How to work inclusion in Product Development?
Written by — Rodrigo Passeira, Head of Product at Sinch
INCLUSION IN PRODUCTS
How to work inclusion in product development? This is an important question that every person working in product development should ask when thinking about creating and developing a solution or physical or digital item.
We can approach the inclusion theme in several ways: talking about technological, social, educational inclusion, among others. Regardless of the focus, in all definitions the most emphasized aspect is always worth repeating: the intention is to think about the product with a focus on satisfactorily serving all people who want to use it, without exception by social class, education or any other characteristic of the product. individual. As obvious as it may be, clarity on this point is essential for us to follow.
Exercising with an example, a face detection product must consider, without exception, all ethnicities, being prepared to recognize people with skins of different colors, shades and ethnic traits in a balanced way so that the algorithm does not have bias or fails to detect a face that is not of the X or Y color. It seems obvious, but it is not what actually happens in reality: headlines appear constantly talking about the theme, including names of major players in the technology sector.
But how to reduce this possibility of bias and what are important points to highlight and be considered in the product construction process?
Here are some points that can help companies create more inclusive products.
1. Company Culture
The company’s culture is purposefully the first item in this text, because without a strong, structured culture that values people, thinking about inclusion and diversity in their teams, there will hardly be a positive evolution towards inclusion.
At Sinch, we have our culture focused on people — customers, candidates, employees, and all others! For its importance within the expected behavior of all people who work in the company. Our values are also very important to us, listed in the sequence below:
1- Dream Big
2- Win Together
3- Keep it Simple
4- Make it Happen
All four items above also encourage inclusion in our daily lives, in the most varied processes and activities in the areas. We value Innovation, work with a focus on the customer to deliver valuable experiences to people, and we are data-driven people to achieve our results, reflecting behaviors that can and should involve inclusion issues in day by day.
2. Selection process
In order to maintain the culture and avoid bias, this second topic is important, because it is the point of entry for people in the company.
It is known that the selection processes based on the evaluation of variables such as the curriculum seeking specific training, Faculty X or Y, and other details can also be loaded with the unconscious bias of the people they will recruit. It is common to see patterns being followed in hiring, because people often, often through an unconscious bias, look for people who look like them. This is a big problem and can cause serious impacts on an organization, when forming, for example, little different teams, with similar people, who have the same repertoires and think the same way. When we think about product and recruiting similar people for a team, we will only have a view of the problem, questioning and unfolding hypotheses can become addicted and with that the team ends up being more limited, which can culminate in a delivery with less value for customers.
Selection processes that use structured steps with chatbots, for example, help to eliminate bias, balancing the process, because the question will always be asked in the same way for the person applying, while at the same time for those who recruit it is possible not to show no data from the application that may generate bias. The view of who the person is, their physical characteristics, where they studied, age, gender, etc., can be untied from the answer, so the analysis is only about the skills, knowledge and experience that the company is looking for.
3. Simplify access to product use
Allowing the use of a tech product without the need to use code can also be a way of giving people more access to technology. Developing a chatbot without using a line of code, for example, is a way to bring more people close to the product and help those who are in the beginning of their careers and want to learn more about development and technology, for example, before they have experience and a lot of knowledge in the area.
At Sinch we have a bot development platform that allows people who don’t know how to use a programming language to develop their own bots without the help of an expert. With that, it is possible for people with different backgrounds to be able to develop new bots, and also contribute with new visions for the solutions. This makes the product team better understand how these new people relate to the platform and what kind of pains they are trying to solve. In other words, all this because the more simplicity brought to the development process, the more scalable the product will be.
4. Gender-neutral language (Non-binary)
People should feel comfortable with the product they are going to use. Non-binary language, that is, that uses neutral terms, without distinction by gender, is an important point of attention for the construction of the solution. Inclusive language ensures that everyone feels covered by the product, not just a specific group.
For products that use voice, there are super interesting voice projects without gender, for example, that mix voices of several different frequencies to achieve a mixed result between bass and treble, not being categorized as of a specific genre.
For Chatbots, non-binary language is essential for inclusion and ensuring that anyone who reads feels that the bot is talking to them. Many people do not realize it, but on a daily basis they use several masculine adjectives for definitions, and in a conversational flow, as in any product, it is essential to always think about leaving the text neutral, using gender neutral language.
5. Use of Conversational Channels
Have you ever stopped to think that, in Brazil, WhatsApp is in 99% of cell phones and that this channel is an important point of contact for people with different products and services? Once we consider this information, having a product or being a company that provides channels that communicate massively with the public is also a form of inclusion.
A great example of this is Aprendizap, a chatbot developed jointly by Sinch and Fundação 1Bi, a social institution of the Movile Group. The product is a free tool that proposes to extend the classroom using WhatsApp for public school students. For this, the bot has four study trails: Fundamental II, Enem, Professores e Tecnologia. Considering the target audience of the product, there is an important aspect to consider: most have limited access to the internet, and the use of the data plan can directly impact the decision to use the product or not.
With that in mind, the solution was developed focusing on a communication channel that, in addition to already being present in practically all devices in Brazil, also acted on the basic premise that any content, such as videos and files, should be sent using the messaging application, so that the user would not have their data plan impacted, and even more people could use the product.
To get an idea of the size and importance of bots in education, a survey released this month by Undime (National Union of Municipal Education Directors), with support from Itaú Social and UNICEF, reveals that 96% of municipal education networks are using among other actions, educational platforms that use technology to teach students during the pandemic. Due to the coronavirus, at least 590 thousand students in the country had classes suspended.
The result was incredible, and during the pandemic there was a 5000% increase in the volume of AprendiZap chatbot users.
6 .Data, focusing on people (without exclusions)
In the book Invisible Woman, by the author Caroline Created Perez, there are several examples of how society and studies are usually based on predominantly male samples or ignoring important points of the analysis, which in turn generate bias and even serious problems, such as lack of knowledge of heart attack symptoms in women. Okay, but how does it impact products? A LOT :) If the sampling used to define an algorithm uses only one specific profile, we will have problems in recognizing important points such as heart attack symptoms in women, which may be different from the symptoms of men.
At various times, the way the data is collected and structured can generate a very large deviation in the results, and with that bring problems and fail to include people who should be contemplated. So it is important to think about the product and the database used, taking great care during product development, always listening to the people who will use it to create solutions that meet their needs, a basic process for product development, but that many is often forgotten.
At Sinch we have a case that symbolizes this theme. Using an algorithm that highlighted the main topics that were treated by users with Bot, we noticed important points after a massive communication campaign using text messages via WhatsApp. The number of bot response messages in audio format started to increase, and going into more detail, many of these messages contained the same text after processing the audios and translating them into text: “I can’t read”.
In other words, this showed us that a relevant volume of people was being impacted with text messages and was unable to understand the company’s communication with them effectively. Thinking about how to solve this point, we promptly started the development of a solution to use audio instead of text and whenever we detect the message “I can’t read” in the text, the bot starts to follow the audio stream, helping to include these people who previously couldn’t communicate with the company.
These are some important points to always take into account when creating and developing products. A process of selection and management of a strong culture, with products that serve all people, taking proper care in communication and using the correct data, always thinking about the people who will actually use the product, attending to the pain of these users, always will make the product more inclusive, which will bring more satisfaction and value for customers and users, consequently better performance for the product and for the company.