The Vector Algebra of Tier-2 India

A brief history of pain

So we started this prepaid business a few years ago. Back then it was a challenge explaining prepaid to anyone. Anyone living in a metro that is.

But most folks we tend to know are from around where we are. So no one really understood it. My telco friends also hardly understood it, if you ask me. And that’s because in a telco, your phone is actually a biological extension of your ears and mouth. So its neither prepaid nor postpaid, but price-paid. You’ve sort of paid the price for being in the telecom industry. But coming back to prepaid. Most of my initial introductions were “I am doing a usage tracking app for prepaid users in Tier-2 India”.

I usually lost the attention of the listener somewhere in the 8th nanosecond after the word prepaid. I would then be asked “Oh, its an app for college kids” or “Yeah, my maid/driver also uses a smartphone”. After enough such similar conversations, the view on prepaid seemed to be Made for ABCD {Ayaah – Bai – Carpenter – Driver}. This represented the finest perspective of prepaid users in a rather extensive circle of people I interacted with. My attempts to explain that 95% of this country did not consist of ABCDs and College Kids was met with expressions that seemed to ask if I was raised by a group of homicidal quantitative chimpanzees. On my further explaining that prepaid is not an income issue in India, but a mindset of control, the above facial expressions changed from inquiry (#chimpanzees) to one of genuine concern about my well being. But all that is history.

The Geography of Social

We now move to geography. For those who concentrated beyond the 8th nanosecond of my introduction, came a phrase “Tier-2 India”. This is now in vogue really. Much of the tech community talks of Tier-2 India. In fact the Tier-2 India narrative is now defined by three vectors

  1. The Jio Vector
  2. The Vernacular Vector
  3. The Video Vector

As an entrepreneur, I have the highest respect for every tech company that’s made it big. I also respect those who have not yet made it big, but that’s a different kind of respect for another article someday. The former is like Shock & Awe of Bush Jr. But my true role model for conquering Tier-2 India is rather different. I think Baba Ramdev rocks. No gyaan – no Jio, no video, no vernac, no app, no nothing. Pure desi execution. Thataastu Tier-2. If I thought my Telco days were execution, then my Retail days were Ritualistic Seppukku, captured rather brilliantly by Takeshi Miike here, one of Japan’s finest directors. But that’s also for another story.

Coming back to Tier-2 India & Execution, the advantage of Retail and Telecom as industries were this – You do something in the morning, by a few hours you know if its working or not. If you want to be cautious in drawing conclusions, you have all of 24 hours. Within 24 hours, you typically know if you have succeeded or failed. Be it a new promotion on Rice or a new tariff in Raichur. Retail and Telecom are truly, without exception, the school of hard knocks. But what these two industries teach you is something like Google’s “Don’t be evil” motto. The motto of these would be “Don’t chumma BS”.

So, I want to try and put the Tier-2 market narrative in a bit of perspective. I shamelessly present (#drumroll)

Tier-2 India – A Vector Trilogy in 4 parts

1. The Missing Tier-2 Vector : The Vocabulary Vector

We do not live in this Boolean equation “If not metro, Then rural”. The latest definition of rural is rather telling. This is important because once we agree on what rural really is, then we can revisit some numbers. Like the top 300 cities in India have 200Mn population (#2011Census). But when I refer to Tier-2, I am not really referring to these 300 odd cities. Cities like Manipal, Kakinada and Jamshedpur. No.

My vocabulary shifts from the word Cities to Districts actually, and I start to refer to places like Nagaur District, Khandwa District and Mehsana District. Which brings me to the first real aspect of understanding Tier-2 India. Its the Districts that matter. To understand Tier-2 India, one needs to change the lens and vocabulary from cities to districts. This has implications beyond conversational tones – it has a direct implication on data structures in the backend, as well as analytical paradigms. Lest I fall prey to my own BS, the implication of this dimension of Tier-2 = “Do VLOOKUP and add District” (Yeah, I knew you’d get that!)

2. Jio and Jeene Do : The Jio Vector

Jio is seen as unleashing the data revolution in India. Well…No. What needs to be understood is Dial Up modems of 2003. I was fortunate to be in a transitionary world where we moved as a country from Dial Up to DSL. From a pricing standpoint, we (Airtel) moved the market from 25/Hr of dial-up to 999 Rs/mth of DSL within a year or so. The key learning : Speed is addictive (I mean internet speed, ok?). This single three-word-philosophy is behind most broadband pricing. While this learning seems obvious, to experience what addiction truly means, you need to listen to customer-care calls into the broadband call center which are innocently categorized as “Slow Browsing Complaints”.

History apart, what Jio has done is to provide real high quality speed. And in pricing it free initially, Jio has not allowed any Jeene-Do, and instead reduced the market to an oligopolistic one by edging out all other players. It doesn’t take a PhD in economics to know how pricing works in an Oligopolistic market. So the true impact of Jio on users is not about a data revolution, but

  • Reduction in number of operators
  • Addiction to high speed
  • Change of pricing landscape
  • Death of dual sims
  • Higher arpu for operators

While I promise to do a detailed mathematical excursion into the above, a short cut is this view on the Airtel Share Price over the last 2 years.

3. Meine Type Kiya : The Vernacular Vector

There are various forms of self-torture and punishment. One is to watch Meine Pyaar Kiya – a movie where the heroine sings a song to a pigeon; and this pigeon is shown in the final scene as having a flashback while taking revenge on the villain (Scroll to 08:00 for an understanding in self-pain).

Another is to try typing in Vernacular languages of India. Choose your language. And see how difficult it is to express oneself in writing. Hindi itself has multiple keyboards with some fiend deciding to place “ee ki maatra” in different places. And then one thinks of the pigeon in Meine Pyar Kiya. Of course, Google Umaachi (that’s Google Bhagwaan for those of you North of the Vindhyas) has Google Keyboard now, which transliterates from English. So my Meine Pyar Kiya becomes मैंने प्यार किया.

And that should solve the vernac typing problem no? Let’s now think of the License Inspector of Jamui District & the penetration of written English in the country. You sort of get the point. But hang on, that’s not my point really.

My kid is in 5th grade. She speaks English, learns to read-write Hindi & Sanskrit in school, and also speak Tamil at home. I also get such khujli once in a while (I could say Itch, but khujli sounds far more powerful #vernac) and decide that I will learn German or Russian, and download DuoLingo and put fight. It takes a toll. Learning a language is not easy, and that’s because languages are complex. Indian languages more so, with their verbs and adjectives morphing with gender and age, to say the least.

Which brings us to language support. When we were doing our homework for SuddiKatte (earlier called NewsChat), we quickly realized that Karnataka itself has Kannada, Tulu, Coorgi and Konkani as languages that all use the Kannada script. So it’s not as simple as “I support Kannada fonts”, but a deep understanding of the language and its nuances before one can be truly vernacular.

Because if you don’t understand what you are dishing out in vernacular, then you fall prey to becoming a moron doing Lungi Dance thinking its what Tamilnadu is about. But then comes the rather magical argument about Video! (PS : Check out why Lungi Dance is not about TN here)

4. Video killed the radio star : The Video Vector

“Video crosses the boundary of typing” – I want to kill this line of discussion quickly. Let’s get real. Video is rising in four genres. Cricket . Porn . More Porn . Still More Porn (Now you get why Speed is Addictive). The genre of growth is rather language agnostic, if I may.

So, to randomly choose, the rise in the video usage of the youth of Bankura District does not really mean that they are watching Amazon Prime or Netflix. All you have to do is read the reviews of the over 15+ apps that have more than 1Mn downloads on the Google Playstore under the search keyword “Bhojpuri videos”. The rise in video has nothing to do with Tier-2 but a more universal behavioral pattern.

This doesn’t mean that video is not on the ascendant Star (#punIntended), but let’s just be honest and not get carried away with the MBs being consumed. For more solid proof, thisreally happened.

Epilogue : Life at a Bus Stand

And here I am. Two apps both running in Tier-2 India. One in the Prepaid Space and one in the Content Space. Two years ago, I was explaining what Prepaid is Not. Now I am explaining what Tier-2 India is Not.

The next time you hear a Tier-2 narrative, ask someone to show you a bus ticket for a trip from Raichur District to Bellary District (or choose your pick from any of the 707 Districtsof India).

Otherwise just convert Tier-2 to Teri-Toh! (#punJabi)

Press Ok to continue

I make money by matching people living in Tier-2 India to products, using data. This is my revenue model. My business model has a few more rather uninteresting line items in a spreadsheet, one of which will become relevant later in this article. Key to my model is the leverage of data. To get deeper, we use data on smartphones to classify Tier-2 India users, and then match them with products and services we believe they would pay money for. And I make a commission on such transactions.

Note that I said “data on smartphones“, not data from smartphones. Why does this distinction matter? Because my 11 year old daughter likes to swim. Got it? No? Neither did my friendly neighbor VC when he asked me … at least at first.

A late Friday evening in 2013, when we guys at Mubble were doing intellectual gymnastics with an active data collection framework, my wife called and told me that our then 6 year old daughter should learn swimming in the summer holidays. And she asked me to go to Decathlon and get a swimsuit for her on the weekend (specifically Saturday). Like all other extremely simple instructions from my wife, I managed to complicate this as well.

Proud of my Google Search capabilities, which have been sharpened in hours of useless corporate meetings combined with even more useless interests in quizzing… I decided to search for options on a couple of eCommerce sites. A few clicks of filters such as boys/girls, age 5-8, type – full sleeve etc etc, and I locked down to a swimsuit, which was out of stock, and then I promptly forgot all about it. On Sunday morning, after carefully avoiding to remind me on Saturday, my wife triumphantly proved my uselessness at household chores. She went herself to decathlon and got the swimsuit. It is to be noted that she never even asked me why I did not get it, what I was doing instead etc. Such was her confidence in me.

But my troubles were just about to begin. Later that day, in fact from that day onwards, every single website I went to, would show me ads of scantily clad women in swimsuits. No matter what I did, banner ads beckoned me to buy swimsuits and showed me more and more content with rather well endowed women. Dei Google, if you are copying my search, blaady copy my filters also, no? Now be it office or home, I was in the embarrasing situation to explain to anyone who walked by my PC that … yeah, just read this post from the beginning again, I had to explain all of this. Cookies it seems. Try getting cooked by a Jedi stare from your wife.

You would think this was enough for me to awaken to the challenges posed by weak privacy, isn’t it? No. It took another more sinister blow to my ego for that to happen. I was already thinking about privacy, when I had my Pietermaritzburg moment later… around 2014. An old friend of mine, Sumit Dutta, sent me an email saying he was in town and asked me to call him up when free. So I did call him that evening. I have known Sumit since 9th grade, and we have had our fair share of fun times together, so when he burst out laughing on my calling him, even before I could say hello, I was not too surprised. But when he continued to laugh non stop for longer than logical, I felt something was wrong.

Welcome Truecaller! Here is what happened. For starters, I had not installed Truecaller. But Sumit had. And when I called him, my name showed up as “Motta Swami” in Truecaller. Sumit being the Bengali he is, was in splits because someone had named me Motta Swami on their Truecaller enabled phone, which he (Sumit) translated as Fat Swami. True to myself being the master of pointless pedantry that I am, I begged to correct him – It was not Hindi, it was more likely to be Tamil. In other words, it was Bald Swami, for Motta also means bald in Tamil. Sumit was now laughing in fits with bursts of coughs in between, while it dawned on me that all data about us is no longer ours, but belonged to apps with shady revenue models. Truecaller was my Pietermaritzburg and I had reached my stop.

I then did a mini PhD in Ad-Tech systems, to see how our data was being taken from smartphones, and by whom, and where was all this being used. My mind was reasonably blown on seeing what had come up, and the direction it was taking. I was reminded of Marlinspike Hall… from the outside & from the inside in Red Sea Sharks

Our lives had been taken over completely, and why? Because we press Ok, Ok, Ok, during the permissions screens while installing Android apps. This problem is compounded manifold in Tier-2 India, where users still think EULA is an extinct winged animal from Mohenjodaro, and Permissions are something that come from Gangs of Waseypur. (Did you know that if you manually disable the Microphone permission, a major taxi-app will not work? Did you also know that as part of this permission, the taxi-app can listen to what’s happening in the vicinity of the phone?)

And then back in 2014 and still continuing, is the phenomenal rise of wallets and more. Being called Motta Swami was less damaging to my ego than the prospect of someone in ICICI Bank seeing my declining bank balance (#entrepreneur) and deciding that I was not worth a wealth manager, but more a Hardship Assistant (Siri is Sorry, ‘eh?). Or a software hacker using code written in Tadjikistan or its whereabouts, could drain what little remaining money I had left in my bank. All this led me and my partners to investigate if there were alternate design models using which we could still use the data of our Tier-2 India users, to match them to products and services (#revenueModel) and yet not screw them over by taking their data.

And thus was born our patent pending method of On Device Analytics, a passive data collection framework where we took a bet that Smartphones will become cheaper and cheaper with more and more powerful processing speed every passing year. This meant that we could, in theory do our computing on the phone instead of bringing data to our servers. And therefore, all we had to do was leverage data on the smartphone, make user segments on the smartphone, send key information about products and services to the smartphone, determine a match percentage on the smartphone, and show the advertisement for the product or service if a certain threshold match percentage was crossed, again… yes, on the smartphone

This is quite a leap actually, and not as simple as it sounds. Besides the obvious issues of a super clever engineering framework design, it also did not change our constraints as driven by our Tier-2 India users (back in 2014), viz.

  • App Size < 5MB
  • App should work offline
  • App should not send hajjar data to our servers
  • App should work over a 2G Network
  • App should not consume too much battery

These are really tough engineering constraints to work within, when everything you plan to & want to do is… on the Smartphone. In a way, this was equivalent to a Geocentric Ptolemaic theory being replaced with a Heliocentric Copernican theory, while being forced to still explain the retrograde motion of planets with even more Epicycles and more Equants(#nerdPun)

And this, is what we have built at Mubble. A patent pending passive data collection framework which leverages On Device Analytics in an extremely Privacy Friendly way. No user data is captured and sent to our servers. Even the sample health check probes we run periodically, capture data independent of user identifiable attributes, with an in-built delete-on-expiry mode rather than an archive-on-expiry mode for the data. Thus, even though Mubble takes hajjar permissions from users, our design philosophy ensures that we work on the phone rather than on our servers. And therefore, we are able to meet the demands of our revenue model, viz. Matching users in Tier-2 India to products & services they need.

Remember, somewhere I mentioned about some uninteresting lines in spreadsheets with regard to my business model? Well, the good thing of not taking data to servers is a rather ridiculously low server cost structure as well. Coupled with some funky targeting we do that results in over 90% organic acquisitions (for another article someday), our operating cost breakeven on the prepaid product occurred mid 2017. On track. Despite Jio. In fact Jio has opened up even more spaces for us to play in, and someday I will talk about Suddi Katte, our new Kannada app which is now in Beta.

Again, my friendly neighborhood VC asked me “Dei Ashwin, I understand the power of vernacular when it comes to Tier-2 India, but all the guys are doing all India. Why are you doing only Kannada?”.

Once again, my answer “Because my daughter has a really heavy school bag” was not convincing at all. Get it? No?

Neither did my friendly neighbor VC, but that’s for another article some other day!

Back to the basics: Understanding how access to the internet set the stage for India’s digital transformation

Most scientists believe that before the universe in its current form came into existence with a big, solid bang, there was nothing but super dense, superhot mass of super stuff. The rules of physics, as we know them, quite probably did not apply, and what existed back then can at present only be speculated about with extreme uncertainty. To a casual observer, trying to look at the Indian market beyond the event horizon of the country’s rapid digitisation, accelerated by the entry of Reliance Jio, appears to be a similarly unfathomable situation.

It is hard to imagine a time when people weren’t walking around engrossed in their smartphones, tweeting and sharing and posting stuff on the go, binge-watching entire TV shows on their daily commutes, and shopping for gifts for their partners on the very day of the anniversary. But trust us, dear friends, there existed a time when India was not as digitally-savvy as it is today.

2001-2010: The decade of the (largely dis)connected India

India got its first taste of public internet in August 1995. This phase is what we like to call the dial-up phase. The country’s internet market was then a monopoly owned by the Videsh Sanchar Nigam Limited (VSNL), which provided its customers with the lightning fast connection speed of 56KB/s on its good days. Service disruptions were more common than actually getting connectivity, and heavens forbid if someone decided to make a long call on the phone line which also served as the primary – and only – gateway to the internet. Integrated Services Digital Network (ISDN) was introduced in 1997, but with the infrastructure hardly capable of providing enough bandwidth to its users, things did not improve by much. Users often had to resort to lighting agarbatti and dhoop around their computer screens, praying to whatever digital gods were listening to get some connectivity.

All these hassles came at a high price. During the initial days, the Gateway Internet Access Service (GIAS) was made available to individual users for $160 for 250 hours, while institutional dial-up SLIP/PPP accounts paid $500 for institutional dial-up SLIP/PPP accounts. Leased lines services were priced even higher. Needless to say, the internet – for the average consumer – was hardly worth the trouble, or the money, of getting a dedicated setup installed at their homes. Most internet connections were taken up by business users. Cyber cafés became extremely popular during this phase, with people often waiting in long lines to get their turn to go online.

The internet, at this time, was good for little more than checking emails, randomly browsing through the web, and accessing chat rooms. But Indians are nothing if not tenacious and willing early-adopters. Despite the issues faced with internet connectivity, VSNL signed up 10,000 customers within six months of its launch. By 2001, there were around 7 million internet users across the entire country.

Things started moving in a positive direction in the second half of the decade. With the broadband policy formulated in 2004, the government, in a bid to extend internet connectivity to a larger section of the population, opened up the sector to private players. The increased competitiveness that this enabled played a major role in improving connectivity across the country, as well as in bringing down the internet tariffs. This led to a drastic increase in the adoption of internet connections for personal as well as business purposes.

It was around this time when Indians first started accessing mobile internet on 2G/EDGE networks. Users were using the 2G connectivity provided by networks like Idea and Tata DoCoMo to access Orkut and Facebook, search for information, and download songs and videos on their handheld devices and their computers. With more and more consumers opting for mobile operators that also provided them with decent internet connectivity, other players rushed in to meet this emerging market demand. The auction of the 3G spectrum by the government also made high-speed mobile internet more available and accessible, further consolidating mobile’s position as the primary medium of accessing the internet.

The champions of this evolution were the country’s millennial users, who were finally coming into their own as a consumer segment. What also greatly aided the rise of mobile connectivity was the availability of relatively affordable dual-SIM phones from emerging Chinese companies – established giants such as Nokia and Samsung were late to this party – which allowed users to switch between network operators at will. Indian consumers, long known and hailed for their jugaadability, were tapping into the best of both worlds, using one network for accessing the internet and the other for making calls and sending SMS. By 2010, Gartner estimated that dual-SIM phones accounted for nearly 15% of the overall mobile phone market in India, while the number of mobile subscribers in the country was estimated to be in excess of 617 million by TRAI. This was perhaps the first instance of such a significant shift in consumer behaviour in India being driven by technology – one that set the foundation of the large-scale digital transformation that was to follow.

Interesting, isn’t it, to peel back the layers of India’s staggering metamorphosis into a digital-first nation? We will continue the story of the country’s digitisation and how it has redefined consumer behaviour and market dynamics in the next post. Stay tuned!

The Navy Boys in Data Science!

In my last article, I reflected upon the history of Mubble. On re-reading it, I think I should call these articles as confessions. Like those 80’s movies, minus the spice, plus the social. The Public Confessions of an Entrepreneur, perhaps. Irrespective of what I call it, one of the amazing outcomes of the previous post was phone calls. From friends!

In such times, where social interaction has been reduced to just a click of approbation or a few words typed in, it was truly enjoyable to reconnect with old friends who read my previous article and called me up and we caught up. An entrepreneur lives in a different reality where perspective and destiny play mind games with each other, and therefore such phone calls are akin to plugging back into the matrix.

So in calls that came and the conversations that ensued, it was much more than just fun catching up. In the phone calls, I was soon recounting various specific episodes – much like a kid showing off scars from some wound. As I said earlier, this helps me look back and laugh at my own recollections. One episode that is relevant in today’s Big Data & Machine Learning obsessed world is my own attempts at hiring a Data Scientist!

But some quick background. At Mubble we make money by matching products and services to Tier-2 India using analytics. We began by recommending prepaid packs to our users, based on analyzing their usage and telecom spends. Which required us to interpret large volumes of data. Much of this data was transient unstructured text, in other words the USSD messages sent by operators to their prepaid users (USSD, not SMS). In doing this kind of analysis, we had already developed a sophisticated text-parsing engine, which we described as “Self-Healing”, our tem for what is now Machine-Learning.

Operator USSD messages are an interesting battle ground for 160 characters. In this 160 character Kurushetra, the Pandavas are the finance & customer service teams that want to tell users what the last usage deducted from their prepaid balance is. The Kauravas are a bunch of marketing guys who want to tell users how their destiny is linked to a ringtone. Or a Cashback. Thus, the USSD message undergoes extensive elastic deformation

The first casualty in this Kurukshetra for characters are vowels. “Your Last Call Charge” becomes “ur lst cl chg”, and someone celebrates a victory of 8 characters. An operator who I shall not name, but is shutting down decided that users are extremely concerned about their data usage – so this operator would give information of consumption in bytes. Yes, bytes. And the vengeful fiend who designed this decided to leave the units out. Which resulted in messages like “Data Vol : 3712159”. Clearly this is a person who grew up getting kicks out of reciting the value of Pi to 2000 digits from memory, knowing fully well that in a pre-internet era, there was no way to verify the stream of numbers from him. A third example would make Poirot & Holmes proud. “Your Bal is 28.12.575757 dial karein to get special offers” – a poor computer had to detect that the first period is a decimal point & the second period is a full stop.

And all these formats and syntaxes are extremely dynamic. It’s not as if battle formations remain the same. Thus a bunch of people thought we at Mubble were simply working off a canned set of templates, and it was easy-peasy. Hardly. We had to parse these kind of messages at scale, which were constantly being abused and battered, and thus we developed our Self-Healing system, which is a more desi name for a Machine Learning system that would learn from the changes in message formats & correct itself. I often felt like Ganesha inscribing Veda Vyasa, and suddenly things would become complex with me needing a break. Except I could not stop writing. And there were 176 Veda Vyasas (22 circles & 8 operators).

But this is all just background and past. We created our parser and have been good with it for a while. Now I was making money by matching users to products other than prepaid recharges & pure advertising. We were now selling leads for credit cards, mutual funds, smartphones and more. And now I was trying to group users into segments of relevance for each industry – realtime & at scale. Hence the Data Science imperative. After all, how hard could it be – after potti-kadais (#tamil for kirana store), the most ubiquitous thing today was a Data Scientist, isn’t it?

Welcome the Navy Boys – below are some of the realizations I have got from the guys I met (disclaimer – this is from looking at people in the 1-3 year work ex bracket)

  1. Dei, Coursera is not the same as work experience. And yes, I also did the free Andrew Ng course on Machine Learning. Doesn’t count for sh*t if I am hiring you.
  2. I am not running a coaching center for your moonlighting needs. If you want to work, quit your job and come. Corollary : Kaggle is not a spectator sport
  3. Amazon & Flipkart payment options drop downs do not list “learning”. I am trying to use models to make money. Not to help you learn modeling
  4. I am not a planet for you to achieve the gravitational slingshot effect on your salary. Corollary : A growth of 40% on salary is not the same as asking for 2.5X of what you are presently earning
  5. A baby takes 9 months to make, a decent data science model takes 12 months to make. (So if you tell me you made 3 models last year, I will instead ask you to make the solar system model I make for my daughter’s school)
  6. Corollary to the above : To model data, please understand it first. Remember Descriptive Statistics? Let’s start there. Corollary : This is the starting point, it’s Descriptive Stats, not Destination Stats
  7. Corollary to the above corollary : There is more to modeling than cleaning the data, removing the outliers, removing stopwords, removing spaces and punctuations and so on
  8. Openbook is not the same as Facebook. Exams are not a social media event – The recruitment section titled Homework Assignment is meant to be done alone, and by you just in case you forgot.
  9. There are three kinds of people who wear the Data Science jersey
  • The MBA grad who has found the new fraudzone of opportunity, and knows how to write some code
  • The Comp Sci grad who is starting to realize he cannot make a career in code & data engineering, but is better than most in R & Python
  • The Stats grad who suddenly wants to call himself a Data Scientist, and thereby jack his earnings
  1. The No-Fail-Checks (caveat : this is as effective as those No-Fail-Diet-Plans)
  • Whatever someone says, check if they can convert a continuous data into discrete data using quartiles. Corollary : Check if they can spell discrete
  • And lastly, if someone has gotten by 3 years saying Navy Boys algorithm instead of Naive Bayes, just eat up the Marie Digestive Biscuits on the plate and end the interview in silence (we are a startup, no Dark Fantasy please)

I could go on, but I promised myself that I would only spend only some time on weekends writing, and the more interesting realities begin after hiring, which is for another post some other day.

Lastly, on a more serious note, one would wonder why such stuff as spellings or pronunciations matter. Maybe it doesn’t if you’re in a familiar world, where going to Adiga’s is no different from a Darshini. But in a new world, I believe if you are not careful, you could reach the Americas and think you’ve reached India … like Columbus and his Navy Boys did.

The Mubble Story

I’ve stopped making new year resolutions for a while now. Somewhere along the way, I realized that I have enough task lists as it is. And I did not find it relevant to have yet another task list, sitting on top of all the task lists – One List to Rule Them All. However many folks I know successfully use new year resolutions to undertake pivotal journeys, and this has remained an inspirational mystery to me. Early this year, I met one such friend with whom I was talking about my experiences as an entrepreneur, and the chaotic ride it has been and promises to be.

Strangely such conversations and recollections found me laughing at the times I had found nearly traumatic. Perhaps, we agreed, talking about one’s journeys is a way to deal with the madness that characterizes entrepreneurship. Which is my excuse for what I am beginning in 2018 – an attempt to chronicle the journey at Mubble. This post, however, is more reflective – and it helps to have talent that can convert an unstructured narrative into a visually aesthetic representation.