Thursday, June 16, 2016

The Offline Internet

Hike, India’s homegrown answer to the ubiquitous Whatsapp Messenger, launched Hike Direct in end 2015. Touted by CEO Kavin Mittal as an ‘incredibly powerful’ breakthrough technology, Hike Direct allows two users of the app who are within a 100 metre radius to chat and share pictures, stickers and files at high speed without using their data connection. Hike Direct uses Wi-Di technology that bypasses the slower Bluetooth connections, allowing data transfer speeds of upto 40 Mbps. Effectively, you can share a 100 MB file with a friend in 10 seconds, and neither of you will use a paisa of your data plan.

However, Hike is not the first mover in the offline game. That crown goes to ShareIt, which has consistently ranked amongst the Top 5 apps in the India Google Play Store since the past few years. ShareIt allows users to transfer files, movies, music and even apps to each other without a data connection, phone to phone. And like Hike, it’s completely free to use.

Both of these apps are examples of what I like to call the Offline Internet, which is a real phenomenon to reckon with in the Indian market. The Offline Internet is a network of mobiles that are a repository of content – games, movies, music, pictures – which gets shared between users at minimal data cost. The Offline Internet represents the way that the Next Billion mobile users are discovering, sharing and consuming content, bypassing restrictive data plans.

It’s not hard to understand why the offline internet exists. Put yourself in the shoes of the young, enterprising consumer who has just purchased his/her first smartphone. What is the first thing you need when you buy a new phone? Apps, music, games, movies – all the stuff that makes your device a personal entertainment system. How do you get it? Do you spend further on expensive data recharges to download everything or stream it off the net? Do you make the journey to a shady ‘download store’ to fill up your SD Card? Do you fiddle with a PC and cables? Or do you just ask your cool friend who has everything on his phone already, to share it with you – and in the bargain, help set up your phone with the latest apps that you need to have? Obviously, that’s the easiest and fastest way to get up and running.

The reluctance to use up data, so alien to the affluent always-connected Indian, is real and palpable, even amongst the allegedly internet-addicted Indian youth. The student will switch off data during classes and switch it on when travelling back home. Call centre employees switch off their data at work (where mobile phones are often forbidden). Many people paradoxically switch off data during travel, or at times in the day when they simply do not want to surf. Even Whatsapp and Facebook can wait till it’s the right time to check them again.

India ranks amongst the countries with lowest cost of mobile usage in the world:




(Source : World Bank Blog


However, it’s a different story when we look at the cost of data vs income. The Broadband Commission for Digital Development, sought that by 2015, broadband services should be available in all nations at 5% of Monthly Average Income, or lower, in order to be affordable. 

However, India, like many other nations with a large number of poor people, faces challenges in achieving this target.



(Source : Measuring The Information Society Report, 2015 by International Telecommunications Union)

In a post entitled The Data Trap, the JANA blog points out that at India’s hourly minimum wage of 20 cents/Rs.13,  it would take 17 hours of labor to pay for 500 MB of data. Hopefully, mobile data prices are continuously falling (and minimum wages are rising). Still, these two sources point to the extent of challenge in getting people online – and keeping them online for sustained period of time. Either affordability has to improve, or people have to see a benefit of being online.

What are the implications of the offline internet for marketers and app developers?

  1. Make apps small. According to VC Firm Lightspeed Ventures, the ideal app size for markets like India is below 5 MB, vs 10-15 MB globally.
  2. The crucial elements of your app experience should be able to function offline or in a hyrid offline-online environment. This is however, carries a difficult trade-off – to work offline, an app needs to store data on the phone and most budget smartphones today are space constrained. It’s common for people to bump off apps that they don’t use, especially if they are hogging handset space. Google has acknowledged the need for offline internet by making YouTube an offline functionality in countries like India.
  3. The example of ShareIt demonstrates that shareable apps are a workable concept in India. People who procrastinate on whether to download a new app, may be willing to use it if they can get it free from a friend. For the developer, this means letting go of control – and also losing out on Play Store download stats. But I am sure that it’s still possible to keep a track of shared API’s. Or to partner with ShareIt to be a featured app and keep a track of shares!


No matter what you choose to do, the offline internet is a phenomenon that we cannot ignore, especially as internet penetrates into the lower pop strata.

Monday, June 13, 2016

Artificial Intelligence is the New Black

"Over time, the computer itself — whatever its form factor — will be an intelligent assistant helping you through your day. We will move from mobile first to an AI first world” – says Google CEO Sundar Pichai.

Artificial Intelligence is the technology of the future that Google, IBM, Amazon and Microsoft have been investing in for years. It represents a future that could leave Apple behind if it takes off. But what exactly is Artificial Intelligence? Is it like those computers that beat Vishy Anand at chess? And if that’s the case, how exactly is it going to change my life?

Voice Assistants like Apple’s Siri, Microsoft’s Cortana or Google’s newly announced Assistant represent the tip of the iceberg where Artificial Intelligence is concerned. To be understood – and to trigger actions based on voice – is not as easy as it sounds, given the huge variety of languages, dialects, accents and individual pronunciations that exist on this planet. If we have made progress in this area, it is because of the increase in the global user base, which has made continuous learning and refinement possible.

But where AI can play a much bigger role is in handling the vast, unimaginable dump of data that has been generated across millions of websites, apps, social media sites and internet users, across the last decade or so when internet usage accelerated sharply.

IBM points out on its website that 90% of the world’s data was generated in the last two years, and 80% of that data is unstructured. Whether it is organisations seeking to organize and structure their own internal archives or a website seeking a deeper understanding of their own user data – Artificial Intelligence can help to make sense of it.

For example, IBM’s Watson Health has partnered with the American Diabetes Association to analyse clinical and research data and create apps that aid doctors in managing the disease. The partnership will give IBM access to 300,000 patient records and 66 years of data. The goal is to use this vast data dump to train Watson to understand diabetes and make data-driven recommendations. Watson Health is already working with Medtronic to build ‘cognitive apps’ that monitor blood sugar level and then automatically adjust the insulin dosage in insulin pumps.

Or take a recent development by the Accessibility Team at Facebook called Automatic Alternative Text – which will begin to ‘read out’ the content of photos to visually impaired users, by ‘seeing’ what they contain eg. Three people smiling. Across Instagram, Facebook and Whatsapp, users upload nearly 2 billion photos everyday, and Facebook’s Artificial Intelligence division is building software that recognizes images automatically, using machine learning. The technology is still at an early stage – it can reliably identify concepts in certain categories like transportation, nature, sports, food and people. It can identify a pizza or a selfie, but not the detailing (eg. Pizza with olives and pepperoni). But it’s still a huge step forward, that Facebook can accurately categorise a range of photos with 80% confidence!

The more gimmicky side of AI has always fascinated people and created great sound bytes for media – Frankenstein Robots, participating in reality shows and beating real life contestants. Now, we also have creative AI - for example, a neural network wrote the screenplay for Sunspring, a sci-fi film that is ‘fascinatingly incoherent’. Meanwhile, Google’s Project Magenta aims to create art using AI – spanning music, videos and other visual arts. Google AI was used to write a (Rather creepy) love poem after reading 3500 romantic books; you can read the text here. One of the line goes like this;

There is no one else in the world.
there is no one else in sight.
they were the only ones who mattered.
they were the only ones left.
he had to be with me.
she had to be with him.
I had to do this.
I wanted to kill him.


However, the new age AI is moving away from gimmicks into a more hardworking space – making sense of complex data, answering more complex queries from that data, and learning to analyse patterns in a more human like, intuitive way. Also, the results of AI are actually directly impacting people’s lives rather than being utilized only at an institutional level. Mobile apps represent a way in which the benefits of AI can literally reach people’s pockets today. We don’t need to read about it, or watch it on TV. We can actually experience it – like the visually impaired people who can now ‘hear’ Facebook photos!