If you have infinite resources, and the year is 2025, how would you improve Google Translate?

  Google
Add Your Answer
Answers (5)

Framework – Ask clarifying questions, identify customer and understand/prioritize pain points, Pick a pain-point and provide solutions and evaluate trade-offs.

Clarifying questions:

  1. Any limitations? – No
  2. Are we allowed to integrate with other Google products or expose features using API or just focus on Google translate home page? – You choose
  3. Can we focus on mobile/web and any device – sure..
  4. Explain why Google translate is important and why Google is investing in it. I will align this with Google vision and make sure my understanding is correct.
  5. Goal? – Improve customer satisfaction with translate

Customer segments:

  1. Students
  2. International Travlers
  3. Professionals

For this exercise, I want to focus on International travelers..

Pain points:

  1. Difficult understanding local languages (both oral and written) and communication barrier
  2. Difficult understanding idioms and gestures that are different in different part of the world.
  3. Difficult navigating websites that are in foreign languages for booking tickets and gathering info that is needed as part of traveling.

For this exercise, I want to focus on 1st and 2nd pain point. Also, it will be difficult to attack all languages at the same time. So I would start with top 5-10 languages used around the world and then, extend it to other languages slowly.

Here are some of the solution to solve this pain-point:

WHAT is Google Translate today?

GT is a free service that instantly translates words, phrases, and web pages between English and over 100 other languages. It has a standalone web app and is embedded in the Google Search, including Google Lens (Image Search), Chrome Browser, Gmail and Google Assistant, etc.

WHO uses Google Translate and WHEN?

If we look at the user by their roles:

  • Students, when they learn new vocabulary, they sometimes use GT or Search
  • Professionals, when they read or put together a document in a foreign language
  • Travelers, when they are browsing websites or information in the local language when planning for the trip, and when they are traveling in foreign countries.
  • For those who are viewing video content in a foreign language on Youtube, Netflix, translation algorithm can also be involved when users chose a subtitle title in a different language

If we look at users by geography:

  • Existing users in America, EU, Asia: Currently Google Translate is primarily used by people who can access the internet, and since the majority of content on the internet is in English, Google Translate is primarily used by people whose first language is NOT English to access English content and services.
  • Billions of new users from emerging countries such as Africa, South America where access to the internet will be dramatically improved in the next 5-10 years, whose native language is not English. Their first device will be primarily a mobile phone, instead of a laptop which is less portable and more expensive.

Are there any pain points for users today?

  • Translation of local slangs in casual conversations is not accurate enough; language is changing at a much higher velocity today and the translation needs to keep up with it
  • Cannot distinguish translating for formal use vs casual settings
  • When translating the entire website, the layout/UI often got messed up
  • For travelers in foreign countries, it takes too long for users to translate while they talk to local people who speak a different language
  • Translation for videos, audios, images are not widely available on Youtube, Netflix, Podcast, etc
  • Reading translated subtitles while watching videos is not easy

As PM for Google Translate, I would like to focus on the last two pain point as if solved it will generate the highest impact over the largest number of users, who are primarily

Ideas to Improve Google Translate

  1. Enable more language translation in subtitles for videos on Youtube
  2. Provide API services to allow videos, audios on other platforms to be translated into a different language by machine
  3. Allow user feedback to improve the accuracy of translation
  4. Make the google assistant on the phone a simultaneous translator from video/audio to text while the audio clip is playing in any apps

I would like to prioritize the idea with four criteria:

  • Reach
  • Impact
  • Engineering Effort

I would prioritize the Google Assistant idea because it comes with each Android phone by default, and it’s very accessible in nature. If the translate capability can be seamlessly embedded into Google Assistant, it will dramatically improve the user experience.

And if Google Assistant become the go-to application for any language translation, we will be able to create a richer profile of the users’ language capability and language needs, so that we can further improve our algorithm and provide a more personalized experience of the user.

I’ll start with a few clarifying questions.

  • When we say infinite resources, can we assume these resources in terms of both monetary and technology advancements? – Assume both
  • Do I have to think about how would I improve Google Translate by 2025 or that we are in 2025 right now and need to improve Google Translate from there on? – Assume later
  • If we are already in 2025, I would love to understand where do we stand currently – how does the product look like at the moment? – Text box translation, web page translation

Next, I would like to come up with a framework. Google Translate can be improved in two directions:

Expanding the languages to cover the entire world

Making the translation available for all languages in the world, including regional dialects and accents – this will lead to seamless use of Google Translate in any part of the world

Expanding the use cases within a language Let’s think of the core use case of Google Translate – it breaks down barrier and essentially translates one language to the other

I would like to drive improvements across two domains – functional use-cases and practical / emotional nuances

a) Functional use-cases

From a users’ perspective, we engage in communication in two ways – delivering and consuming. I would like to break down the delivery and consumption in their respective modes

Consumption

  • Reading
    • Hardware – We can power the Google glass with Google Translate so that when anyone is reading a different language, the text immediately gets translated. Assume a world where a person lands into an another country and is able to engage with the surroundings seamlessly.
    • Software – Google Translate would enable the entire web including websites as well as private applications like email etc. On top of that, it can also not just plain text but also what’s embedded in images for comprehensive reading experience.
  • Listening
    • Hardware – Google Translate would enable smartphones / headphones that can be used to physically talk to a person from a different country with real-time language translation and no latency.
    • Software – I would make Google Translate perform real-time translation of voice over internet. This would include listening to podcasts, taking audio calls without any language barrier whatsoever.
  • Watching – Google Translate to perform all three translations – text, audio, video – simultaneously. Imagine an American person watching a Korean movie, in which the characters would speaking in English, the background images/text in the movie will be in English. This would also include translating not just stationary content but also moving.

Delivery

  • Writing – Translate the entire files like documents – typed / scanned. This will enable ease of cross-border business.
  • Speaking – Real-time transcription of my voice, translation of the underlying text and converting resulting text back into voice. I would broadcast my message by just recoding once in one language and then just translating in different languages. One use case could be world-wide addresses by political leaders / businessmen.

b) Practical / Emotional nuances Beyond functional aspects of a language, I would like to improve Google Translate to cover emotional aspects of the language and communication.

  • Instead of word-to-word translation, I would make Google Translate intelligent to choose the appropriate word in second language, depending on the emotions person wants to convey.
  • I would also ensure that the real emotions remain preserved especially in audio and video translation. Anger / happiness / sadness is expressed slightly different in different language so I would ensure that Google Translate also takes care of this.

This is the future I see for Google Translate with unlimited resources.

Clarifying questions –

  1. In the question you say, its 2025, is it fair to assume technological advancements, faster, cheaper and accurate AI, ML – [Interviewer] Yes.
  2. Are there any form of media (text, Audio, Video) where you would like me to concentrate on ? – [Interviewer] I will let you choose.
  3. While google has infinite resources, is it safe to assume that we have identified Google translate to be a product where we want to invest in this year ? If that is the case, I will skip the opportunity assessment part of the analysis. Is that ok? [Interviewer] – Sure.

High-level Strategic analysis –

  1. Googles mission is to “organize worlds data and make it universally accessible and useful”.
  2. Google Translate as a product aligns very well with the mission, as at its core, it not only parses data in various languages, it also makes it accessible to the world population. Not only that, it gives google access to data which was unknown to one part of the world and now can be used to improve worlds understanding of various phenomenon as well as for commercial activities. (Type format)
  3. From competitors stand-point, today, there are dictionaries, country specific local apps, courses that teach people language which help the cause of translations. If built properly, Google translate has a chance to be the primary thing the world requires to translate languages. (Hardware play)

Approach –

  1. With that understanding, I will start with thinking about different user groups that will benefit from using Google Translate.
  2. Consider their pain points and priorities them.
  3. Find solutions to elevate these pain points.
  4. Then I would priorities implementation and think through few trade-offs / repercussions of the decisions made
  5. Finally think of some success metrics.

Users –

When I think of users who would benefit from using Google Translate the following come to mind –

  1. International Tourists – These users do not understand the spoken or written language, Cannot read signs, Do not know what to order in a restaurants, Often get confused with public transport setup in a new country
  2. Business professionals that work across boundaries – These users normally are working on mission critical projects which affect there everyday life, their personal growth and confidence. The ill-effects of miscommunication or failure here are high both for the user and their employer.
  3. Translators who do the work of translating books and other media – A better translate product has a potential to take away the jobs of these people and they would be good critics to work with to understand the nuances of translating. These folks can improve translate product well by teaching it emotion, tone of voice and other cultural nuances.
  4. Students – These users are learning a n ew language and if the translate product does well, they might choose to pursue some other interest.

Of the following sets of users, I would want to focus on the business professional user as the impact of good work here improves the society in general, it brings economic growth and helps individual feel better, unlocks talent that was surpassed due to language barriers and helps companies reduce cost and improve productivity. It also aligns well with goals mission to make world data universally accessible and useful. Lets consider a few pain points of this user base.

Pain points –

  1. Some employees who are not well-versed in a foreign language do not speak up there ideas in a meeting.
  2. If they do, they are not articulate enough for those ideas to be well understood.
  3. To avoid awkwardness, shame and time required for them to explain the ideas these users normally, explain the idea to colleagues in a local language which is then translated by this third person in the way they understand it. Sometimes the way the idea is gets explained is –
    1. Different than the original intent
    2. Fails to communicate the emotion, passion or the tone with which the idea was originally proposed
    3. Creates a language barrier between team members and arbutus free flow of information
    4. Important information gets filtered
    5. Creates a dependency on a translator (in this case a college) to forward the conversation.
  4. A person who not well versed in a foreign language always has to work twice as hard than a native person to understand documents, audio and videos recorded. The same energy if used for solving problems then merely understanding them can create wonders.

Given that I want to focus on improving conversation when multiple team members are not present in person and are communicating via text ( Google docs / wiki collaboration), audio or video (conference calls).

Solutions –

  1. Integrate google translate in products like Hangouts such that it allows each user to set their preferred language and automatically allows them to speak in a language of their choice as well as listen in the language of their choice. The software should automatically detect the language of the input and based on user language preference provide appropriate output.
  2. Translate should also capture emotion, passion, anger, frustration in users voice and try and convey in to the other end.
  3. It will also help if the product kept a log of all the inputs and their equivalent translations.
  4. Integrate google translate in products like Google docs, text messaging etc such that it allows to set their preferred language such that a user can type in any language of their choice and read comments in any language of their choice.
  5. Finally translate should create a API, which any third party tool can integrate with and use translate in text, voice and video formats.

Trade-offs

  1. The feature might not work really well creating further barriers in communication where the world will stop learning multiple language and all future interactions would be through machines making humans incapable to function without them in international setting.
  2. This feature, might replace a lot of professions around linguistics, an optional language education etc
  3. If the feature works really well, it might fill in the gaps and make an ill-thought plan/argument look good and cause churn which could simply be solved by asking to think more clearly and deeply before an idea is accepted.

Metrics

Since our goal is to improve translate , some of the measurements that would be helpful are –

  1. Accuracy of translation
  2. Level of confidence around how good was the product able to capture different emotions
  3. Speed of translation

Great question, I’ll take a moment to organize my thoughts before answering – Sure

I have a few clarification questions and some assumptions – Go ahead

  • 5 years and infinite resources, I have no constraints in terms of making technology advancements – Yes, no constraint
  • I’ll focus on improving the lives of people and changing the ways people use Google Translate and may expand the use to different products.. not just existing interface – Cool

Let’s take a step back and think why people need translation – okay

  • For business communication
  • To communicate with people who speak a different language, not in business but casually
  • Watch movies, plays etc., basically to enjoy and learn the art and culture of different countries with different languages
  • Understand signs or menus or books in different languages
  • Communicate our feelings and understand someone else’s to/of a person speaking different language

Basically we need translation to communicate and understand different languages.

(Philosophically and passionately) If we could solve this problem we can actually UNITE the world – So what are you proposing (poker face)

Sure I’ll take a little time to jot down some ideas – Okay

Okay, so my proposal for the future of Google Translate is “One World”. The vision of One World is to “Unite the world by removing the language barrier”. Let me explain to you what this product will do – I am excited

Features:

  • Translate the audio from any language to any language (Not subtitles but dubbing)
  • Translate the script of any language to any language (Reading and writing)
  • It will have different settings for example business, casual, urban etc. based on who you are communicating with.

Let’s me walk you through the design and a few use-cases of this product. But before that just wanted to check-in and see if there is any confusion – I think I’ll let you explain the use-cases and design and see if I have any question

Sure, so One World will be integrated with headphones, microphones, and eye-glasses. This way you can speak, understand, read, and type different languages. Now imagine:

  • You are in China and go to a restaurant. You are wearing One World earphones w/microphones. Server comes to take the order and speak fluent Mandarin, but you don’t hear mandarin, you hear casual American english.. No struggle, no confusion… You respond in English and the server hears Mandarin .. friendly, casual, no confusion… Both of you are using One World products – Interesting
  • Someone in a rural part of India.. who only understands or speaks their native language. He can watch any video in any language because he is using One World headset, everything seems natural.. The voice, the dubbing, context.. Emotions, Feelings, Tones don’t get lost in the translation if you use One World
  • You go to Italy and have One World eye glasses.. But you don’t see Italian script at the airport or street signs or restaurant menus.. You just see the language you prefer to read..

In essence, One World will remove the barrier of language from the world… When I imagine One World, I think of a world where you are not lost or need any external help as you can read, write, speak, understand, and communicate in any language.

It’s in contact lenses, eyeglasses, sun-glasses, head-phones, wireless ear-plugs., you even have it in your smartphone..

You can read any book or watch any video.. No subtitles required.. The experience will be extremely seamless.. For eg. all the external noise will be canceled and you hear only the language you selected.

You can talk and write using different settings.. Business settings for businesses, Urban setting to communicate with your international friends in school.. so many possibilities

This is the future I see for Google Translate with unlimited resources in next 5 years – It’s actually a bold vision

Yeah but there could be downsides. The downside I think of:

  • Latency in translation will nullify the experience.. Reading or listening
  • Mistakes in translations during business meeting or even in casual conversations can have very negative consequences for the users
  • It has to be perfect… The only incremental release possible is the addition of languages and the feature of hearing the translation in the voice of the person speaking..