Deep Dive into Virality & Network Effect

ProductCamp SV Logo

I recently delivered a talk “Deep Dive into Virality and Network Effect” at ProductCamp Silicon Valley held at Santa Clara University. My presentation is now available on Slideshare (enabled for download too). I had Excel models too but that would be too complicated for someone without a discussion.

It was the 8th year for ProductCamp Silicon Valley, which is biggest of all the ProductCamps anywhere in the world. It was fantastic to see dozens of people at the sight as early as 8AM on Saturday.

Session Voting Rank

Tom Gilheany introduced the event to the audience in a very interesting way including an exercise to demonstrate how regular practice leads to a habit. Out of 27 on-the-fly suggested topics, my session managed to gather 19 votes to stand 9th and get a place in the second round of sessions. The entire list of sessions can be seen here. The topics were extremely interesting, diverse, and delivered by extremely well rounded Product Management professionals in the valley.

The top voted topics in the first session were

  1. How Product Managers Decide What to Build (7 Ways to Prioritize)
  2. What Directors and VPs of Product Management do?
  3. WTF is HDFS? – Big Data for PMs
  4. Beyond the Tech Cube: the 7 steps to building a thriving marketing consulting practice
  5. Our Product Managers Are Not Strategic & Other Complaints
  6. The Lean Product Playbook

As I looked at the choice of top topics closely, it revealed some interesting facts about the audience

  1. The confusion of PMs around what to build can only be appreciated by PMs. Other professionals wouldn’t probably relate that well. Indicates that hardcore Product Managers were present at the event.
  2. The second most voted event attracted double the capacity of the room. It was obvious that most participants were ambitious mid-level PM folks looking for career promotion.
  3. Big Data is the current Mega trend in valley and the pace at which the space is evolving and the startups are getting funded, PMs definitely need to upgrade themselves about it. The seen on HDFS was definitely an important one and voted rightly so.
  4. 4th and 5th sessions confirmed my belief that the large audience was mid level PM community. This is the obvious stage when PMs get frustrated with the status quo at their company and start evaluating options for self practice, freedom, and larger opportunity portfolio.

It was delight to see that some participants pulled bigger audience with even ad-hoc sessions in the open area. The feel of event was pretty addictive and reminded me of Barcamps. There were quite a few Job postings on the wall. Silicon Valley is definitely a great Product ecosystem and this event reflects it so well.

UBER – The best designed product of mobile era

UBER LogoIt is exciting to see products that change people’s behavior at the mass scale. Products like facebook, instagram, and whatsapp are undoubtedly some of the change leaders but people take time to appreciate these products. However, there are products that take our breath away the first time we use them. UBER is one such product that is inspirational for its design, purpose, and nearly flawless implementation. I am deeply influenced by UBER and end up discussing it invariably in most product related discussions with fellow friends. Before I dive into describing why I truly admire about UBER, let me share some experiences that helped me build the perspective.

First Experience

In June 2014, I used UBER in Gurgaon (India) for the first time, primarily because it extended an opportunity to travel in high-end cars like BMW 7 series, which I could not afford otherwise. It was a premium experience; I loved to talk about it within my network, and observed how I opened up from being cost conscious to experience conscious customer. Multiple WhatsApp groups were abuzz with discussions every time someone used the service. The community around me, which I believed would find it expensive, proved me wrong. People were open to shell out a few extra bucks for the UBER experience. The zero barrier trial with the first-two free rides and word-of-mouth virality were driving people crazy. Entire experience from onboarding to payment was redefined and was incredibly seamless.

Second Experience

In August 2014, I relocated to California. At San Francisco Airport, I wanted to use UBER to drive down to Sunnyvale. Unfortunately, I did not have a US number activated, so I was forced to take a traditional taxi. I wondered why the hell do they not have a web interface or let alone a counter at SFO airport; it could drive so much business after all. I paid $127 to the taxi driver and later discovered that I overpaid at least $40 for not using UBER. This experience drove me crazy to try UBER even more. The first thing I wanted after having a local number activated was to use UBER for my routine travel between home and NASA campus.

Regular Experiences

Until I bought my car, I used UBER for over 30 trips between Sunnyvale and NASA Research center and conversed with almost every driver out of curiosity to discover how do they feel post UBER era. Taxi drivers ranged from 25-year-old professionals upto 60 years old retired men and women. What I discovered was truly transformational. Particularly, a recent immigrant from east coast, who joined a real-estate agency, used UBER to familiarize herself with the new geography in addition to making a few extra bucks. An Asian retired person admired the sophisticated and yet easy-to-use technology interface that empowered him to make upto $2500 a month easily, a sum that helped him fulfill his family needs in the growing cost structures of bay area. Every driver had a unique and exciting story to share. I could closely feel the inclusive social impact that UBER was driving.

Macro View

At the macro level, it is a clean unification of three heterogeneous infrastructures

  1. Physical Layer – Drivers and Personal cars
  2. Data Layer – Advanced Geographical Information System
  3. Application Layer – Intelligent dispatching software system

The resulting infrastructure is the classic two-sided marketplace of car drivers (the producers) and the end consumers with a trustworthy distribution network that is really hard to replicate.

Product Design

From the design standpoint, I think following decisions were crucial to deliver the optimal and consistent experience

  1. Native smartphone application not HTML5 web application. It is easy to fall trap for the opportunity loss due to SEO driven traffic, but web could actually not deliver the connected and engaging experience that native application can. It is a hard choice but an immaculate decision to deliver superlative experience.
  2. Leverage existing physical infrastructure in developed countries, but create own infrastructure (UBER leased a fleet in India) in developing countries. This is a very capital-intensive decision but went long way in ensuring the consistency in experience.
  3. Most applications ignore the importance of last-mile information delivery. UBER does a fantastic job of providing visibility into the last-mile dispatching of car in real-time. This is the disruptive UX and certainly requires heavy engineering investment. The only place you witness such an experience is in the international flights.
  4. While every transactional business was trying to bring the payments experience down to a single click, it was probably the first time I experienced a zero click payment interface. Having worked closely with payments industry, I can vouch that UBER’s payment experience is absolutely stunning. I believe that the absence of the explicit act-of-payment will become the de-facto standard going forward.

Growth View

UBER’s success depends not only upon how soon can it attract both sides on its platform, the drivers and the travelers but also on the levels of participation of both the sides. A higher participation from drivers is useful only if it results in higher availability, and consequently, lower waiting time for travelers. Similarly, a higher participation from passengers is useful to drivers only when it means lower down time and, potentially, the ability to charge higher prices, thanks to Uber’s much-maligned surge pricing. (Surge pricing increases rates as demand overtakes supply)

Demand Cycle

Each driver is unique and different and therefore is the weakest link in spoiling the entire experience for the customers. Therefore, driver’s onboarding is crucial to enforcing compliance, uniformity, and customer delight. A few challenges I can think of are

  1. Background verification of each driver to establish trustworthiness
  2. Training of each driver to deliver the same level of experience
  3. Compliance enforcement at each node for every transaction
  4. Quick turnaround for onboarding
  5. Sophisticated commission computation and disbursal system
  6. Transaction analysis to detect patterns for unfair practices
  7. Fraud detection and resolution mechanisms

I have not signed up as a driver but understand that the driver’s onboarding process is entirely automated, seamless, and quick. Considering the complexity and current scale, I think the driver’s onboarding process is designed remarkably, which in my opinion is very crucial for leveraging the network effect.

From the customer onboarding standpoint it is important to lower down the barrier for trial and hence adoption. Assuming that the product is well designed and delivers aspirational experience, customers actually end up spreading the word-of-mouth leading to huge virality. Four important metrics that are crucial to understand the user acquisition story are

  1. Virality coefficient
  2. Virality Decay
  3. Virality Cycle Time

I wouldn’t be surprised to know if the virality coefficient ranges from 0.5 to 0.7. Two schemes in my opinion would lead to such a high value

  1. First ride free – This lowers the barriers for new users to try the service and talk about it.
  2. Referral Bonus to both the parties – Both the referrer and referee receive $20 bonus credits on the first ride ensuring that each person invites at least one person.

Considering the virality decay is 50% overall, The Virality Coefficient above 0.5 would lead to a hockey stick curve growth, which looks pretty likely. My wild guess is that the decay would be more than 80% within 2-3 days for infrequent travelers.

The virality cycle time would be really low because the probability of newly acquired customer experiencing the service immediately would be low. I would imagine that the average cycle time could range between 5-10 days.

Renting a taxi is transactional in nature. As the need is fulfilled, users tend to move out of engagement loop. But the length of the transaction in this particular case is large, which works in favor as the customer spends more time thinking about the service. Traditional taxi renting experience clearly needed improvement both on customer onboarding and offloading. UBER’s entire transaction experience flow is single purpose and designed to engage customers extensively.

Customer Focus

UBER somehow makes you believe that it is in a constant pursuit of driving the travel costs down to zero eventually some day. The launch of UBERX signaled the effort and the launch of UBERPOOL was even more convincing. From the user retention standpoint, it gained the trust of busy professionals who demand great user experience along with a dependable service.

On the New Year eve, I received an email around how prices will be shot up around the peak load time. This proactive pursuit of informing the users goes a long way in winning the trust further.



I’d recommend two features that I believe would go a long way in retaining the customers

  1. Onboarding duration accuracy – As soon as the nearest taxi driver accepts the request, the application flashes the approximate duration (t1) within which the taxi is expected to reach the customer. But that duration is not reliable as the net duration (t2) taken by the taxi driver could be X times higher than it first flashed on the screen. Because of this I don’t trust UBER in critical situations. Let’s say the application also flashes a static timer that computes the net duration to onboard the customer, and the driver is incentivized for reaching earlier than approximate expected time, it will significantly enhance overall onboarding duration accuracy. For the delay of more than 2 minutes over expected time, the customer should be compensated suitably.
  2. People Discovery – It would be great to get the application social. If I can discover who else travels the same route within the half-mile of both the source and the destination, I may be willing to work around my travel time to leverage the UBER POOL benefits. This service can lock the users from going to any other similar service provider and shall hugely contribute to retention of existing customers.

Jerry & Filo Framework

I belong the generation that grew watching Yahoo! grow big, used Y! messenger endlessly, and carried a dream to create something as simple and as big as Yahoo. During my Undergrad school in 2004, I came across a small video advertisement that portrayed Both Jerry Yang and David Filo in a casual setting.

This ad was the first of its kind I had ever seen. I connected with them almost instantly. Their chemistry was so smooth that it was deeply imprinted on my mind. Whenever I looked at someone to work or start a company with, I always imagined whether we could behave like the way Jerry and Filo did in the ad above. I called it as Jerry & Filo framework for myself. The framework was simple, but very effective. It helped me choose some amazing people I have worked with so far. It works even today, all the time, and I carry huge respect to both of them.

Following years Yahoo had its fallout and Jerry resigned as CEO in 2012. It was emotionally breaking for an insignificant admirer that I am. Last year through a friend, who now reports to Jerry, I discovered that they are based in Palo Alto. I came from India, visited their office, but returned back without meeting Jerry. He was probably sitting in an adjacent room. I was so close to one of my role models yet stayed away. Of course I’m sure about meeting him some day soon but with the right opportunity.

Today, Forbes published an insider story Finding Alibaba: How Jerry Yang Made The Most Lucrative Bet In Silicon Valley History. The story outlines the professional journey of Jerry, the epic role of Mayasoshi Son (Softbank), and the bonding between Jerry and jack Ma. An unexpected piece of information that really touched me

… But guess who’s getting a seat on Alibaba’s board post-IPO: nobody affiliated with Yahoo except Yang.

It is great to see genuine credits to Jerry Yang and hear about him in the valley again. I highly appreciate Jack Ma for this decision. The article has a video, The return of Jerry Yang, embedded at the bottom that shows he is back in action. I thoroughly enjoyed the story and seeing him in action again. It gives me hope, energy, and focus again.

Happy to have relocated to valley this year, don’t know why but it feels like coming home!

Rocket Execution

“It is not the idea, but the execution that matters”

For a long long time, the above principle has been the key to huge success, but for innovative entrepreneurs the above principle is seeing a generational shift.

I am reading a book called ‘The Dhandho Investor‘, written by Mohnish Pabrai - an iconic figure in the investment world, and the Chapter 14 in his book is titled, “Invest in Copycats rather than the innovators“. Simply because copycats are likely to show multiple degrees stronger financial returns to the investors than the innovators.

Despite active participation of people globally in consuming technology innovation at a higher rate than the previous year, stock markets continue to behave absolutely opposite. Such a contrast it is!

Having seen two pioneering software product startup cycles, and having witnessed both web and mobile evolutions very closely over last decade, I have come to believe that it is not the execution that wins the innovators anymore.

It is the “Rocket Execution” that keeps the innovator alive in today’s times.

Rocket Execution

Rocket Execution

I often state the term “Rocket Execution” in my conversations, however, people end up misunderstanding/ misinterpreting the term every single time. In a quick succession since yesterday, I read news about two young companies, which I truly admire.

Yesterday, I read about the next round of funding news of InMobi. InMobi team started early 2007 and pivoted towards current idea after a failed attempt. Let’s look at their statistics

165 countries (presence),
1000+ employees
$400 million revenue
750 million monthly active users
2.6 billion app downloads
126 billion monthly ad impressions (which is roughly 170 million impressions per second).

The company was valued at $1b in 2011 (within 4 years of existence). It is aiming to double the valuation in 2014 (within 3 years). Hopefully the next billion would be added in 2016 (within next 2 years). This is a classic example of ‘Rocket Execution’ because if you do not grow at that rate, technology innovation elsewhere will outpace you.

Today, I read about the $19 million funding news of A company that started around June 2012. Let’s look at their statistics

25 cities covered in India
12 co-founders (Whoa!)
900+ Employees
2000 houses listed per day
50,000 houses mapped in Bangalore
80,000 houses mapped in Mumbai

If you notice the quality of listings on their website, you will understand the value of listing 2000 houses per day. At this rate, the company will list 700K houses annually, which is not at all a bad number. But just 2 years in operation, the company is aiming to list 10,000 houses per day, 5 times the rate at which it lists today i.e. 3.5 million listings annually. This is another fantastic example of rocket execution. Scale before another technology innovation trumps your innovation.

If that was not enough, an interesting tweet from Navalkant validates the emergence of Rocket Execution little scientifically. It states

1999 – $5M to launch a product, 30M serious computer users.
2014 – $5K to launch a product, 3B serious phone users.
Leverage per $ is up 100,000x

So don’t get excited if you got a million users within a year, until the next million happened in 6 months and the next million in 3 months and so on. Rocket execution is the only way to leverage the never-existed-before power of the connected world.

Disclaimer – The published data is gathered through public sources and it does not reflect actual number that may be relevant now.

How twenty-somethings can change India!

Youth is the future!

That’s one statement you would hear from every next person anywhere in the world and wonder how could that be true for India since there are so many problems in our country and our youth is absolutely disguised. However, there are certain stories that put the above statement in the right perspective.

I recently learnt about a very young company, led by a team with an average age in 20s. Below presentation starts with an eye-catching statement, “Why the hell are you spending your twenties at …?”

I congratulate the team on their successful entrepreneurial journey. They are definitely weaving an inspirational story for twenty, thirty, forty … somethings in India. Hats-off!

4 Steps framework to cultivate Product Thinking

Product Thinking? You probably think that the answer is obvious but like the term “strategy”, the term “product thinking” is elastic. As an entrepreneur I often use this term and sub-consciously evaluate product thinking quotient of people I meet, only to find that most do not understand or still care.

What exactly is Product Thinking?

A product acknowledges that teams are really working to ensure superior experience. It encourages the delivery of small features, frequent releases, continuous flow and releases that cover more than one “project”.

Every product provides some sort of a service to its user. Facebook is a product but it provides communication, photo sharing as service. Mobile handset is a product but it provides ubiquitous communication as service.

Services have always been differentiated by its quality and customer experience e.g. Dell and Apple both are examples of fantastic service quality and customer experience respectively.

In order to ensure highest standard of service quality and customer experience, the organization structure, systems and processes need to be efficient, smooth and must have a feedback loop to feed improvisation, leading to a cheaper, faster and better output. The fact that the service design is now a never ending process brings it closer to product design process. The ability to apply product design framework is precisely the product thinking.

Confused? Read on.

How to cultivate Product Thinking?

More often, the cue comes from user experience by breaking it into as many distinct logical steps. Watch out for the 4 steps framework below to design better experience

  1. Optimize the time and motion in one or multiple steps
  2. Embrace obvious ergonomic convenience
  3. Eliminate dependencies on external agents
  4. Eliminate maximum possible manual effort

Let’s take the example of commonly used “hand soap” used for washing hands and try to break down its user experience in distinct logical steps with respective exit points.

S.No. Action Exit


Open the tap Water supply begins


Pick the soap Wet hands/soap


Rub the soap Soap layer between palms


Leave the soap n/a


Rub the palms Lather creation


Hands under water Lather washed


Close the tap Water supply seized


  • Soap has an ergonomic issue with multiple people using the same bar.
  • The external agent here is water.
  • User engages with the soap in steps 2, 3 and 4 and hence require manual effort.
  • Exit column identifies necessary outputs through sub-processes and it is this column that matters more than the Action column.

Now look at the picture below


Think about it, isn’t it interesting how someone would have thought of replacing soap bar with the liquid soap in dispenser.

As a product thinker, we can safely begin with user experience and one could take varied approaches. Keeping the above 4 steps framework in mind, I came up with following questions

  1. Can the three steps (2, 3, 4) be reduced to one or two steps to optimize the time and motion?
  2. Can we remove dependency on the external agent (water) in 2nd step completely?
  3. Step 3rd requires specialized manual effort, can we completely eliminate it? If yes, then step 4 could possibly be eliminated, being an extension to step 3.
  4. Water supply begins in 1st step and seizes only at the end of process, hinting at the wastage of a potential resource.

While this is the way to think, actual insight can come through any of the above questions. Whosoever came up with the idea of liquid soap using dispenser achieved few important things

  1. It eliminates ergonomic problem associated with the soap bar, providing key mental convenience.
  2. External agent is eliminated in the early part and gets introduced later, ensuring efficient use of a resource.
  3. Eliminates step 3 completely, improving the time and motion.
  4. Overall manual effort reduced.

You can apply similar steps on existing systems and services around you and it is guaranteed that over a period of time you’ll start to conceive not only innovative but scalable product ideas.

An aspect that I have purposefully not covered is the cost-effectiveness of solution. This has got a lot to do with its market size and perception and the subject requires detailed discussion. I will probably write a separate post on it soon.

How to get Psychometric tests to help you?

Recently, I was interviewing people and it was averaging 3 Skype calls and 2 in person meetings every day. Quite crazy it was! Soon I drifted in a mode, where one would look at macro behaviors to conclude people’s personality but I can confidently say that it is dangerous! It is humanly impossible to make consistent conclusions in such a high pressure circumstance and your company runs the risk of limited understanding and personal biases of one person.

I immediately looked for psychometric solutions to save us from this menace and also provide scientific ways to understand people’s personality. Soon, an award-winning and well-known solution provider got in touch and their sales guy did a brilliant job of convincing me why they are best suited to us. Soon I got a demo account and as a natural step forward, I wanted to validate accuracy of results.

I attempted the test myself, the “Situation Judgment Test” score was quite impressive, so I din’t bother much but “Personality Test” result was not that straight forward. Since it was me, I could drive logical conclusions, which made the challenge of validating the accuracy of results even bigger.

  • Can the score be taken at its face value to make hiring decisions?
  • If not, how to drive logical conclusions for new people?

While I immediately got a senior colleague to take the test for the help but the confusion persisted; we had to still apply mind to make logical conclusions. What do you do in such a case?

And then I decided to take the test again, to check the degree of variation in my scores itself. The outcome is really interesting and helped me a lot in driving better conclusions for new hires. Here is what I found

Big-5 Model for Personality Assessment

E = Extraversion
C = Conscientiousness
ES = Emotional Stability
O = Openness
A = Agreeableness

Both the test show how I’m low on Extraversion (i.e. reclusive, shy, silent, introvert etc.) in both the cases. It actually depends on what kind of group I’m in but on a broad level I found it okay. I was also okay with three more trait scores such as ES, O and A in the first result but conscientiousness score improved in the second result. Below is the comparative analysis and conclusions that could be driven

  1. Top (O) and Bottom (E) traits came out constant sharply (although actual rating has some deviation). I concluded that the algorithm to compute both O & E score has some stability.
  2. Two Traits (ES and A) jumped one level each. Deviation in A is not high although it jumped a level being a boundary case. Therefore, approach to compute A’s score still seems fine.
  3. Deviation in ES score is significant. It definitely needs attention due to significant deviation.
  4. Conscientiousness is the only trait which jumped two levels with high deviation in score. This particular trait I feel is not properly computed and needs attention; for now I’m ignoring this trait for all the candidates being tested.

Having tested these tests over multiple hires and potential recruits, I absolutely understand that statistical models are not fool proof but they do provide some indicative personality traits. It is highly suggested that the tests be used for some time over a bigger sample of people to either understand their accuracy patterns or reject the solution for it to be any useful. Otherwise, it is highly likely that despite the availability of such scientific solutions, we will end up making incorrect conclusions.