ASO Conference 2023 Part 1 Recap: ASO Fundamentals
On June 1, 2023, growth agency Phiture hosted the 6th ASO Conference in Berlin, an annual event dedicated to advanced ASO and gathering many of the top ASO practitioners. As such, the AppTweak team was not only in attendance but also a proud sponsor of the event. We were delighted to meet with other members of the ASO community and discuss the recent evolutions in ASO. Impressed by the quality of presentations and conversations all day long, we thought to share our summary of the various interventions that took place on the main stage.
Staples for impactful App Store Optimization
With 14 different presentations on the main stage, the conference was a great opportunity to hear about both recurring topics at the core of ASO, and new stories that developed since the last ASO conference. In this first blog of our two-part series, we have shared a few highlights of the 7 presentations that touched on ASO fundamentals – A/B testing, alignment with paid UA, Apple Search Ads, localization, and prioritization.
Identify & plan an ASO strategy
The day began with a presentation from Anton Tatarynovich, Senior ASO Manager at Freelytics and former ASO strategist at Phiture, who shared advice on how to organize and plan an ASO roadmap based on his experience.
Having transitioned from a consultant role among a large team with many resources to a role as a single operator dedicated to growing one specific app, Anton pointed to how a prioritization framework is essential to avoid lacking focus or investing in low-impact projects. Quickly after starting at Freelytics, he applied the following 5-step process to identify and report where the early focus should be put:
- Looking inward: After identifying key markets that deserve early attention based on the user base size and revenue share, Anton advised to start with a health check of store pages per market to see which ones have been left unattended the longest and/or have the worse-looking creatives, metadata, and reviews & ratings. His advice was also to consider the app’s category ranking, not so much for ASO performance purposes, but rather as a soft signal of where your app sits in the market and items that can help grab top executive’s attention.
- Localization overview: Anton’s second step was to review in more details what localization efforts had already been made. This step is a good complement to the first one, and again focuses on research rather than immediate action. That’s because identifying areas with strong localization potential can shift priorities from bigger markets to smaller ones, where your app is further from reaching its maturity.
- Marketing benchmark: Only after doing a thorough inner review comes the time to look at competition. Anton warned the audience of how being aware of competitors is essential, but looking into them too early can bias your own efforts and lead you to prioritize changes that are less relevant for your particular app. Discussing further the challenges of reviewing the competitive landscape, Anton recommended reducing your list of competitors from 10+ to only 3 to 5 based on comparative shares of downloads. Smaller apps should not follow market leaders too closely, but focus first on competitors in similar notoriety and ASO effort ranges.
- Visibility analysis: The fourth step Anton shared is to analyze your traffic by source type in each key market to understand which items matter most. Especially when it comes to keyword optimization, he underlined that it’s important to understand the share Search traffic holds for your app and your category, as well as breaking down how much of search traffic comes from brand keywords compared to generic. This is to avoid lengthy keyword optimization cycles when consumers are not used to search for your type of product.
- Conversion analysis: Finally, reviewing differences in conversion rates per traffic source to identify where the biggest gaps are is key to determine where to start with conversion optimization efforts. Keyword-level data available in Google Play Developer Console as well as Apple Search Ads campaigns statistics can also help further break down your app’s conversion rate per search term or keyword type.
At the end of this process, Anton’s methodology is to finalize the audit with a color coding system, which helps visualize which areas to tackle first.
Apple Search Ads
The second presentation of the day was delivered by independent growth expert Thomas Petit, who offered to share his “hard truths” about Apple Search Ads (ASA). While Thomas shared many specific insights about Apple Search Ads, one of his observations particularly stood out to us – due to the differences in user behavior measurement between Apple Search Ads and other ad networks, he underlined that Apple Search Ads should be set aside from other paid UA channels as comparisons cannot fully take those into account.
More practically, Thomas also shared many useful observations regarding Apple Search Ads targeting challenges, budget management, and specificities of ASA:
- Avoid using Search Ads basic and monitor any broad match campaigns by checking on search terms.
Thomas illustrated examples of campaigns meant to target the keyword “Shazam” that led to bids on search terms relating to a TV guide app instead.
He underlined how even exact match campaigns can lead to surprising bids due to the “close variants” rule, sharing an instance of a campaign for the keyword “diet app” that ended up placing some bids on “diy app.” In all these cases, he highlighted how such issues cannot be avoided but should be corrected quickly and easily by adding negative keywords in your campaign settings.
- Search result campaigns are, by nature, likely to spend the vast majority of your budget on only a small number of keywords.
Thomas pointed out that this is because the App Store search essentially focuses on the “short tail,” while broad match and discovery campaigns can serve more as insights providers, with discovery campaigns potentially informing developers of how the App Store algorithm perceives their apps.
- Limitations of Apple’s campaigns audience settings.
Last but not least, Thomas warned that with only 22% of users opting in for personalized advertising on iOS, most gender-based or age-specific campaigns actually exclude more than three quarters of traffic.
On the other hand, the notion of new and returning users don’t offer enough clarity, and campaigns targeted at new users can still end up delivering redownloads. Even country targeting can sometimes create surprises (especially when analyzing campaign performances through the lens of MMP data), as Apple considers the App Store country selected by a user rather than the location of their IP address.
When it comes to budget management, Thomas highlighted several times the reality-check that is sometimes needed with Apple Search Ads budgets. With the intense competition, auction system, and opportunity to access a very large pool of consumers, it’s only logical that costs per install can be higher than with other ad networks. With that being said, he advised to not blindly trust Apple’s bid recommendations (especially with a few examples of fanciful figures), but also debunked the idea of Apple Search Ads working as a simple second-bidder price setting system.
With examples of campaigns providing very similar numbers of impressions and installs after drastically increasing or decreasing bids, Thomas indicated that the relevancy of the bidding app is also a key factor. He recommended apps with high tap-through-rates and share of voice for particular keywords test reducing their bids.
Specificities of Apple Search Ads
Finally, Thomas identified some of the specificities of Apple Search Ads in the overall paid UA mix.
- His biggest advice was to always separate brand from non-brand keywords in campaigns, both when piloting budget and analyzing performances.
- He encouraged ASO practitioners to analyze the lift and cannibalization effect of campaigns by pausing campaigns every once in a while. He also asked not to shy away from the cannibalization question, which is a cost that can be worth paying for the other insights you can get from your Apple Search Ads campaigns.
- Somewhat along the lines of this topic, Thomas also insisted that while search is a limited channel because of how hard it is to drive more people to search for a specific term, the size of Apple Search Ads and success from certain advertisers prove that it is possible to scale with the channel, provided you accept that the biggest limit to scaling is your own return goals.
Alignment with paid UA
Throughout the day, two more speakers added notes to how ASO and paid UA can and should work together:
Measure & predict the organic lift from paid UA campaigns
Talia Hermesh from Playtika first underlined how marketers can identify if paid UA impacts Browse traffic (store featurings as well as traffic coming from similar apps), generic search, or brand search, before highlighting how statistical regression can help forecast the organic lift from different types of paid UA campaigns.
She recommended running these analyses in countries where apps have high traffic, a strong K factor between organic and paid downloads, and good user quality, reminding audience members that “we cannot all spend our money in the US.”
Walking the audience through her methodology, Talia shared case studies from two Playtika games:
- A rewarded video campaign in Brazil delivered a +0.1 organic download lift per paid download to Bingo Blitz
- A native ads campaign in France helped House of Fun get 0.5 organic download per paid download thanks to a surge in brand search
She suggested marketers to get used to adding organic uplift to the KPIs of their paid campaigns. Whenever apps scale down on paid UA, they do so progressively to limit the impact on organic traffic.
Best practices for aligning ASO and paid UA
Later on, Luca Stefanutti from Adidas took to the stage to share a few best practices and observations for how ASO and paid UA campaigns can better work together:
- His first piece of advice was to address how ASO is not always meant to only focus on the long-term aspect of user acquisition, but can also make timely changes to support the short time nature of paid UA campaigns. To that end, the first tactic he presented regarded the amplification of paid UA campaigns, which can be done through in-app events and promotional content, adding a promotional text, running Today Tab campaigns in Apple Search Ads, and reaching out to Apple and Google for featurings in their respective stores.
- Regarding events, Luca noted that in his experience, promotional content on Google Play is about four times more effective than in-app events on the App Store. A second tactic Luca shared was to improve conversion through the use of store ads, events, and custom pages. He was also quick to warn that the latter do not always have a high impact, and need to be made a priority for qualified users and touchpoints.
- Finally, Luca shared a few formulas that ASO practitioners should use to calculate the impact of ASO on paid UA campaigns in order to more easily show value to top stakeholders who do not realize how ASO interacts with paid UA.
The afternoon at the conference began with a presentation by Mateusz Wrzeszcz (Klarna) who covered in riveting details some of the crucial yet rarely known considerations for efficient store A/B testing.
As an introduction, Mat underlined how A/B testing using tools such as Google Play store listing experiments or App Store product page optimization differs from A/B testing creatives with Facebook ads. This is because contrary to Facebook ads, store A/B testing focuses on predicting future differences rather than “merely” comparing real time data.
- Consequently, Mat’s first piece of advice was to give research the same importance as testing itself, as a test can only be as good as the hypothesis it is trying to prove. This should translate in experiments’ prioritization, which should focus first on hypotheses based on internal data and insights, second on best practices, benchmarks and heuristics, and third on what-ifs and creative adjustments.
These prioritization principles should also help favor bolder tests with better Minimum Detectable Effect. Without going too deep into statistics, Mateusz explained how experiments reach better statistical power when they run for longer, working like nets that become tighter as the sample size grows and can, therefore, catch smaller changes in conversion. As many apps don’t necessarily have that much time available to run a single test, it’s important to test first the experiments that have the best chances to deliver a conversion change that can be caught early by an A/B test.
- So Mat’s second key piece of advice is for every ASO practitioner to identify what is the Minimum Detectable Effect they should go after per app, country and storefront. They should also use sample size calculators to make sure they avoid testing hypotheses that would take too long to be properly validated to make the test worth the investment. To that end, Mat made his own sample size calculator and recommended ASO practitioners group their markets into three tiers for traffic and for CVR levels to easily visualize the markets where all, some, or no experiments are worth it.
- Finally, Mat recommended that all ASO practitioners take the habit of testing radical changes at least once per quarter, and aim to test one hypothesis at a time rather than only one visual change at a time.
- He also reminded the audience to prioritize changing assets that appear in the first impression frame, and don’t systematically test the same change in all markets as markets are not all equal to one another. Once again, everyone should remember that test results are only directional and matter less than the quality of the hypothesis that led to the creative changes.
Last but not least, when it comes to fundamentals of ASO, Davronbek Makhkamov (Ulysses) and Anastasiia Baranchikova (Prequel) each came to the stage to share insights and recommendations for how to address the topic of localization in the app industry.
A clear roadmap for successful localization
Davronbek emphasized on how localization is an extremely impactful area for growing your apps, as it requires comparatively fewer tech resources than adding new features or porting your product on more platforms. For this reason, he elaborated a very simple visual framework to help him formulate his own roadmap to taking any app international:
|Level 0 = globalization
|Level 1 = translation
|Level 2 = localization
|Level 3 = culturalization
Simply by comparing the level of localization of an app in the various countries it is available in, you can quickly identify areas of improvement. You can also factor in whether you prefer to further improve your localization quality in a given country, or prefer starting from scratch in a new one – the ultimate goal being to take the app to the highest level of localization in each of the countries where it is available.
To facilitate the decision between improving localization and expanding to another language, Davronbek advised to consider the resources needed for specific action items corresponding to a deeper-level localization. Also important to evaluate is how sensitive a certain population may be to localized products vs. global products. Specifically, he suggested the following action items when tackling proper localization and culturalization:
- Localization action items: Researching local keyword search patterns and implementing them in your app store metadata, translating your website in the local language, testing localized pricing, and translating your support materials and FAQs.
- Culturalization action items: Adjusting your app title to include a local name extension or even transcribing your brand (for instance, in languages using different alphabets), testing and implementing culturally appealing creatives, featuring local content in your app and website, hiring a local to handle all communication channels and community platforms, setting up a local marketing plan, and offer full customer support in the local language.
Importance of recognizing cultural differences
Right after Davronbek’s presentation, Anastasiia took to the stage and expanded on the theories that can be used to power better culturalization and more efficient marketing.
She first identified a challenge in how the Internet has led to a belief about an emerging “global culture,” when, in reality, cultural differences remain very real. It needs to be understood that different cultures do not experience values the same way, instead of thinking different cultures necessarily promote different values.
- As a result, Anastasiia recommended that ASO practitioners learn about multiple culture differences frameworks to better understand how to tailor their messages to key markets based on dimensions that change how particular values apply in different cultures. As a concrete example, she illustrated how messages like “the whole world is watching” and “one app for all your sounds” are likely to resonate better respectively in cultures that are more collectivistic or individualistic.
To properly leverage cultural differences in ASO, she recommended the given process:
- Select a value that can be connected to the use of your app, particularly via the needs and pains consumers experience before using it. For instance, the need to belong being a strong motivator for using a photo editing app like Prequel.
- Research cultural differences models that can help measure the different attitudes towards the value you have selected. In the case of Prequel, the need to belong may be expressed differently according to gender roles and/or power distance.
- Identify which elements in your store creatives can be used to display the user benefit (depending on context intensity, you may turn to text in your screenshots, real life imagery, subtle highlights, etc). Draft a backlog of hypotheses for what kind of creative experiments to conduct and where to conduct them, while being sensitive to the fact that some markets may require you to go for a different core value that your product addresses.
- Finally, Anastasiia proposed to push the application of cultural differences frameworks in ASO. Sharing her own experience with localization in South America, she highlighted how it is important to create creatives that reflect how local consumers see themselves rather than how you may see them.
For Prequel, that meant understanding that consumers in Brazil don’t see themselves through the lens of carnival, despite it being an important seasonal event. Furthermore, she suggested that apps can also attempt to leverage clichés to take a more personal tone. They can mix elements of their own culture and cliché about it with illustrations of how the product addresses a particular value.
Stay tuned for more!
We’ve divided this lengthy blog post into two separate articles in order to effectively communicate the insights and key takeaways from the ASO Conference. The second part will explore the emerging topics in ASO, where we’ll touch upon interesting talks on artificial intelligence (AI), ASO for crypto platforms, third-party app stores in iOS, and much more. So don’t forget to check out the second part of the series!