Case Study: Achieving 66% growth through App Store Optimization
This is an in-depth case study by Jiri Chochlik, an ASO Expert at AppAgent
Good mobile marketing managers are relentlessly looking for new ways to improve their user acquisition efforts. They spend time finding the best strategies to follow, hoping that they will give them an advantage over the competition. If youâre a big publisher or an indie developer, ASO is an essential strategy that you can use to grow your app business.
In this case study, we will give you an in-depth insight into the steps we took to achieve a 66% growth in organic downloads for Kiwi.com. We hope it will offer some useful app store optimization strategies that can use for your own app.
In this case study you will learn:
1. How we defined opportunities for growth through ASO
2. What our ASO strategy for Kiwi.com looked like
3. Our keyword optimization process for US and non-US countries
4. Which assets have been tested and how we achieved an incredible 82% increase in the conversion rate from over 30 experiments.
Kiwi.com is one of the five biggest online sellers of airline tickets in Europe, and the fastest growing startup in the CEE region, according to Deloitte Technology Fast 50. The companyâs turnover increases massively every year, reaching an impressive âŹ700 million in 2017. Kiwi.com currently employs over 1500 people and operates in nine countries across the world.
AppAgent has acted as the full-service mobile marketing team for Kiwi.com for nearly two years. Weâve built up an incredible relationship with the business and the brand during this time. Itâs a powerful partnership thatâs delivering results.
IDENTIFIYNG GROWTH OPPORTUNITIES
When we first assessed Kiwiâs app store presence, we found some common issues with both keywords and store conversions:
1. Keywords werenât properly structured, and they were frequently placed in the inappropriate field (title, subtitle, keywords)
2. Many keywords were not generating decent traffic
3. In major non-US countries, keywords had been just simply translated without considering - and properly researching - local search queries
1. There had been very few A/B tests and store experiments in the past
2. Screenshot design was rather âdullâ, did not stand out in the search results, and was due for an immediate update
At a strategic level, while Kiwi.com offers flight tickets to travelers around the globe, its focus was mainly on the US store. This meant that they were missing key opportunities in foreign markets including France, Japan, Spain and the U.K.
ASO consists of two distinct, but interconnected, areas: conversion rate optimization of the store listing and keyword optimization.
In order for us to have a solid understanding of the landscape, the first step we took was performing an in-depth analysis of Kiwi.comâs competitors. We needed to understand how they present their services; their visual and language style; whether they use video or not; and their store listing experiments. All of these would need to be considered when we created the strategy, and calculated potential rankings for Kiwi.com.
In order to identify the countries where additional investment into ASO could be profitable, our data analyst prepared a list of top markets using available sales and market data. To narrow down the list of countries we looked primarily at three metrics: the average revenue per user; top-10 countries in terms of revenue; and market size (downloads potential etc.)
Being âon brandâ, while keeping sufficient space for experimentation, is a never-ending challenge for any big brand. There are always keywords that could work well volume-wise in the title/subtitle, but which fall foul of the brand guidelines. To ensure we were always on point, Kiwiâs Head of Mobile Karel Soucek was always engaged in our discussions, giving his view on our messaging.
The whole optimization process was split into two phases:
1. ASO for US market both on GP and iOS (including keyword optimization + store listing)
2. Localization for the most attractive non-US countries
1) KEYWORD OPTIMIZATION IN THE US MARKET
A) Identifying competitors and Kiwi positioning
In order to understand what the possibility of achieving a high ranking (within the top 10) we first needed to assess Kiwiâs place in the travel category, as well as identifying the best apps in this niche.
The ranking potential is strongly dependent on aspects like the reach, brand strength, visibility, number of reviews etc. As the travel category is a very competitive one, we knew that Kiwi.com wouldnât be able to outrank the best apps like Skyscanner or Kayak for the most searched for keywords (those with Top 10 rankings).
Our strategy was to look for less competitive keywords with less traffic but much higher probability for Kiwi to rank in top 10 - and ideally in top three - as thatâs where most installs happen.
The travel category, and especially its niche - offering cheap flight tickets - is tricky as users are more likely to search directly for well-known brands rather than for generic phrases like âflight ticketsâ.
However, our assumption was that there is also a considerable number of users who are not familiar with big brands. I personally confirmed this hypothesis by simply searching and recording how many brands within this category were unknown to me.
This diagram explains the whole keyword optimization process:
Start with an in-depth research to create a long list of possible keywords that are relevant, have decent traffic, and to assess if the app already ranks for them.
Evaluate our chances of reaching a top-10 or higher position based on the level of competition.
Select the final keywords set(s).
Submit the new keyword set.
Measure, iterate, measureâŠ
Note: We always check all of the keywords that we select manually by using multiple techniques and tools; we know that we cannot rely on any single tool for this crucial part of the process.
2) KEYWORDS OPTIMIZATION IN NON-US MARKETS
The level of localization ranged from simple keyword translation to a full localization of the store listing, including screenshots.
For keyword research in the most important countries, we utilized our network of native-speakers including experts in Germany, France, UK, Spain, China, and Japan. For the rest of the world, we did just simple translations, as the ratio of time investment and growth opportunity didnât justify a more complex approach.
Instead of including more countries on iOS, we expanded localizations to Google Play. Another hard learning point was about the accuracy of translations. After finding a few mistakes using paid online translation services, we decided to hire at least two translators per country to double-check the final keyword sets.
Checking typos or spotting nonsensical words is impossible without a local knowledge of the language. We wanted to absolutely nullify the chance for any grammar errors on Kiwi.comâs store page and therefore invested in building a network of verified translators who were also trained in keyword research.
In the end, when we examined results, an interesting difference between iOS and Google Play appeared. On iOS, we saw a great uplift across all metrics. However, ASO on Google Play had been much tougher to implement and we didnât see such big growth - even though Kiwi ranked #1 for one of the main keywords âflight ticketsâ for some time. (While not as successful, we still consider a 31% increase in app downloads within the first year as a great result.)
We also saw that keyword rankings were moving up and down all the time. You can see how dramatic these rises and falls are below in the Google Play Store through this âvisibility scoreâ diagram generated by Apptweak:
Don't split your forces - Find out the platform on which your app performs better in terms of revenue and downloads, and focus on it when you start with ASO.
Donât go head against the wall - Keep in mind that the competition in US App Stores is the fiercest and that shifting your focus to other territories can bring quick wins. Itâs important to learn what the best performing countries are and assess whether localization would be effective there.
No shortcuts - If you want to see real benefits in keyword optimization, you need to invest in in-depth keyword research, ideally using locals who can brainstorm the most frequently used search queries and typical phrases.
Itâs important to note that downloads also increased as a result of the conversion optimization work we complete, as you will see in the next section.
1. Store listing analysis and how to stand out
Have you ever asked yourself whether itâs better to start with keywords or conversion rate optimization? So did weâŠ
Our thought process was as follows: If Kiwi.com doesnât rank well, then the rankings need to be improved first so we can learn about the true performance of the default store listing.
That performance then served as a baseline for our experiments with store assets. Itâs crucial to run just one test at a time, because doing so allows you to evaluate it properly. Thatâs the reason we stopped iterating on different keyword sets once listing experiments were running, and vice-versa.
We completed a large number of tests on both iOS and GP. This diagram demonstrates how the process works. Below we discuss the most interesting findings.
In the beginning, Kiwi.comâs screenshots looked generic, which resulted in a low conversion rate. At that time, with the release of iOS11 fast approaching, we knew that having landscape screenshots in search results offered us an opportunity to capture the attention of visitors. The goal was to take advantage of the new features of iOS11 and create something a bit different that would stand out from the crowd.
Another experiment we came up with was the idea of using some of the great testimonials the app had received from trustworthy outlets. Kiwi.com has had positive reviews from across the world, including publications like the New York Times. Once we checked screenshots of our competitors and saw that no one else was leveraging such social proof, we decided to give it a try.
We saw an incredible 30% increase, even after such a minor change.
This feedback encouraged us to experiment more; this time with an app preview video which was auto-played in iOS11.
The problem was that we wanted to show the New York Times testimonial in the search results as well as the new preview video at the same time. Therefore, we created new versions of the screenshots in portrait orientation while keeping same colors. We also included additional testimonials, including one from Business Insider, that would address frequent business travelers.
This experiment resulted in an additional 29% increase in conversions.
c) Further tests and learnings
We focused extra attention on the localization of screenshots as well. This time, besides just translating the dominant one-liners, we also decided to fully localize the appâs contents as they were visible on-screen; also showing regional airlines and relevant currencies.
One of our experiments in France and Japan also included culturalization, where we attempted to tailor the listing to local preferences. It was a great idea, but the results were inconclusive, so we didnât consider it worth pursuing further.
Dwell on details - Even a seemingly minor change like using a testimonial from a trustworthy outlet can have a big impact on conversion rate.
Think outside of the box - Many apps are just following store listing of others which creates something like a loop and results in all of them looking very similar. This can be a great opportunity for you to create something that will stand out.
Donât stop - If some of your experiments have little impact on performance, you didnât fail, you just found what doesnât work. Keep generating new ideas and donât be scared to test. Sometimes they will work, sometimes they wonât.
Take your time  - Our experience suggests that you need to run a decent amount of experiments to find out what works best for your app. Factor this into your plans.
The potential impact of App Store Optimization strongly depends on the niche you are in, if it has active user base or not, if users are actually searching for apps like yours and the level of competition.
Being aware of the quality of your app is crucial - ASO is influenced by reviews, ratings and other in-app factors!
Keep in mind that ASO is not just about keywords and screenshots - itâs about optimizing everything. You need to improve the user journey between the store listing visit and a successful app install. If you have a game, look on your APK/IPA size to increase the download rate. When you start with a new product, include keywords in the bundle ID and in the Play Store. There are many small wins that compound if you take your time with them.
ASO is a great strategy for both big and small publishers. Some apps can leverage it more and some less, but for many itâs a great strategy for attracting the most relevant users.
Even when you stop actively doing ASO, you will still benefit from free downloads which can last for months or years. Itâs a great advantage when compared to paid campaigns.
We would like to thank Kiwi.com for giving us such a free hand to test different strategies and grow their business. Because testing is what ASO is about!
Jiri Chochlik and the AppAgent team
