Let’s talk about decision making

July 7th, 2017

Slides from a talk I gave about decision making earlier this month at UC Berkeley.

Lets talk about decision making – UC Berkeley from Indus Khaitan

Systems of intelligence–the new software moat

April 24th, 2017

Jerry Chen of Greylock wrote a fantastic piece on moats. He gave a refresher on traditional defensible moats used by technology companies (sustained by scale and network effects) and introduced the Systems of Intelligence (SOI) as a new defensible moat within the enterprise stack. The SOI is sandwiched between the Systems of Engagement (web, chat, mobile) at the top and the System of Record (CRM, HCM, ERP, ITSM) at the bottom.

A system of intelligence aggregates, analyzes, and acts on data passing through it–reduces human intervention, and takes decisions on their collective behalf. Extending Jerry’s thought, a defensible moat is not just AI; it is automated custom actions, stellar visualizations, and an engine to accelerate the system itself.

Let’s imagine what a System of Intelligence would mean in the world of enterprise cloud SaaS products. Today, when you want to buy best-of-breed cloud products you buy a vertical SaaS. But even though most vertical SaaS products supplanted many individual “modules” in the vintage single instance and single vendor on premise software stack, they remain highly specific and fragmented.

For example, there’s a SaaS product for sending mass marketing e-mails to consumers, sending out post-purchase surveys, or engaging users on a site. There are a half-a-dozen products one has to use to analyze a customer workflow (or touchpoint) across channels. At each stage of the workflow, data is generated and then stitched to get a state transition value.

In enterprise SaaS, the System of Intelligence is an analytics mesh, which coalesces many individual data-points and provides actionable insights. This gives birth to a new kind of defensible moat because the vertical SaaS is now serving as data sensors that are tightly coupled with the enterprise user on every device.

BI and analytics products such as Tableau, Domo, Qlik and other tools are precursor to tomorrow’s Systems of Intelligence. Today, they do a good job in providing a unified data aggregation and analytics layer on top of customer, employee, and prospect workflows.

We will continue to see disintermediation of general purpose analytics products, and domain-specific intelligence tools will win over generic aggregation. AI will help stitch the data together and create automatic actions, which is when intelligence will drive system acceleration.

SaaS: Where are you in this 2×2?

March 15th, 2017

Some SaaS ventures lead to category leadership while some lead to imaginary frozen quadrants. Here’s a little 2X2 to assess where you are in your journey to SaaS nirvana. When amazing products are sold in amazing ways, it produces the almost mystical flywheel effect.

2x2 Product vs Sales

Let’s dissect this.

Red: Weak Product and a Weak/Average Sales Team

This is a highly incremental quadrant where a single provider may be serving the exact needs of a handful of customers. It’s an equilibrium that doesn’t last too long. I’ll leave this quadrant at that.

This is a highly incremental quadrant where a single provider may be serving the exact needs of a handful of customers. It’s an equilibrium that doesn’t last too long. I’ll leave this quadrant at that.

Blue: Weak Product and a Strong Sales Team

When people say “that company is sales driven” this is what they are referring to. Founders of companies in this quadrant have a knack of story-telling and projecting a product market fit before a product is actually ready. What happens next is catastrophic. Sales drives the company’s culture, narrative, and product building. Both product and engineering go into a wild-balancing act of fixing the problems while trying to add features in a near random fashion.

It is unsustainable. It bloats customer service, support, and pre-sales. Lack of a strong product causes politicking, confusion, and populism in every department, which leads to relationship-driven rather than value-proposition driven outcomes. Unless a startup iterates on product rapidly or brings in a disciplined and creative leader, there’s a significant risk of revenues plateauing at $5m-$10m mark.

So why is it blue? Because it is fairly cushioned for a while though sales > everything is bad karma.

Yellow: Strong Product and a Weak/Average Sales Team

This quadrant probably causes hackers amongst us the most heartburn. A lot of strong products start with nobody focused on sales. They continue to write amazing code, design amazing screens, and setup amazing data pipelines, but they just don’t know how to position, craft a story people remember, distinguish themselves from 99 other guys who may have had the same idea. Many product founders suck at sales and often hire the first person who blinks.

Even for successful startups, this can be a transient stage, but successful founders realize their mistakes and then quickly hire a sales leader and move to the next quadrant. The good news if you’re yellow is that just like in real life, you can cross the traffic light before too much damage happens.

Green: Strong Product and Strong Sales Team

This is jazz improvisation zone. You can have a strong product and sales culture. It all starts with respect for both and it certainly involves finding the right talent that can craft what really works uniquely for you.

That’s why very few founders get there. A scalable sales model is crucial. A product alone can take you so far. For every Dheeraj Pandey ringing the IPO bell, there’s a Sudheesh Nair driving the quota home. For every Jyoti Bansal getting acquired at $3.7bn, there’s a Dali Rajic digging into sales capacity, and for every Jason Lemkin, there’s a Brendon Cassidy. When phenomenal founders and product builders pair up with their sales counterparts, to accomplish a sight to behold – a startup on a flywheel across the sky.

In each of these cases, the sales counterparts were able to hit their targets, because of a product which was able to either create demand or was superior to incumbents. If there was a product market fit, based on the narrative, the product scaled to bring in a perpetual stream of renewals and sources of new revenue.

Hopping in the 2×2

I hope you’ve found your color by this point. So how do we transition from a shitty part of the quadrant to an awesome one?

If you are Blue or Yellow, scale to Green quick. Here are some things that increase your chances in a hop

  • Listen to early feedback from customers and employees and suppliers. Setup key feedback loops
  • Iterate the product every week, every day, every hour. Continuous Beta. A living element
  • Once you cross $10 million, press the gas pedal. Go. Go. Go

Good luck and let me know if you think of additional colors.


Thanks to Leena for flywheeling this post!

A better way to do 1-on-1 meetings with your team

December 27th, 2016

It’s been almost 4 months at SirionLabs. I started as a Chief Marketing Officer, to drive a data-driven marketing culture. Within a month, I started working on activities which were ancillary to what a marketer is supposed to do—soon my R&R got expanded (more on this, in future).

My immediate challenge was how do I get acquainted myself to a large cross-functional team. I don’t like sending e-mails to a large distribution saying, “Meet your new boss”. Being on the other side of the table, I always hated those e-mails coming from someone who was many hops away.

A better idea was I thought, why not simply organize 1-on-1 meetings.

The big question was how do I get them excited to talk, get them engaged with me, still have a formal communication, but break-the-ice right away.

In the past, I had mixed success with 1-on-1s with large teams–they were a hit or miss. When my former bosses did 1-on-1s with me I always thought, “Is he even listening to me?”, “Does he even know what I’m working on?”.

I did not want my extended team think about me the same way. Plus, I wanted to deep dive into their inner-workings, what they do, what excites them, what demotivates them, etc. etc. The theory was that if the discussion started within the context of their day-to-day work, the answers would be easily found. I hypothesized.

A better way to do 1-on-1 is to start with a demo, get the team member present what they are working on, present an idea, and then do the formal 1-on-1.

Here’s a screen grab of the agenda of many such 1-on-1s I did last week, this week and many more to go the next week.


I’m going to continue to use this format in future. I found it engaging and made both the parties to open up quickly and start at a common ground.

PS> It’s a different topic altogether that the 1-on-1s are not an easy feat by itself when you have dozens of people who are 13 and half timezones away!

The down day hack–Reminiscing your last day e-mail from an earlier job

October 13th, 2016

The startup experience is a collection of up and down moments–whether you are a founder or not. The moment you commit yourself to a cause, it’s no longer a compartment of your brain which can be turned off. A few days ago, I had a down day which lasted 24-ish hours, but luckily I chanced upon something which brought me back from gloom.

While looking for a random message, I found an e-mail which I sent while leaving Oracle, earlier this year in March. Reading that thread and re-collecting the love people gave after I sent that note warmed my heart. Thanks to the lovely messages, I was back to drinking virgin Pina Colada!

I never thought I’d ever publish this e-mail outside, but here it is, slightly redacted.


Messaging Apps And Revenge Of The Computer Science Nerd

February 7th, 2016

Guest post on TechCrunch.

The Portability Angle Of Tech Immigration

February 1st, 2015

Wrote something after a looong time. Guest post on immigration, in TechCrunch.

Bootstrapping versus raising venture money

November 5th, 2013

Wrote a guest post on NextBigWhat opining that bootstrapping is just a foundation stone for building a large business.

How I hired in India

September 28th, 2013

Bangalore Team watching cricket teams sparIn 2010, Bitzer Mobile’s Bangalore R&D centre came to life in a living room in Marathahalli. Ravi (first team member in India) and I wrote much of the first version of Bitzer’s governance console on our respective couches and coffee-tables at Cha Bar in Leela Palace, before we moved into our own office.

In a year, we got a team of kick-ass engineers in place. They built critical parts of the product. However, the first few hires were not easy and was rife with people who kicked tires without joining us. Apart from engineers, QA, office boy, accountant, office manager positions, I also on-boarded an amazing head of engineering which took 12 months!

In 18 months, reviewed close to 2,000 resumes, conducted roughly 400 phone calls and code-jam sessions, followed by approximately 50 on-site interviews, leading to the 15 people we have in Bangalore. I am lucky to work with them.

Here is some gyaan about hiring your first 15 key team members in India:

Always be hiring

Hiring is the other full-time job of a founder, apart from writing code, doing sales and marketing or speaking at events. Most startups start the hiring process, do a few critical hires and then get bureaucratic. I have seen startup founders who start it well and then delegate it to their direct reports or to “managers”. There is a single rule for always-on hiring process–The founders and the first five hires should hire the next 500 and be part of every freaking hiring decision making process.

Use the tools

We started using Recruiterbox from the get-go. I can brag that we were one of the early adopters of this application. We got close to 3,000 resumes in 18 months. Every resume reviewed, annotated and passed around within the application itself.  In the first 12 months, every candidate who submitted a resume and was not a fit got a response. However, we stopped the response due to increased spam and sheer volume, but made a point to respond to anybody with whom we had a call.

Most hiring managers fail at the tools. They continue to rely on e-mail for managing the resume queue. Moreover, there is no context if the resume moves around and gets passed as a solitary reaper. The interviewer has to start all over again! Wastes time and energy.

Go deep beyond the interview

Go beyond the technical “chops”. Discover the person you are hiring — pose him left-handed questions, personally take the candidate out for lunch/dinner, let your team take him out for lunch/dinner, have casual conversations, spend time with the candidate away from office. Get as much facetime as possible before pulling the trigger.

You are a startup, find a right match

Just like there are rules of marriage (some archaic), there are rules for early startup hires. It’s difficult if not impossible to get an engineer from top R&D organizations in India to come and join your founding team.  We were lucky to have applicants from these companies but a match could not be made. These were some amazing engineers, and to my surprise they were ready to take a massive pay cut. However, based on my gut feeling, as an early stage startup, we were not ready for them.

Don’t boss the interview feedback

We lost couple of solid candidates, which the team felt otherwise, whereas I wanted them on-board. The rule was simple, if there is one negative and the argument cannot be one in favour of candidate, then it’s a NO. In one situation, I rescinded a verbal offer, because we jumped the gun as the fifth interviewer went out for a walk and we did not take his feedback into consideration before making a verbal offer. The candidate was shocked that we changed our mind in less than 60 minutes. It was a tough conversation. This one simple thing gave me a solid trust with the team. I lost a few good candidates (well, I still think they were good!) but that built a foundation for the hiring process.

Hiring senior talent–Not everybody can build Adam’s bridge

Head of engineering is one of the toughest position I filled. I had an odd criteria in mind viz. 10 years of experience, hands-on, startup-guy, big company guy, product experience. Combine this that we are an unsexy reasonably funded startup, and that too in India, this made the job of finding the right person even tough. An engineering manager who is responsible for people’s career has to be humble, full of empathy and a kick-ass developer himself. Three things had to be ensured around his happiness and outlook (a) Not a boss (b) Keeps his family happy (c) Still hungry. It took me close to 12 months to fill this position. The CEO still chides me whether I was really hiring for this position or my progress reports were “cooked-up”. A wrong engineering manager is like arsenic, there are no visible bruises, but gnaws the company from inside, slowly. The engineering leader could make or break your company without anybody knowing about it. Compared to a developer, there are no straight-forward objective tests for a engineering manager.

There is one new rule I followed–Evaluated if the person continues to amaze after the third meeting (maybe a dinner at candidate’s home), fifth meeting (maybe a lunch with my family). Candidates who moved beyond my filter were then interviewed by pretty much everybody whom he is going to work with. Took time–But, I think, I got the best guy on board.

Count the intent (or rather lack of it)

In one situation, with the offer, I had to send our incorporation certificate, proving that we are a real company! In another, a candidate wanted to talk to the CEO, before doing a technical round. None of them joined. Whereas, motivated candidates would pick-up the phone, follow-up, show up on time. The biggest data point I have–from interview to offer, if the candidate has been punctual in taking calls, coming for onsite interviews, turning back the assignments, etc., it’s likely that he would accept the offer. Whereas, candidates who continue to move things around are just plain kicking the tires with you!

There are a lot of tactical things which cannot be disclosed publicly on this blog–without context it would misleading. However, there is only one goal–hire the people who like working with each other and can solve any problem given the right motivation and tools.

Happy hiring!

9 reasons why VPN is not the right tool for remote access from Mobile

September 25th, 2013

Wrote a quick (but long!) post at Bitzer Mobile’s blog on VPN and why it ain’t not the right tool for remote access from mobile device and tablet. Read it here.