ASOdesk talked with Ilya Kukharev (pictured), Aviasales ASO optimiser. Ilya described what tools are used, what to read to learn ASO and whether it is worth chasing popular requests.
Ilya, hello! Tell me about yourself, how did you get into Aviasales?
For about 10 years I have been freelance translator, translating applications, localisation, that’s all. Last year, Aviasales began to develop the second Jetradar application. This is almost identical to the Aviasales application, but it is designed for Western, Eastern, Asian and many other markets. Why did they choose this? Jetradar is more understandable for a foreign user than Aviasales.
So, friends invited me to Aviasales, and I always wanted to do translations, and here is a mobile, such a company. Naturally, I immediately agreed. For almost a year now we have been working on this. In my department I am alone, marketers and a large team of freelancers who help me translate everything I need help with. This is about three hundred phrases, and my task is with the help of services, to pick up words that we will use in the Store. Here you need not to confuse translation and localisation. Localisation includes translation, but much other work on adapting the application for a particular market.
To select words I use Sensor Tower, this is the main service. Also, I use ASOdesk, it’s convenient because no one else shows the number of users by the keyword. Other services provide a certain traffic factor, and there are no specifics. For statistics I use AppAnnie.
What were happening to ASO in Aviasales before you?
With the Aviasales application, everything was not bad, since they worked only in Russian. Clever people selected keywords. With my arrival, the figures improved, just because a separate person for ASO appeared. As for Jetradar, they were not engaged in anybody at all, everything was terrible. Everything was in the state of “made the release of the application, entered the words that first came to mind and thats it.” There was no traffic, nothing happened. A sad situation, but I managed to fix it.
Oh, share successful cases!
In the autumn of last year, we decided to go to the terrible markets for China, Japan and others. Scary concerning written signs – hieroglyphics. So, in China, the application exploded and the set of keywords that was selected brings a large number of downloads, say, 20-30% lower than Aviasales receives in Russia. At the same time, there was not a dollar spent on marketing in China. The only cost is localization of the application. This is with a relatively small spread of iOS in China – about 25%. The rest is Android, and not on Google Play, but on the Chinese system with other stores. I’m proud of this case – worked for a couple of months and got good results. Also, I’m continually testing different items in a stack. By the way, without tests it is better not to change anything at all. At me tests go practically without a stop. I recently tested a short description for Hotellook on Google Play, the test showed an increase of 20%. We will see.
Also changed the screenshot of the issuance in Aviasales, set the route only in one direction, from this the price is lower and put different airlines. Accordingly, the user sees a reasonable price and choice of the carrier – so you can download.
Is there a travel-specificity in ASO? Does the set of words vary depending on the season?
Most often not, especially since itunesconnect does not offer as much space for keywords. Although we see the dynamics of downloads, which varies depending on the season. For example, in January, a good growth, now (March) is already falling. In the summer it will be higher, in autumn it will fall again. For keywords, I do not take into account the seasonality, maybe worth a try.
Tell us about the struggle with competitors.
I try to fight with competitors. They are not very many, in contrast to the gaming industry, where on request “poker” crawls out a few thousand games. On request, “tickets” crawls out 10 applications with which you can compete, but it’s hard. Also constantly get out applications on request Aviasales, most often games. We complain about them and they take our company out of the keys. There are applications made based on our SDK – they also use Aviasales for promotion. This can not be done and we make sure that they do not use this key. As for the competitors, for example, in Jetradar I did not take such high-frequency words as “air tickets”. I have specified [their words] and collected all the traffic that goes alongside the request “air tickets”. For example, Skyscanner can afford not to write anything extra in the name and get all the necessary traffic. We have to choose less high-frequency words, collect traffic from them and build up the base. For example, not flights, but airline tickets. Only then go out somewhere higher. The same situation with our Hotellook application, where I encounter serious competition. The users got the impression that hotels are only Booking. We try to destroy this stereotype, to show that booking is not always cheaper. We show more interesting prices and from this we play.
How and where did you learn ASO optimisation?
I’m more than confident that 100 people in Russia had experience in ASO. Unfortunately, some people do not even believe in it, although in the US this sphere is developing and has moved to the advanced stage already five years ago. When I was offered to do this, I prepared for another month. I read a lot of material, including ASO Bible (Gabriel Machuret, 2014) – the very first book in this field. If someone wants to do this, then this is the only book worth reading. The algorithm is always changing, so the book may already have irrelevant information. I also signed a video course on Gabriel for ASO. I read everything that is possible by searching ASO in Google. I gained the basic knowledge base and continue to recruit it. It seems to me that ASO is such a topic that you can not learn completely. Every Friday I open my bookmarks and read new articles, starting with the blog Sensor Tower, Trademob, AppAnnie, AppFollow, Moburst, etc.
Look, with ASO-optimization for iOS everything is clear how to track keywords and how to influence their progress. What about optimising on Google Play?
I started playing Google Play about 3 months ago because the App Store was more interesting to me. Google uses old-school SEO with keyword density. It seems to me that you need to work better with the icon and the name because this is the first thing the user sees, there is nothing to lure. Even if everything is done according to the rules, there is no guarantee that it will work. It is necessary to increase the reference mass, increase the authority of the page, update the application as often as possible. These are the classic tips that we use, the results are there. Again, add localisation. I suddenly discovered Brazil, and Germany, on the contrary, did not go very well, because it uses iOS more. Here the main patience, there will be no immediate effect.
What, in your opinion, are the myths and mistakes of ASO?
At the current stage of the development of mobile shops, the mistake is not to deal with ASO at all. I’m surprised, but such people still exist. There is the opposite myth that if you make ASO, then immediately you will go to the top and everything will be fine. No, it will not. One ASO here can not do, you need a competent marketing strategy. Especially ASO never ends, they need to engage regularly, then there will be a result. Very often the developers score on the reviews. On Google Play, they can be answered, and there is a chance that the user will correct their score. It is necessary to respond to the negative and help users if they have something that does not work in the application. There is also the myth that if you found a popular query, anyone, then you can use it, even if it is irrelevant. As for us, for example, the query “aviasales” use games. Irrelevant inquiries will not do anything well, and you may be refused at the stage of consideration.
Thank you very much for an exciting interview and good luck!
Interviewer: Diana Sedunova.