Hello! My name is Natasha and I am a UX researcher at a design, design and research company. In addition to participating in Russian-language projects (Rocketbank, Tochka and many others), we are also trying to enter the foreign market.
In this article I will tell you what you should pay attention to if you have a desire to bring your project outside the CIS or do something right away with an emphasis on English-speaking users, and what is better to refrain from as factors due to which you just waste your time and money.
About research of foreign audiences and useful tools, about approaches to interviews and the choice of respondents, about the stages of this path, about our personal experience - under the cut.
Iβll make a reservation right away that we ourselves are still in the process - we are increasing the audience in , we write about our cases and processes, but so far we are entering the foreign market mainly with the help of familiar guys who either do their own projects there, or know those who do it. Therefore, we cannot specifically tell about the ways of entering the local market. I will describe the steps of directly studying the market, conducting research and designing, if you are doing a project for a foreign audience.
Market research
At this step, there are two main ways - to conduct in-depth interviews and not to conduct in-depth interviews. Ideally, if you have the budget and strength to do so. Because in-depth interviews give you a very good understanding of all the specifics of the market in general and the perception of your product in particular.
If you are a little limited in funds, or there are none for this, then you can work without in-depth interviews. In such cases, we do not talk with users according to a pre-prepared methodology in order to find out in detail their path, identify problems, and then, based on this, assemble the structure of the service's functionality. This is where the methodology of desk market research comes into play (read: use of available sources).
The result of this stage is the behavioral portraits of users and current CJM - either of some process or use of the product.
How portraits are made
To create a correct user portrait, you need to understand the specifics of the market (especially foreign ones). When you communicate with real users, you can ask them questions about their experience and problems, clarify how they use the product, where they stumble, what would they advise to improve, and so on.
But this is an ideal situation, and it happens that this is not possible. And then you have to use the resources that are at hand. These are all kinds of forums where users of similar services discuss problems, these are collections of reviews for similar ones to your product (and if the user does not like something, and strongly, he will not regret a couple of minutes to write a review about it). And, of course, there is nowhere without word of mouth and communication with friends in the subject.
It turns out that there are many sources, they are rather scattered, and this is more a quantitative search than a qualitative one. Therefore, in order to get really adequate information in search of portraits, you will have to shovel a very impressive amount of information, and not the most relevant one.
We did one project for the American market. First of all, we talked with friends who moved to America, the guys told us how their friends now use similar services, what they are satisfied with and what problems they face. And at the very top level, it helped us define user groups.
But a group of users is one thing, and portraits are another thing, portraits of more realistic people filled with problems, motivation, values. To do this, we additionally analyzed tons of reviews about similar products, questions and answers about data protection and other issues on the forums.
Where to get useful data
First, you can use specialized question and answer services, such as Quora and the like. Secondly, you can (and should) use what the user himself will use for search - Google. For example, you are making a service for data protection, and you fill in queries that an upset user can fill in when problems arise. The output is a list of sites and forums where the audience you need lives and discusses similar problems.
Remember to use Google's advertising tools to analyze the frequency of use of certain keywords and understand how relevant the problem is. And you also need to analyze not only the questions that users ask on such forums, but also the answers - how complete they are, whether they solve the problem or not. It is also important to look at it in terms of time, if you are doing a more or less technological service, then questions and reviews older than two years can already be considered outdated information.
In general, the criterion for the freshness of such information is highly dependent on the industry. If this is something that changes dynamically (the same fintech), then a year and a half is still fresh. If it's something a little more conservative, like certain aspects of tax or insurance legislation around which you want to build your product, then two-year-old forum threads will still work here.
In general, we have collected information. What's next?
An example of information analysis for one of the requests
Further, all these reviews, requests to the search engine, questions and answers on the forums are divided into groups, brought to one or another common denominator, which helps to fill the portraits with life experience and details.
Their morals
There is another very important thing here. If you are into prototyping, interfaces, research, and so on, then you already have experience. A good experience that allows you to do your job well.
That's what you need to forget about. At all. When you work with a different culture, make products for people with a different mentality, use the data that you have collected, but not your own experience, disconnect from it.
Why is it important. In the case of a VPN service, what is our usual audience for such products? That's right, people who need to bypass the blocking of some site, which for a variety of reasons is now inaccessible from the Russian Federation. Well, IT people and people are more or less aware of the need to raise a tunnel for work or something.
And here is what we have in the portraits of American users - "Concerned Mother". That is, there VPN is one of the tools with which mom solves security problems. She worries about her children and does not want to give a potential attacker the ability to track their location or access data and network activity. And there are a lot of similar requests from users of this category, which allows us to highlight them in a portrait.
Yes, she doesnβt quite look like a worried mother of 40, but we are already tired of looking for a suitable photo on stocks
What does βConcerned Motherβ usually look like in relation to mobile applications in our country? Hardly the same. Rather, it will be a person who actively sits in parental chats and is indignant about the fact that it seems like a month ago they handed over money for linoleum, and tomorrow they need it again. Quite far from a VPN, in general.
Could we get such a portrait in principle? No. And if we started from experience and did not study the market, we would miss the appearance of such a portrait on it.
A behavioral portrait is a thing that is usually formed after research, this is a logical next step. But in fact, even at the research stage, you can benefit from building a service, understanding how people will behave with it. You can immediately highlight the main unique product offerings that will attract a set of portraits. You begin to understand what fears and distrust people have, what sources they trust when solving problems, and so on. All this helps, among other things, to form a textual presentation of the material - you can immediately understand which phrases to use on the landing page of your product. And what is also important - what phrases you should definitely not use.
By the way, about phrases.
Language problems
We did one project with a direct focus on the American market, and not only for IT people, but also for quite ordinary users. This means that the textual presentation should be such that everyone, both IT people and non-IT people, understands and perceives normally, so that a person without any technical background can understand why he needs this product at all and how to use it, how it will solve problems.
Here we conducted in-depth research, this is a standard methodology, you highlight the main characteristics of user groups. But even here there are problems. For example, with a recruit. A foreign user for research costs twice as much as a Russian recruit. And it would be nice if only money - you have to be prepared for the fact that the recruit will slip for research not the American user you need, but those who recently came to live from Russia to America. That completely knocks down the scope of the study.
Therefore, it is necessary to carefully discuss all the conditions and exceptions - which user is needed for the study, how many years he must spend in America, and so on. Therefore, in addition to the usual characteristics for the study, it is also necessary to enter detailed requirements for the respondent in terms of the country itself. Here you can directly prescribe that you are looking for people with such and such characteristics and interests, while they should not be immigrants, should not speak Russian, and so on. If this is not noted immediately, then the recruit will follow the path of least resistance and push you to study former compatriots. It is, of course, good, but it will lower the quality of the research - after all, you are making a product aimed specifically at Americans.
Data-driven CJM - the current process for addressing data protection in the US and EU
Language is also not so simple. We know English well, but we can still miss some moments just because we are not native speakers. And if you make a product not in English, but in some other language, it is even more difficult. Hiring a foreign freelance writer is not an option. We once hired a translator from Thai to work. Good experience. Now we know for sure that we will not do this again. It took us 3 times more time, we collected 5 times less information. It works like a broken phone - half of the information is lost, another half does not reach, there is no time left to deepen the questions. When you have a lot of free time and nowhere to put money, that's it.
Therefore, in such cases, when you are preparing something for a similar market, it also helps to study issues in English - its universality makes the same English-language resources the main source of information for such countries. As a result, you can successfully get portraits, and CJM, which the user goes through, and actions within each stage, and problems.
CJM based on a full study - one of the portraits of B2B exporters, ASIA
Studying problems is important in principle, because people go to discuss a situation in which they pay money for a service, but they continue to encounter problems. Therefore, if you make a similar paid service, but without such problems, in general, you understand.
In addition to problems, you should always remember about the possibilities of the service. There are features that form the framework of your service as a whole. There are some non-critical features, additional goodies. Something that can become that small advantage, because of which, when choosing from similar products, yours will be chosen.
Design
We have portraits and CJM. We are starting to build a story map, a product part, about how the user will walk inside the service, what functions to receive in what order - all the way from the first recognition to getting benefits and recommendations to friends. Here we are working on the presentation of information, from landing to advertising: we describe what terms and what you need to talk about with the user, what catches him, what he believes in.
Then we build an information scheme based on the story map.
Part of the functionality of the product - one of the scenarios in the infochart
Yes, by the way, about design, there is an important trifle. If you are making an application or site not only in English, but for several at once, start designing with the most visual βuglyβ language. When we made services for Americans, Europeans and Asia, we designed all elements first in Russian, with Russian names for all elements and Russian texts. It always looks worse, but if you designed in Russian so that everything turned out fine, then in English your interface will generally be excellent.
The well-known property of English works here, that it is simpler, shorter and more capacious at the same time. In addition, it has a lot of native elements and button names, they are well-established, people are used to them and perceive them very unambiguously, without discrepancies. And there is no need to invent anything, because such inventing creates barriers.
If there are large text blocks in the interface, then all this is necessarily read as a native. Here you can find people on sites like Italki, and ideally build yourself a base of people who will help with this. There is a cool person who knows the rules of the language, grammar and so on - great, let him help with the text as a whole, help correct the little things, point out that βThat's not how they say itβ, check idioms and phraseological units. And there are still people who are exactly in the topic of the industry in which you are making a product, and it is also important that your product speaks with people in the same language and in terms of the characteristics of the industry.
We usually use both approaches - the text proofreads the native, and then a person from the industry helps to land it on the product area. Ideally - two in one, if a person is from the sphere and at the same time with the education of a teacher and good grammar. But he is one in five thousand.
If you have done a good research, the most typical and used phrases and expressions will already be in your CJM and portraits.
Prototype
The result is a designed prototype, a very detailed communication scheme (all errors, fields, push notifications, emails), all this needs to be worked out in order to give users a product.
What do designers usually do? Give multiple screen states. We create a holistic experience by carefully working through all the texts. Let's say we have a field in which 5 different errors can occur, because we know the logic of how users work with these fields well and we know exactly where they can make mistakes. Therefore, we can understand how to validate the field and what kind of phrases to communicate with them for each error.
Ideally, one team would work through your entire communication scheme. This will allow you to keep the experience consistent across channels.
When checking texts, it is important to understand that there is a researcher who was involved in building a portrait and CJM, and there is a designer who does not always have the experience of a researcher. In this case, the researcher should look at the texts, evaluate the logic and give feedback on whether something should be corrected, or everything is OK. Because he can try on the resulting portraits.
And this is one of the portraits for the EU financial service, created on the basis of interviews with users
For the same service with a more creative bent
Some are accustomed to making a design right away instead of a prototype, Iβll tell you why itβs still a prototype first.
There is a person who thinks through the logic, and there is a person who does it beautifully. And everything would be fine, but between logic and beauty, there is usually the fact that the customer rarely provides full technical specifications. Therefore, we most often have a prototype - this is a kind of task for analysts or those who will program the product. In this case, you can understand some technical limitations, understand how to make a product for the user, and then communicate with customers on this topic, convey to them what things can be considered critical for the user.
Such negotiations are always a search for a compromise. Therefore, the designer is not the one who took it and made it awesome for the user, but the one who managed to find a compromise between the business with its capabilities and the limitations and desires of the user. For example, banks have restrictions that cannot be bypassed - as a rule, it is not very convenient for the user to fill in 50 payment fields, without them it is more convenient, but the Bank's Security Council and internal regulations will not allow to completely move away from this.
And after all the changes in the prototype, a design is made that will not undergo any major changes, because you fixed everything at the prototype stage.
Usability test
No matter how well we research our audience, we still test designs on users. And in the case of English users, there are also some peculiarities here.
For the simplest portrait of a foreign user, recruiting agencies charge 13 rubles or more. And again, we must prepare for the fact that they can slip someone who does not meet the requirements for this money. I repeat, it is critical for respondents to have a cultural code and features of a native.
To do this, we tried to use several sources. First Upwork, but there were too many narrow specialists and few people in search of not the most skilled work. Plus, everything is strict there with requests, when we wrote directly that we need people of a certain age or gender (there should be a distribution in the samples and characteristics - so many of these, so many of these) - we snatched out bans for ageism and sexism.
As a result, you get a double filter - first you find those who meet the given characteristics, and then you manually filter out those who do not fit by gender and age, for example.
Then we went to craigslist. Waste of time, strange quality, no one was taken.
A little desperate, we started using dating services. When people realized that we didn't want exactly what they wanted, they complained about us as spammers.
In general, recruitment agencies are the most efficient option. But if you get around its high cost, then it's easier to stop at word of mouth, which we did. We asked our friends to put up ads on university campuses, this is a normal practice there. From there, the main respondents were recruited, and someone asked colleagues for more serious portraits.
In terms of the number of respondents, we usually recruit 5 people for each dedicated user group. Eat Nielsen Noman, which shows that even testing on groups, each of which has about 5 high-quality respondents (representative), removes 85% of interface errors.
And we must also take into account that we conducted the testing remotely. It is easier for you personally to establish contact with the respondent, you follow his emotional manifestations, notice the reaction to the product. Remotely it's more difficult, but there are pluses. The difficulty is that even at a conference call with Russian guys, people constantly interrupt each other, someone may have communication problems, someone did not understand that the interlocutor would start talking right now, and started talking himself, and so on.
Pros - during remote testing, the user is in his usual environment, where and how he will use your application, with his usual smartphone. This is not an atmosphere of experiment, where one way or another it will be a little unusual and uncomfortable for him.
A sudden discovery was the use for tests and product demonstrations through . One of the problems with product testing is that we can't just share it with the user - NDAs and the like. You can't give a prototype directly. You can't send a link. In principle, there are a number of services that allow you to connect a string and simultaneously record the user's actions from the screen and his reaction to it, but they have disadvantages. Firstly, they work only on Apple technology, and you need to test not only for it. Secondly, they cost significantly (about $ 1000 per month). Thirdly, at the same time, they can also suddenly become stupid. We tested them, and sometimes it happened that you were doing such a usability test, and then all of a sudden, after a minute, you werenβt doing it anymore, because everything suddenly fell off.
Zoom allows you to share the screen with the user and give him control. On one screen you see his actions in the site interface, on the other - his face and reaction at the same time. Killer feature - at any moment you take control and return the person to the stage you need for a more detailed study.
In general, while this is all I wanted to talk about in this post. If you have any questions, I'll be glad to answer them. Oh, and a little cheat sheet.
Tips
- Study the market in any case, with or without a budget. Even asking Google, as a potential user of your service would do, will help collect useful data - what people are looking for and asking, what annoys them, what they are afraid of.
- Chat with experts. It all depends on social capital, whether you have people around you who can help validate your ideas. I somehow had an idea, I was going to write an article, collect feedback and test the product, but I asked 3-4 questions to a familiar expert. And I realized that I should not write anything.
- Make interfaces first in an "ugly" language.
- Proofread with natives not only grammar and so on, but also the relevance of the industry in which you launch the product.
Tools
- for testing.
- for infoschemes and design.
- β service analogue of gravedit for English.
- Google for Market Understanding and Requests
- (formerly RealtimeBoard) for story map
- Social networks and social capital to find respondents.
Source: habr.com