data team. They also sign the IOs with the media sources they plan to promote the app on and then integrate those media sources with the MMP.
(11:10) by Magda
💎 The pre-launch state is the phase to work closely with the creative team to align on the creative strategy for the different media sources.
(11:49) by Magda
2. Soft launch phase
💎 The goal of the soft launch is validation by real users. You have a functional version of the product, and you need to understand if the game has a “pulse”: does it have a healthy economy, are the core loops effective, is the game stable? At this stage, the Growth org is a data provider for Product to validate all those things.
(12:25) by Warren
💎 It is key that you set ahead of time, as a cross-functional team, what you think the success criteria and goals are for each phase. This helps you know what you’re validating and have objective standards to reach before each next phase.
(44:48) by Mike
Warren explained that if you know from experience the steps to take to improve a specific area (e.g., XYZ action should help improve ad monetization by 30%), you should factor in this when deciding if you’re ready to move to the next phase.
UA
💎 Phases for the soft launch from the UA team’s perspective.
Tech launch - make sure the app runs smoothly by identifying bugs and verifying all the integrations. This phase is usually quite short and is relatively cheap (often with Facebook app install campaigns in Brazil and the Philippines).
Soft launch - evaluate the app’s potential to generate revenue but also user stickiness. At this phase, they usually run Facebook and another SDK network like Unity or AppLovin for countries where users behave similarly to the target country (e.g., with the US as a target market, use another tier 1 English-speaking country like Australia/Canada/New Zealand). They run basic mobile app campaigns for the first 2-3 weeks, then move to more sophisticated optimization types.
(13:30) by Magda
Why not the US directly? Because you might expose your target market to a product that is not yet ready, and might not get a second chance to make a good impression.
At this stage, they don’t have clear KPIs yet and don’t know the payback window, so they use benchmarks. The main difference between soft launch and scaling is that soft launch is focused on data collection and analysis.
In the Q&A, Warren explained that you want to look at the progression of your ARPU curve, and where people drop off in the funnel. Regarding ROAS, some games don’t have great retention but are great at monetizing: the LTV curve might not look amazing, but if the activation happens early, you can still have a successful game and adapt.
Creatives
💎 During the soft launch phase, the creative team produces the first UA creatives and starts getting some data back. They usually go with creatives that have a wide appeal and focus on the game's key moments. You start building your brand and work toward consistency, but it is also an important time to stay flexible and open-minded based on the data coming in.
(17:08) by Ava
💎 Soft launch is not the time to get widely experimental with creatives yet, but you do want to test various unique concepts: at least 3 videos and a few statics. Since the creatives are used in multiple countries, you don’t want text-heavy concepts that rely on English fluency.
(17:50) by Ava
Data
💎 The data team needs to test everything to make sure that there is no leak in the pipes, that all the data is flowing, that the data from all the different data sources (MMP, networks, etc.) is normalized, and you can understand the aggregate data. The data needs to be meaningful for the UA and creative teams to start understanding what works/doesn’t and set goals and expectations.
(18:55) by Mike
You want a setup/tool that allows you to compare media sources as much as possible “apples to apples”, and a buying model letting you see channel by channel and country by country breakdowns.
💎 At this phase, you want something that allows you to demonstrate that you are successful, the game is viable, and investing more is a good decision.
(21:10) by Mike
💎 Some larger developers need to present the marketing plan for a year or more before you have the soft launch data. If that’s the case, define profitability targets and create 3 scenarios and “budget pools” amongst which to choose once you have the data:
Above profitability targets? Use the larger budget pool.
Average? Use the mid-case budget pool.
Below profitability targets? Use the barebones budget while you continue to improve the product
(21:35) by Warren
3. Scaled growth phase
💎 The scale phase is about turning the game into a profitable business. Can you put $1 in the marketing machine and measurably get more than $1 out? It’s a simple question but has a lot of moving pieces.
(23:00) Warren
Creatives
💎 By the time of the actual launch, you want your creative portfolio to have about 10 different concepts: status, gameplay videos, and UGC videos. You use the best-performing assets from the soft launch phase and add some widely new and different concepts. You start broad, cast a wide net, and then zero in on what’s working as you collect the data.
(23:50) by Ava
By then, you should also have done some ASO testing and be confident in your app store assets.
💎 Divide your creative production cycles with a 60/40 split:
60% on iterations (new variables, get granular on characters to show, best end cards, CTAs, etc.)
40% on new concepts to find creative breakout winners (give designers a lot of freedom)
(24:20) by Ava
I’ve typically seen people mention 80% of iterations and 20% of new concepts. I’m just speculating, but the change here might be due to the restrictions brought by privacy policies and the fact that it’s harder to understand the impact of smaller creative iterations.
💎 When you scale, you want to be OK with failing and take higher-risk creative ideas that aren’t copies of your competitors’ ideas because this allows you to find new creative winners while having a foundation of ad creatives to fall back on (the ones from soft launch).
(25:04) by Warren
Here is an example Upptic shared for a “out-of-the-box” creative:
UA
💎 At the scale phase, you know the app and have KPIs based on the soft-launch results. During the first few months, you start calculating the actual payback window.
(26:10) by Magda
💎 As the budget gets bigger, you get more opportunities to test and find the best sources for the app for your target country (you can then explore other countries). Start with media sources used by competitors (usually FB, TikTok, Google, and SDK networks like Unity/AppLovin) so you can understand which ones are a good fit for your app.
(26:30) by Magda
💎 Once the app is mature and you know your payback window and your KPIs, you can start experimenting with media sources like incentivized traffic or DSPs.
(27:28) by Magda
💎 The scaling phase is also when you move to more sophisticated campaign optimization (value or event optimization) as soon as possible.
(27:45) by Magda
Data
💎 At this point, the data scales. You’ve “normalized” the signals during the soft launch, and you now look at tools for budget management and forecasts. You also add new events to track user progression/engagement so the UA team can optimize their spend and creatives.
(29:00) by Mike
💎 Due to privacy changes, you need a whole new model that builds on top of the standardized data based on SKAN (and soon Google) to understand SKAN optimization, cohorts, and ROAS. It’s critical to be able to compare the data from a specific source/platform to the model.
(30:15) by Mike
💎 Make sure you have a flexible infrastructure that allows the UA and creative teams to constantly turn on new channels, test new things, and an easy way to compare past results to current results (apples to apples) even over a long period.
(31:17) by Warren
Before I leave, here is a quote on getting clarity around the marketing P&L:
“A company can go bankrupt buying traffic profitably against a long enough ROAS window” - Eric Seufert (Independent Analyst & Investor)
See you next time. Stay curious!
⛏️ Sylvain
🔗 Sources: