iPhone

For a long time I eschewed iPhones. The form factor didn’t appeal to me. They were too fat for their size. And so I went with a series of Androids that started slowing down insanely after one OS update. And then the iPhone 6 changed that.

This had a remarkably different form factor to its predecessors. It was thin. It was big (not even the Max version). I saw some relatives using it at a family function and knew that it was the right time to try an iPhone. I bought an iPhone 6S, and still continue to use it, and have no problems with it at all.

My wife had bought an iPhone 6S at the same time as me, and that was doing well as well, until a freak accident a couple of months back. That meant she needed a new phone, and having never used an Android (she jumped “directly” from a cheap Nokia to iPhone 4), decided to get an iPhone.

The iPhone 11 arrived on Friday, brought to us by the sister-in-law. It’s big. Much bigger than my 6S. It has many cameras, and very evidently there is a significant amount of software processing that goes into shooting each photo.

And the photos are brilliant. Through the long weekend (Friday was a state holiday) we were at a wedding, and I kept borrowing this phone from my wife to take pictures (even the sister-in-law, who uses its predecessor XR, kept borrowing the 11 to take pictures)  – often enough to annoy the wife.

But I don’t know what it is with this iPhone, but it seems to “look like an Android”. Maybe it’s the case that we got in a hurry (cheapest on on Amazon). Maybe it’s that we haven’t yet removed the film covering the front. Maybe it’s the size. Maybe it’s the Wi-Fi indicator on the top right rather than top left. But for now I’m yet to “accept” it as an iPhone.

As things stand now, I intend to continue with my 6S for as long as it goes. Hopefully this won’t have a freak accident like my wife’s 6S.

PS: I also treated myself to a pair of AirPods. So far they’re decent, but I find them less effective in shutting off outside noise than some random earphones I used earlier. Maybe they aren’t optimised for my ear?

But I love the technology, though! And the product design.

Ganesha Workflow

I have a problem with productivity. It’s because I follow what I call the “Ganesha Workflow”.

Basically there are times when I “get into flow”, and at those times I ideally want to just keep going, working ad infinitum, until I get really tired and lose focus. The problem, however, is that it is not so easy to “get into flow”. And this makes it really hard for me to plan life and schedule my day.

So where does Ganesha come into this? I realise that my workflow is similar to the story of how Ganesha wrote the Mahabharata.

As the story goes, Vyasa was looking for a scribe to write down the Mahabharata, which he knew was going to be a super-long epic. And he came across Ganesha, who agreed to write it all down under one condition – that if Vyasa ever stopped dictating, Ganesha would put his pen down and the rest of the epic would remain unwritten.

So Ganesha Workflow is basically the workflow where as long as you are going, you go strong, but the moment you have an interruption, it is really hard to pick up again. Putting it another way, when you are in Ganesha Workflow, context switches are really expensive.

This means the standard corporate process of drawing up a calendar and earmarking times of day for certain tasks doesn’t really work. One workaround I have made to accommodate my Ganesha Workflow is that I have “meeting days” – days that are filled with meetings and when I don’t do any other work. On other days I actively avoid meetings so that my workflow is not disturbed.

While this works a fair bit, I’m still not satisfied with how well I’m able to organise my work life. For one, having a small child means that the earlier process of hitting “Ganesha mode” at home doesn’t work any more – it’s impossible to prevent context switches on the child’s account. The other thing is that there is a lot more to coordinate with the wife in terms of daily household activities, which means things on the calendar every day. And those will provide an interruption whether I like it or not.

I’m wondering what else I can do to accommodate my “Ganesha working style” into “normal work and family life”. If you have any suggestions, please let me know!

I miss Google Talk

From the time it launched in 2005 until 2012 or so when it was folded into Hangouts, I pretty much lived my life on Google Talk. The standard operating procedure when I came home from work (back then, I used to have jobs) was to open up my computer, open Google Talk and then ping 5-6 people who were online then.

The beauty of Google Talk was that you know who exactly was online at what point in time. So when you sprayed around your “hey how are you”s, you could target them at people who you knew were (or at least appeared to be) available to chat. This meant this had a high hit rate, and you could have productive conversations.

The problem with other chat mechanisms such as WhatsApp is that there is no “online” and “offline” mode. For example, if i’m working and I’m getting stuck and could use a quick conversation, turning my status to “available” on a chat room can then be a magnet for people to ping me. Or I can see which of my friends is online using their statuses, and then ping them.

With WhatsApp, I need to guess who might be available for conversation at this point in time. And that means lots of messages sent out which get responded to at the most inopportune of times (when I’ve got back my flow of work thought, for example).

In yesterday’s business standard I read about this app (whose name I now forget about ) which is trying to restart this kind of chats, signalling “availability” and chat rooms. Hopefully something like that will take off!

Reliance Jio Tariffs Seem Stupid

For the longest time I used a post-paid mobile phone. The hassle of recharging regularly, combined with the attractive rates available on post-paid “corporate” plans, meant that right from the time I graduated business school (in 2006) till I moved to England in 2017, I almost wholly used postpaid phones.

And then in England, I got a prepaid sim upon landing, and then soon discovered that it wasn’t more expensive than a postpaid (and there was no paperwork), and I kept the prepaid. Upon returning to India earlier this year, I’ve continued with a prepaid phone, with a Reliance Jio number. A few months back, I took an annual plan with Jio, paying for a year what I used to pay Airtel in less than two months before I moved out of India.

One of the reasons I don’t really mind having a pre-paid now is that it is far less of a hassle than postpaid. I have a 12 month program, for which I paid once, and until next May I don’t need to worry at all. There is one less bill to be paid each month.

And one thing that makes this “hassle-free” is that I don’t need to check my usage at all, either in terms of voice and data. It is a nicely bundled plan, with zero marginal cost for either calling or using data (the latter up to a (very high) limit). When there is a per-call charge, the balance notification at the end of each call places a mental cost (even if it is a low marginal cost), and you sometimes wonder if you need to call at all, or when you need to recharge.

The “current” zero marginal cost plan by Jio (I had a similar plan from EE in the UK) means that there is no such mental cost, and you can treat your prepaid mobile like you used to a postpaid.

Now things are changing. There are regulatory issues in India – on the “inter connection charge”. When a Jio customer calls an Airtel customer, Jio has to pay Airtel 6 paise per minute for Airtel’s service of completing the call on its network. This was earlier 14 paise a minute, which came down to 6 thanks to Jio’s lobbying, and was supposed to go away entirely in 2020.

When the entire market has settled on a zero marginal cost plan, like it is the case in the UK, inter connection charges don’t really matter. In India, however, there is massive asymmetry. People on older plans from Airtel and Vodafone still pay a lot for their calls, so they don’t mind paying a high interconnection charge, and want to receive a high inter-connection charge.

So over the last couple of months you’ve had massive lobbying, and hilarious exchanges like the debate among the major telcos regarding “missed calls” and how long the phone should ring.

Anyway, it appears that the inter connection charges won’t go away next year as planned. Jio is not happy. And in order to show its spite, it has decided to start charging for calling. A marginal charge of 6 paise a minute is going to be applied on Jio customers calling non-Jio phones.

I don’t see how this is going to be good for the Jio customer (I’m protected since I’d bought a long term plan earlier this year). The mental cost of calling comes back. You need to start worrying about what network the person receiving your call uses. You start getting that balance notification at the end of your call. You might need to recharge before your validity is over, creating more mental cost.

In other words, it seems like a rather dumb move by Jio. While it has clearly been taken to show that the operator is pissed off with the competition and the regulators, it is likely to hurt Ji0’s own business and drive its customers to the competition.

There were several smarter ways to handle this. Basically the problem is that Jio’s costs aren’t coming down as expected, so it needs to charge more. And there are several ways of charging more without imposing a mental cost.

One, the price point itself can be increased. Instead of Rs. 150 a month, it can charge Rs. 160. Second, instead of “unlimited free calls”, they can offer “1000 minutes of free calling per month” or something like that, with a different plan offering 2000 minutes of free calling per month. And so on.

But no. Reliance is more interested in making a statement than serving its own customers. And so it comes up with hare-brained schemes like charging per call “outside the network”. It will be interesting to see how their growth goes like over the next few months.

Dunzo and Urbanclap

I realise that Dunzo and Urbanclap (and many other apps) grew in a particular way. Initially they weren’t sure of the exact problem that they were solving, and instead focussed on a particular “problem class”.

And then over time, based on pattern recognition and segmentation/cluster analysis of the kind of problems that people were using these apps to solve, they started providing more targeted solutions that made better business sense.

Dunzo started off as a “we’ll do anything for you” app. People making fun of the company would talk about a Dunzo executive who would come home, collect your bean bag, get the beans refilled and bring it back to you, and only charge for the beans.

I’m pretty sure that there were many other such weird use cases in which people sort of abused Dunzo in its early days. However, most of the users of the app, I’m guessing, used it for sending packages across town, and to fetch stuff for them from shops and restaurants. And now, four years down the line, Dunzo highlights these specific streamlined use cases in the app, and has figured out a good way of charging for each of them.

It’s similar with Urbanclap. While I didn’t use them in the early days, I used their competitor HouseJoy. I used the app to request for “a plumber”. A plumber duly arrived and did all sorts of odd jobs in our apartment building, some of which were dangerous. And then at the end we paid him in cash, and he told us that “if someone from the app calls, tell them you paid me only 200 rupees” (we had paid him 2000).

Soon, after being a marketplace for all sorts of odd jobs, Urbanclap and its ilk noticed patterns and started specific services. So last week we got someone from Urbanclap to “repair our water heater” (this had a fixed fee on the app). It is another set of such specific services that UrbanClap offers.

I may not have said much new in this post, but it’s basically a crystallisation of some of my thoughts of late – sometimes it’s okay to not have a particularly precise business plan as long as you know what problem class you’re tackling. If you manage to get funded and are willing to burn money, you can learn the best set of problems from the market (within your identified class).

It’s an expensive process for sure, since until you figure this out you’ll be spending a lot of time and money doing random shit, but if you and your investors are willing to bear this kind of expense, it might be worth it.

The worst thing that can happen to you, though, is that after you’ve burnt your company’s money in learning about the market’s precise problem statement, another well-capitalised firm moves faster than you to address this specific market. The question is how well you can put to use your learnings from the early period for later on.

Instagram targeting

Instagram is really good at what I call “one dimensional psychographic targeting”.

Essentially, based on the photos and videos (more likely hashtags) that you see, spend time on, like and comment, the platform figures out some of your interests and targets at you advertisements of products that serve these interests. And instagram manages to combine this with demographic information (where you live, etc.) to target advertisements better at you.

For example, of late I’ve been looking at a lot of weightlifting stuff on Instagram – I follow most of the coaches at my gym, and a few other handles that post fitness stuff. I’ve even posted a video of myself deadlifting.

As a result, Instagram has been following me with advertisements related to fitness, and the combination with demographics means I’m being served stuff I can get in Bangalore. For example, last two days I’ve been seeing ads of my own gym (!!). There are ads for whey proteins and healthy foods of all kinds as well.

This targeting is not perfect – for the last few months, ever since I returned to India, I’ve been bombarded on Instagram with advertisements asking me to emigrate to Canada (I don’t know what makes it think I want to move abroad again given I’ve just moved back home). The seemingly un-targeted mattress advertisements are everywhere. The shirt advertisements as well (though recently I uploaded a picture of my wardrobe on Instagram).

Nevertheless, this is a massive step up from what marketers were able to do a generation ago, where they could at best target based on a demographic. Marketers might have created elaborate psychographic or behavioural profiles of their target audiences, but when it came to advertising, the media available (newspaper, television and outdoors) meant that they had to collapse it into a demographic profile.

Instagram is not perfect, though. To the best of my knowledge, it can only target me on one “psychographic dimension” (“interested in weight lifting”, “interested in coloured chinos”, “likes Bangalore”) along with a multitude of demographic dimensions (I’m sure it’s figured out my gender, age group and maybe even caste, even if it exists in some vector somewhere and no human knows these classifications).

However, when you have created elaborate psychographic profiles, collapsing them into one dimension is still a simplification process. And so you get a reasonable degree of error in targeting. So I’m wondering what can be done that can enable advertisers to target me with more specific products that I might be interested in.

Finally, really how much are the likes of Charles Tyrwhitt, and some mattress brand whose name I don’t recall, willing to pay for their campaigns, given that their untargeted campaigns have beaten the highly targeted campaigns of the fitness guys and coffee companies to reach my eyeballs?

RSVP

I’m reminded of this anecdote from class 11. A girl in my class had invited me to her birthday party. Knowing that there was a clash, I had immediately responded to her saying that I was sorry but I wouldn’t be able to make it. She immediately got offended – that I had told her directly that I wouldn’t come. She would possibly have been less offended had I told her I would come and then not showed up.

A lot of people in India don’t get the concept of how to reply to invitations. Like my old friend, these people think it’s a sort of insult to tell someone that they can’t make it for an event or a function. And so they end up giving false responses or non-responses which doesn’t leave the host any wiser. That leads to massively messed up planning, and possible wastage of food and gifts.

I must say I’ve been guilty of this in the past as well – maybe affected by that 11th standard incident, I have started giving non-committal responses to events that I know I won’t go to. And messed up my hosts’ planning. Having been on the other side multiple times in the last one month, however, I hereby undertake that I will give accurate responses to any invite I get, as far as things are under my control.

Over the last ten days, the wife had kept a massive doll display at home on the occasion of Dasara. We had made an elaborate plan of calling people from different “sides” on different days – in the interest of not mixing groups, which can have a massive negative effect on conversation.

And then some people threatened to destroy these carefully made plans by asking if they could come at a time when they were not invited! Some people were nice enough to tell us that the time when we had invited them was not convenient for them, and requested us right there to give them an alternate slot which we did. Others, however, responded in the affirmative, failed to show up and then wanted to come on a day when we weren’t prepared to receive guests (or worse, on days when were expected other guests from other “sides”).

The other side is also a bit painful here – when people give you an open invitation and tell you to “come any time”. While this gives you greater optionality than a specific slot, this also creates greater pressure on you to accept the invitation. And I’m guilty of responding vaguely to some of these invitations as well. Next time someone gives me an open invite, I will either say no, or try to tell them as soon as possible a specific date and time when I’ll be there.

PS: Of late I’ve started becoming actively (but subconsciously) rude to people who show up at my door unannounced. It throws me off massively. Sometimes my wife wonders why I bothered coming back from England at all!

 

Prodigy

Sometimes I wonder if being a prodigy is more of a curse than a blessing. The sense of having “achieved something” fairly early in life leads to a lowering of objectives, not being excited by anything, and a sort of satisfaction of having “arrived” that reduces motivation to do anything else in life.

A few prodigies keep up the fight and make a successful career for themselves as adults (eg. Sachin Tendulkar). Most fall by the wayside. And find it struggle to come to terms of having become ordinary. And find being an adult incredibly hard, and then get into all sorts of issues.

Five years ago, the Guardian identified the “best young player from each Premier League club“, and they’ve kept at monitoring the progress. Five years later, the results aren’t encouraging.

Out of our 20 players from the English top flight in that 2014-15 season, only three are playing at Premier League clubs now: Marcus Rashford, Dominic Solanke and Hamza Choudhury.

That may not sound very impressive but some others are at Premier League clubs but on loan in the Championship. Six of the 20 are playing second-tier football (five in England and one, Harley Willard, in Iceland) so nearly half are playing at a very high level. On the other hand, two of the 20 – as far as we are aware – are not playing football any more.

While it is natural for parents to push their kids and get them to “achieve something” at a young age, such achievements in most cases don’t result in any lasting advantage as adults. Instead, children who achieve something get labelled as “prodigies” or “gifted” or “talented” and these labels only seek to increase pressures on them as they grow up, rather than helping them build sustainable careers.

OK I might be ranting so I’ll stop here.

Menu Design

Yet another family function yesterday, and we skipped lunch entirely. While it was at a temple and it was well known that lunch would be served rather late (two red flags already), it was more of scheduling issues that we decided to go there for breakfast instead.

Breakfast was pretty good (the wife was pleasantly surprised – she has completely given up on function meals), though I started feeling hungry earlier than I would have wanted to.

In any case, coming back to my original rant on quality of function meals going down, I have a new hypothesis related to an old one. Basically, it’s the increasing bargaining power of the caterers.

Until just about ten years ago, my family eschewed “caterers” and instead employed cooks whose job was to cook with the ingredients provided. The cook, upon being given a menu, would give a list of ingredients and we would procure them. Based on the list, they would bring the appropriate number of cooks, who would be paid on a person hour basis.

It was in the 1990s, I think, along with liberalisation (when you could easily buy groceries in the open market), that cooks moved up the value chain to become caterers. They spared the hosts of the problem of procuring raw materials, and started providing meals, and charging on a per-plate basis. It was just that our family was late to adopt to this practice.

Soon, caterers started providing all-in-one service. The guy who catered for our wedding, for example, also provided the photography services, pooja materials, decoration of the wedding hall and all other sundries. In fact, he would have also been willing to provide for the priests, had we so demanded. “I have set up my business such that the parties getting married don’t need to do anything. They can just turn up and get married”, he had once told us.

And as caterers moved further up the value chain, they became superstars. Moreover, their operations became more process driven which meant that there needed to be standardisation. And standardisation meant less customisation, and they started pushing back.

You would say, “one sweet is enough”, and they would push back with “no, you need two. Our experience suggests that’s the best”. You might ask for some “exotic” item, but they would provide a valid-sounding reason as to why that was not possible.

And so it would go – nowadays if you engage any of these superstar caterers, you have very little control over the menu. You get your choice of sweets and stuff, but in terms of the overall menu, the caterer makes most of the decisions. So even if you are particularly inclined to provide nutritious food to your guests, there is a good chance your caterer will overrule.

Now I make a leap of faith – by hypothesising that this standardisation of the menu is responsible for the declining quality of food and menu choice in most functions and weddings. In other words, now that we are at the Nash equilibrium of caterer control and a certain menu that isn’t nutritious, there isn’t much we can do to improve the quality of food served at functions.

I guess I’ll just stick to eating at home before going to functions, especially when it’s going to be food served on a banana leaf.

Segmentation and machine learning

For best results, use machine learning to do customer segmentation, but then get humans with domain knowledge to validate the segments

There are two common ways in which people do customer segmentation. The “traditional” method is to manually define the axes through which the customers will get segmented, and then simply look through the data to find the characteristics and size of each segment.

Then there is the “data science” way of doing it, which is to ignore all intuition, and simply use some method such as K-means clustering and “do gymnastics” with the data and find the clusters.

A quantitative extreme of this method is to do gymnastics with your data, get segments out of it, and quantitatively “take action” on it without really bothering to figure out what each clusters represent. Loosely speaking, this is how a lot of recommendation systems nowadays work – some algorithm somewhere finds people similar to you based on your behaviour, and recommends to you what they liked.

I usually prefer a sort of middle ground. I like to let the algorithms (k-means easily being my favourite) to come up with the segments based on the data, and then have a bunch of humans look at the segments and make sense of it.

Basically whatever segments are thrown up by the algorithm need to be validated by human intuition. Getting counterintuitive clusters is also not a problem – on several occasions, people I’ve validated the clusters by (usually clients) have used the counterintuitive clusters to discover bugs, gaps in the data  or patterns that they didn’t know of earlier.

Also, in terms of validation of clusters, it is always useful to get people with domain knowledge to validate the clusters. And this also means that whatever clusters you’ve generated you are able to represent them in a human-readable format. The best way of doing that is to use the cluster centres and then represent them somehow in a “physical” manner.

I started writing this post some three days ago and am only getting to finish it now. Unfortunately, in the meantime I’ve forgotten the exact motivation of why I started writing this. If i recall that, I’ll maybe do another post.