The best way to build a business is to pick a metric that you would like to optimize, and focus singularly on growing it. Through five years of doing a startup, I’ve seen that the choice of how you define your growth matters as much as your strategy decisions or your operational efforts. And it’s not just me — this approach is followed religiously at YCombinator, where startups usually pick a single metric and attempt to amplify it week-on-week.
I’ve seen the same pattern hold up in my personal life too. In years when I’ve set broad goals like “meet more people” or “read more”, I’ve fallen terribly short of my targets. When I’ve switched to quantitative targets like “meet 1 person each week” or “read 3 books a month”, I’ve not only been forced to think through my goals in greater detail, I’ve also enjoyed far more success in hitting my targets.
Can the same mantra be applied to our public lives?
You’ve framed your problem, prepared your datasets, designed your models and revved up your GPUs. With bated breath, you start training your neural network, hoping to return in a few days to great results.
When you do return though, you find yourself faced with a very different picture. Your network seems to do no better than random selection. Or, if it is a classification model, has curiously learned to classify all entries to a single dominant category. You scratch your head wondering what went wrong, and hit a wall. What’s more, since you’re programming at a higher layer of abstraction, you have no intuitive sense for what’s going on with your matrices and activation functions.
This isn’t a problem faced only by beginners. Empirically, it happens to even the more experienced among us, especially as the complexity of models, the dataset and the core problem increases. So if you find yourself in this situation, don’t fret. To tackle this, we’ve put together a little checklist that might help you find a way out of this hole. This was written specifically in the context of image classification, but the advice is generic enough to apply to all types of networks.
In India, we’re often satisfied with letting go of bits and pieces of our civic rights.
Potholes on roads ? “No problem, we’ll drive around them”! Garbage dump piling up? “We’ll take a walk elsewhere”! Broken footpaths? “Wait, what are footpaths”?
Against the backdrop of the far greater challenges that we face, such issues are viewed as inconveniences and luxuries that we can make do without, rather than a more fundamental loss of rights. That’s why I’m really excited to write about a case of civic action that I was privy to, in which a government body acted surprisingly swiftly and thoroughly.
This is the first in a series of articles in which we use Semantics3 Brand Reportsto learn more about consumer brands and the markets they operate in.
Take a look at the image below —
Personally, until recently, I would’ve reacted with revulsion, since this volume of choice feels like chaos and inelegance to me. That was, until my inner data geek led me to dig deeper into an Excel report generated by my team, and discover a whole new world.
Gathering data about how your FMCG brand is performing can be terribly hard.
If you’ve been through the grind, you’re probably used to working with month old Nielsen data manually gathered from a small group of physical retailers, even as your costs for acquiring these reports pile up. What’s more, as a result of the hassle involved, you’ve probably not given e-commerce much of a thought, especially since it still represents a small percentage of your total sales volume.
We understand this pain … so we’re glad to let you know that we’ve found a solution!
It’s been a couple of years since I made the switch from iOS to Android, and I’ve had little cause for complaint. Play Store, Google Now, Now on Tap, voice search, folder-level controls, system control apps … nearly every Android feature feels preferable to its iOS equivalent (or iOS compromise😛).
But there is one issue that had tempted me to go back to iOS: battery life.
Wander into the kitchen at any home or office in urban India, and you’ll probably find a reverse osmosis (RO) water purification device ready to dispense drinking water. These devices have seen a significant uptake over the last few years, since they offer the promise of health and hygiene at little cost. Unfortunately, these advantages come at the cost of massive water wastage, which I believe can be avoided with a bit of imagination.
Yesterday, I exercised my right to vote for the first time ever! A full 8 years after I gained enfranchisement, I finally cast my first vote during the Tamil Nadu state elections. The experience proved to be quite revealing — here are my thoughts on the matter.
Earlier this week, I watched this TED talk by Hans Rosling. During this talk, Rosling asks three questions:
1) Has the number of deaths in the last century doubled, remained the same or been halved?
2) How many years in school has the average woman spent in school (given that men spend 8 years on an average)?
3) Has the number of people living in extreme poverty doubled, remained the same or been halved?
New Years is a usually time for optimism. Heading into 2016 though, I’m far more wary than I usually am.
A Tale of Four Cities
I call four cities/regions of the world home. Here’s a look at what each one of them endured last year:
Chennai: Hit by the worst floods seen in decades, the city came to a standstill for several days. Some of my nearest and dearest were in mortal danger, communication lines were cut off for days, homes became swimming pools, and my own family began rationing food and water since supply lines were cut off. 300+ people lost their lives and many more lost their life savings.
San Francisco / California: Faced with prolonged drought, California is in a state of emergency. Regular wildfires, rising food prices (projected to rise 6% in 2016), popular foods under threat (almonds!), groundwater drilling frenzies, hundreds of millions of dollars of losses, and more. Farmers have been the most affected to date and it might not be long before urban dwellers are hit where it hurts.