Jun 152011

I gave a talk last night at Ignite at the Velocity 2011 conference in Santa Clara, CA.  My topic was a humorous attempt at capturing the difficulty and competitive nature of finding a place to live in the city of San Francisco. The event was livestreamed and the video is below:


My talk starts at 13:01. You can view the presentation and speaker notes here:



Click on the “View on Slideshare” button to see the presentation on Slideshare with the speaker notes. Overall, the presentation went better than expected and my timing was spot on for the 15 second auto-advance. 

UPDATE: All the videos are here and my talk starts at 13:01.


 Posted by at 5:41 pm
Apr 152011

Side Note: I’ve realized that my blog posts have become rather lengthy. I’ll try to include shorter articles starting with this one.

Finding co-founders and starting a business is difficult as it is, but starting a company with remote co-founders adds an order of magnitude of complexity. Jambool had 3 co-founders all living in 3 different cities. I lived in Seattle, Vikas lived in San Francisco and Nissim Harels, lived in Israel. You can imagine the difficulty in communication, coordination and overall function. Nissim worked with us several months on Business Plan #1 before he decided to return to academia. That left Vikas & I to execute on the rest of the product.

Coordinating with 2 remote co-founders is a lot easier than 3. But it was still an insurmountable task at that time. So how did we do it? Before I go into details I should stop you here and share with you my first piece of advice:

Do Not Start a Company with Remote Co-Founders

Why not? Quite simply, choosing a co-founder can be difficult to begin with and unless you know and trust your partner very well, then you will be signing up for unnecessary challenges and increasing the chances of failure. You have to make extra effort of discipline, over and above, the normal challenges of working with a co-founder. Nevertheless, working with a co-founder is still better than working alone. If you go down this route and decide to work with someone remotely, then these are my survival tips:

1. Establish a designated workspace

Photo courtesy of Flickr user Jóhannes G

We literally worked from our bedrooms for over 18 months. Technically I worked mostly from my breakfast table and Vikas worked from his guest bedroom that had a desk. Working from home takes incredible discipline. There are far too many distractions, and especially if you have kids or other family members around. Allow yourself to designate a work zone in a specific, consistent location in your home. If you step away, leave your laptop behind on your makeshift desk. That way you will not be tempted to slouch away to the sofa during the work day with your laptop and be distracted by your TV or something else.

2. Set regular sync-ups

Sync up regularly

Our routine was to get onto a Skype video call at 9am every morning. That forced us to be up well before that, but not necessarily dressed for business. Often these morning calls would last 30 minutes, sometimes 2 hours. And usually we would also set an end of day sync up to check on progress. These sync ups were good checks and balances to ensure that communication and progress was being made on both sides.

3. Set daily goals

Set achievable daily goals

One important agenda item of the morning sync up was to set an achievable daily goal. Whether it’s a piece of coding functionality, design document, updating a pitch deck, no task was too big to set an end of day milestone. Committing to a deliverable, forces one to work towards it when you know someone else is expecting that to be completed. It’s much easier to procrastinate if the next milestone is more than a day away, or if nobody cares about it. It also forced us to breakdown a task into day size chunks.

4. Document everything

Draw, write and share documents (Image from Flickr user mattyeo)

In a regular shared workspace, a whiteboard will suffice for expressing ideas and design. When your co-founder is remote, collaboration on ideas must still take place. In our case, before each major feature was implemented, a design document was written. We had pretty crude tools at that time, and would sometimes draw our UX designs on paper, scan and emailed to each other. Skype later added screen sharing and that worked really well. Use Google Docs, Gmail, Dropbox, or your favorite collaboration tool.

5. Travel often

Travel often

As much as your budget allows, travel to meet your co-founder as often as possible. Spend time working on your startup, but also spend time getting to know them outside of work. In my case, I would travel at least once a month pre-funding, and Vikas would come visit every month as well. We would stay at each other’s home to save on costs, and would usually end up working way past bedtime. Once we were funded, I was travelling every 2 weeks. Employees who travel up from San Francisco were invited to stay at my house. In the 2 year period that Social Gold ran as an independent entity, I found myself heading to Seattle’s Seatac airport every single week to either pick up someone or be flying out.

6. Work around other entrepreneurs


Surround yourself with like-minded people

Don’t work from your home all the time. Work from a coffee shop if you can. Especially one that is popular with other entrepreneurs. Simply being in an environment surrounded by like-minded individuals will have an amazing impact on morale. If possible, co-work with other entrepreneurs. I spent much time co-working with Ajit Banerjee, co-founder & CTO of Tracksimple, where we both worked from each other’s houses. It was not relevant that Jambool and Tracksimple had completely different business goals, but simply working with someone else was more fun than being alone.

7. Be completely transparent

Be completely transparent. Image from Flickr user P Shanks

This is perhaps the most important advice of all. There are a number of non-technical tasks that co-founders will work separately on when building your startup. These include a meeting with a customer, investor, potential hire, etc. Regardless of whether the meeting was successful, co-founders need to make an extra effort to over-communicate and be completely transparent. This requires a large effort to keep other in sync to replace the absence of hallway conversations. For example, always get on the phone immediately to debrief on how that meeting went and share that experience. It will keep your partner involved. Keep your IM program running all the time and let the other party know when you are stepping away for a short break. You can never be over-communicating.

Have you started a company with a remote business partner? What are some other tips can you share?


 Posted by at 8:19 pm
Apr 112011

So what exactly is a lifestyle business? A lifestyle business is a business that generates sufficient revenue that allows the owners the flexibility to comfortably adjust their work/life ratio. It sounds like surefire income but like any other business, the risk for failure is just as high. The phase of Jambool in building and monetizing Facebook apps is best described as a lifestyle business that allowed Vikas & I the luxury of having a lot of fun, earning an income and yet having the flexibility to take time off whenever we wanted just because we could. The following is the evolution of how Jambool’s 2nd business plan evolved into a lifestyle business and a top Facebook app developer.

Building on a repeatable platform

November 2007 was the turning point in Jambool history when we realized we were actually onto something big. When we built Hug Me, we wisely decided to build a generic gifting app engine so that we could launch other gifting apps based on other themes. By doing so it allowed us to launch new apps in the future simply by adding configuration and custom images.

Much of the rest of the month of November and December 2007 was a blur; we spent the month scaling out our systems to handle the surge in traffic, including switching dedicated server providers, adding memcache, tuning and partitioning our MySQL database. This was a far cry from a few months before when we were on the brink of shutting down. When we had time to think creatively, we scrambled to think of other clever specialized gifting themes so we could generate new clones of Hug Me.

We launched Send Good Karma and Smiles towards the end of December 2007 and watched usage sky rocket through cross-promotion between the 3 apps. Much like the name suggests, Send Good Karma allowed users to send virtual karma, such as Hope, Happiness, Peace, etc. Smiles allowed you to send funky types of smiles. Other than the icons, names and adjectives, the 3 apps were completely identical. It turns out a fair number of users who send virtual hugs would most likely send virtual karmas as well.

We repeated the formula, and launched additional gifting apps over the next few weeks, based on the Hug Me engine, that specialized in sending specialized emotions such as Wishes, Cheers, Massages, Spanks and the ever classic Fart You. If any conclusions can be drawn by the number of users who liked sending virtual emotions, then one can safely say that more Facebook users liked sending virtual expressions of flatulence than virtual cheers. :)

Early Monetization of Reach Apps

Now that we had a bunch of pretty viral apps, we decided to focus on monetizing them. We placed Google ads on our applications and found that it produced very little revenue. Next, we switched to using Facebook specific ad companies, such as SocialMedia and RockYou ads and saw revenue increased tremendously. The difference being that these latter types of ads cross promoted other Facebook apps, and kept the user in context on Facebook, whereas Google ads produced links that took the user away from Facebook.

Soon it became evident that our ad revenue increased because of the direct correlation between the number of Daily Active Users (DAU), and thus page views and revenue. The industry called these reach apps because they are extremely viral, reach a lot of users, but generate few page views before users bounce off distracted by something else interesting. By simply placing advertising within these apps, there was a high likelihood the users will click on an ad.

We generated about $3,000 in advertising revenue in November 2007 and $8,000 the following month from these apps and ad revenue steadily climbed over the next few months. Later on, we found other sources of monetizing these apps including a sponsorship by Time, Inc, to help build their fan base on their Health.com Facebook page by a partnership with Send Good Karma. Technically, we were profitable that first month as we generated enough to pay for the servers, and it certainly helped when we weren’t paying ourselves a salary. But that was about to change and not a moment too soon.

Our first paycheck

Vikas last took a salary when he left Amazon.com in early 2006, and by late 2007 his once patient wife had given him an ultimatum to get a real paying job within a few months. When it became clear that we would be generating enough income to support the operations we decided to pay ourselves a salary of a whopping $2000 a month. Our first paycheck arrived in our bank accounts on January 31, 2008, just days before Vikas hit his 2 year anniversary of having no income and when the ultimatum expired. If we had waited another month to experiment on the Facebook platform we may not have continued down this route.

Beginnings of Virtual Currency monetization

In early 2008, a new kind of monetization technique started to gain popularity on Facebook. This was the concept of Offer Walls in which Cost Per Action (CPA) based advertising allowed end users to be rewarded with virtual currency in exchange for performing an action, such as subscribing to Netflix, filling up a survey. The advertiser would pay for each completed action, and the revenue would be shared with the ad network and the game developer.We introduced virtual currency on our gifting apps and it did provide an additional channel of income through the use of Offer Walls. We were selling virtual emotions for virtual currency. The portion of revenue earned through virtual currency was still lower than straightforward CPA ads. Unsatisfied with the performance of how virtual currency worked in those apps, we decided to think of new apps that would have virtual currency built in from the start. My next blog article will describe the birth of these engaging apps.

The global reach of Jambool Apps

By January 2008 we had all but abandoned the Social Collaboration site (Business Plan #1). Jambool the Facebook app (Business Plan #2) company was up and running and we had about a dozen Facebook apps all clones of Hug Me. These reach apps propelled Jambool into the Top 15 app developers later that quarter by number of users. Over the lifetime of Hug Me, Send Good Karma and Smiles, they attracted a total of 14 million unique users, most of it during the first 6 months of launch. Facebook acquired 100 million users by August 2008 and one could comfortably say that one in 10 Facebook users that year had used a Jambool app by then. My hypothesis was confirmed by a pseudo-work/holiday trip I took to Brisbane, Australia in February 2008, when I met some friends of my friend Noel, half of his friends had used one of our apps. The moment of recognition was still overshadowed by having to explain what I really did for a living. Uhh. yeah I write apps for Facebook that allows you to send virtual hugs to your friends.

The reach of Jambool apps expanded far beyond Facebook users when we took a portion of the revenue earned from Send Good Karma and Hug Me to make micro-loans on Kiva. Over the life of these apps, Jambool made almost 500 loans totaling almost $13,000 to small businesses around the world. Everytime someone spent virtual currency on our apps, we contributed into a pool of funds for Kiva. This was our way of indirectly translating virtual karma into real world goodness.

Time for an exit?

The early success we had with these apps attracted a variety of attention. Aside from partnerships from other apps for cross-promoting apps, there was more serious inbound interest to acquire our apps and also the talent. We could’ve taken an early exit, made back some of our lost income and moved onto something else. We decided against selling out, as we were having a lot of fun being in control and even getting some attention from the press.

Within a year’s time from the launch of our very first Facebook app, we did eventually decide to abandon the Facebook app business and pivot onto something else. At the peak of our success, it seemed unthinkable to wake up one day and decide to abandon the easy income, the flexible lifestyle and being one of the recognized names in your field. But that is exactly what we did, and more on that in a future blog post.Early ConclusionsUp to this phase in our Business Plan #2, there much of the success was quite accidental and the business had not grown long enough to draw major conclusions, but nevertheless there are early takeaways about a lifestyle business:

  • Build a platform - if we didn’t build a generic gifting engine we wouldn’t have been able to theme Hug Me and branch out to other apps. Send Good Karma ended up being more successful and recognized than Hug Me.
  • It’s still hard work – While most of the time it was fun, there were still many late nights spent scaling the technology and sweating about the loss of users and income.
  • Have fun – We were having a lot of fun thinking up new apps and riding the Facebook app fever. This was probably more important than seeing your revenue grow.

If you’ve built a tech lifestyle business, how it evolve? I’d love to hear your stories.

 Posted by at 7:58 am
Mar 112011

A big concern for every founder of a startup is making a bad hire. In fact, according to Neil Patel’s survey many experienced entrepreneurs cite bad hires their biggest mistake. I think there is a lot of emphasis placed on asking the right kinds of questions during the actual interview to avoid this situation, and insufficient time on streamlining the post-interview debrief to arrive at an accurate and unanimous decision. One of the unique hiring traits at Jambool was the process of arriving at a hire/no-hire decision.

The debrief largely counts on first impressions, not unlike like Malcom Gladwell’s principles in his book Blink. When a candidate leaves the building after their interview, we conducted the post-interview debrief as soon as possible.  We usually didn’t even wait until the following day, so that our impressions and memories are freshest.

The Thumb Test: Photo courtesy of Flickr user: Noam, Jemima & Lila

The debrief ritual goes like this:

  1. Stand Up: Similar to the purpose of stand up meetings in Agile, we make interviewers stand to force them to pay attention and to keep it short. Most debriefs take less than 30 minutes, on average 15 minutes and in extreme cases, up to an hour.
  2. The Thumb Vote: On the count of three, everyone simultaneously flips out their thumb indicating their vote. The angle of the thumb can be in 4 positions: fully down, fully up, slightly down and slightly up. Nobody can remain neutral and must pick a side. This solidifies the position of the interviewer based on his or her impression only. For added drama, we would hide our fist behind our backs and flip it out much like a game of Rock, Paper, Scissors.
  3. Give Feedback: Next, we go around the room having the interviewers explain
    • What questions they asked
    • How the candidate solved it
    • Why they picked their thumb position

    Other interviewers start probing and asking the interviewer more detail on his impression if necessary. Opinions may start to form, but interviewers are forbidden from trying to sway a vote at this time. This part is much like a regular standard debrief.

  4. The Thumb Re-Vote: After all interviewers have given feedback, we go around the room again and ask interviewers if they would like to change the angle of their thumb. Usually there will be one or two folks who switch their angle. Lobbying and swaying may take place, but by this time most interviewers would have formed an unified opinion. The hiring executive makes a decision based on this outcome and seeks to get everyone to a unanimous decision.
  5. Hire/No-Hire Statement: The final step is to affirm the hiring decision.  The hiring executive makes a hire/no-hire statement and goes around the room and gets every interview to agree to the decision. Getting 100% buy in from the team was important as this showed this was a team decision and not any single person.

This technique worked very well and very efficiently, allowing the team to arrive at a decision almost before the candidate has had a chance to reach home. It focuses on the decision making and not on the content of the interview.

Debriefs can be done remotely

It’s preferable if all the interviewers were present in the same room, but we were also to do a debrief over a phone / conference call as long as we stayed with the preceding steps. At Jambool, we emphasized transparency and often conducted the debrief in the big team room where non-interviewers could listen in. Non-interviewers could only observe and interject when they felt we were going off on a tangent or straying from the rules above, but they weren’t allowed to be have a vote.

Limit the stand up to an hour

In very, very few instances were we unable to reach a decision after an hour. In one particular case we were debriefing on a candidate who had a similar but slightly different background. After an hour of heated debate, we postponed any decision making. The next day, after a good night’s sleep, the team met again and repeated step (4) again. Much to my surprise everyone was more positively inclined now even though there was uncertainty the day before. We made that hire.

It works for every role

Our first few hires using this method were software engineers, but we found that the same debrief process worked for every role we hired, regardless of full-time, part-time or contract positions.

A fast decision will impress the candidate

Once a Hire decision has been made, call the candidate that same day. We found the speed at which you can give positive acknowledgment that you are interested will bowl the candidate over. Most companies, especially the larger ones, take a long time to reach a Hire decision. This is where the agility of being able to present an immediate offer is to your advantage. Almost all of the hires at Jambool accepted were hired this way, and accepted on the spot or within a day.

Reference checks won’t change a positive decision

In almost all cases, references didn’t fundamentally change the Hire decision. Candidates are inclined to provide only positive references anyway. We went through the reference checks as a formality, and usually found out more about the referee and made mental notes to recruit them in the future. After all, a stellar candidate will have provide references whom they respect and are likely as good as themselves.

This quick decision making still depends on having a skilled team that understands their role in the interview process and having the proper skill to interview others in the first place. I’ll cover the actual interview process in another post.

 Posted by at 10:46 am
Mar 032011

In my earlier post I described the product reasons that caused the collapse of Business Plan #1. Today I will cover the change that led us out of Business Plan #1 and into Business Plan #2.

The end of the beginning

By the end of August 2007, it was pretty clear that our social collaboration site idea was going nowhere. Up to this point, throughout the 6 months that worked on the Social Collaboration site, we went through a roller coaster of emotions. One of the reasons I left Amazon.com was due to the stagnation in AMZN stock price, so as did many other talented employees who went on to start their own companies. As fate would have it, the week I left Amazon was week that AMZN stock price started its incredible ascent from about $40 to about $80 by the end of August that year. The opportunity cost for leaving my stable job now had real monetary value, having to forgo any future stock based incentives that was left on the table when I resigned. Moreover, both co-founders were not drawing a salary which was heaving more salt into the wound.

In addition, I had just returned from a vacation from Malaysia and spent time relaxing at Redang Island - perhaps one of the best beaches in the world with pure white sand and calm warm water. Life was good personally, but there was some serious consideration about what to do about the startup path that we took. I personally, gave myself 2 years to work on a startup before I gave up if it failed. By this time, Vikas had been tinkering around on startup ideas for over a year, and time was running out for him.

During the entire span of the startup journey, this was the moment of absolute rock bottom.

An early hand drawn mock-up of the header of the Social Collaboration site


Pivoting on your platform

Earlier that year in May 2007, Facebook, opened up its APIs to let applications tap into it’s user base. In a moment of “what have we got to lose,” Vikas decided it was a good idea to see if we could build a Facebook application built on top of our social collaboration platform. We launched the application Shared Memories soon after Labor Day 2007 in early September. The idea was to allow friends to create memories of the past by uploading pictures or videos and discussing it collaboratively. The underlying data objects and storage was based on the same objects used by the Social Collaboration site. The presentation layer was new and custom for the Facebook platform. We were hoping to build a viral application that would create enough users and then funnel them into our standalone website. We spent advertising dollars promoting our application.

We learned a few things quickly about this experience

  • Facebook users like to do fun stuff, and very small percentage felt creating memories was worthwhile their time, as it seemed like real work.
  • When we turned advertising off, users would stop using our app.

Within a few weeks it became apparent that Shared Memories was not going to save us.

Trying to create a funnel

We knew there was something that could be done to harness the (then) 50 million users of Facebook. Surely there has to be a subset of them, no matter how small, that would find the value of our Social Collaboration site useful. It was clear that we needed to create a funnel of some sort.

Our next Facebook app was a more fun gifting app called Balloonz. It was greatly inspired by the flourishing of gifting apps already on Facebook. The app allowed users to upload their own images, and then send these virtual gifts to their friends. The twist was that these gifts were placed into a virtual balloon and the recipient can choose to pass on the balloon if he or she thinks it’s a gag gift or let it explode on their profile. Once the balloon “pops” would the mystery gift be revealed. In order to get this out ASAP, we roped in my friend, Arun, who was one of the guys who went to Redang Island above. Balloonz had several hundred Daily Active Users (DAU) which was about 10 times more than Shared Memories generated.

This was also more users we saw ever using our Social Collaboration site, so we considered it amazing, putting aside the fact that the  profile of the average Balloonz user (one who enjoys sending pictures of poop and other juvenile images to their friends) would likely not be same target user who’d use digital social collaboration to plan trips and tasks.

Several hundred DAU was great, but still not a big hit. Within the Balloonz app we placed our own advertising links to get users to try using our external Social Collaboration site. That generated some users but we realized we needed an even bigger funnel.

Drawing inspiration from others

One of the first conferences we went to as startupers was Community Next in October 2007. This was a conference all about building on social networking platforms. There were a fair number of interesting speakers who described their techniques to viral user growth when building their early Facebook apps. One thing great about the startup community is that when there is success, entrepreneurs are not shy about sharing their learning’s at speaking sessions.

One idea that came out from this conference was that we should try to continue building simple gifting apps but throw in the element of incentives. In layman’s terms this means is to provide a reward to users to keep them using more of your application especially by spreading it to their friends. For a Facebook app, this means rewarding a user with status, points or fame if they send an invite or notification to their friends. Their friends would do the same, and the spiral continues. Affiliate programs had been doing this for years: refer a friend, they’ll give you $10 off the next purchase.

Just Hug Me

The 3rd Facebook app we built was called Hug Me. This was a dead simple application that allowed users to send various assortment of virtual hugs, each one with a cute image.  In order to send different kinds of hugs, users had to first send a certain number of hugs to their friends. We weren’t the first app to send virtual hugs, but one of the first to specialize in just sending hugs. Hug Me was launched in early November 2007 and took off like a rocket, by the end of the first day, we had over a thousand DAU, and by the end of the first week there was over 10K DAU with no advertising!

The Hug Me header with custom ad promoting Shared Memories


Facebook has since shutdown most of the virality channels that made this kind of success possible. This includes the prohibition of incentivizing users and use of notifications. The number of invites have also been clamped down. The Hug Me concept and name has also been copied several times over. Including, RockYou which changed the name of their successful X Me app to Hug Me. There is something oddly poetic about this, since X Me was the one of the top gifting apps on Facebook at that time which was an inspiration for Hug Me in the first place. Oh well.

We placed advertising on Hug Me, and by the end of the first week we were making almost $100 per day in ad revenue. This wasn’t a lot but given that we had just about zero revenue before that, this was incredibly uplifting. Not to mention the continuing spiral of growing users was exciting to watch and actually started to strain the capacity of our single server at that time.

Hug Me alone didn’t save the the startup, but it was definitely the turning point of Business Plan #2. Over the course of the next 10 months we would create 22 more Facebook apps, a MySpace app and even a Bebo app. You’ll have to read about that in another blog post.

Summary of Lessons Learnt

There are a number of takeaways from the experience above for a tech entrepreneur:

  • Keep your core technology and pivot the idea around that: We started building a Facebook app that was based on the same core of the Social Collaboration site. Then we took our learnings from building one app to another app, and the finally onto Hug Me which was vastly different but you could see a clear transition from one to the other.
  • Conferences are useful: Not just for networking but to be around entrepreneurs looking to solve similar problems and to seek ideas.
  • In times of darkness find the thing that keeps you going: For us it was seeing the growth of users and actual revenue, no matter how small.
  • Sometimes people just need a hug: Virtual or not.

Wait, so how did the Facebook app business (aka Business Plan #2) , transition to a virtual currency and payments business? Well, that’s yet for another blog post. Stay tuned.

 Posted by at 7:21 am
Feb 232011

In his 2002 book, The Tipping Point, Malcom Gladwell referred to people who participate in many social circles as connectors. These are folks who have the ability to make friends and acquaintances that span many different worlds. They are highly networked, maintain great relationships with each circle and can join a new circle very easily. Super Connectors have hundreds of such social circles. While neither my co-founder or I were nowhere close to being a Super Connector, but as founders we had to be a Connector of sorts and tapped into our disparate social & professional circles to hire the team.

Have you reached out to everyone in your network? Photo by Flickr user luc legay

One of the rules that we tried to follow in hiring was to not hire friends or family as much as possible. If things go awry, then it makes for an awkward conversation and possible strained friendship. I did end up recruiting my sister-in-law to work part-time for us, but generally, I stayed away from trying to recruit friends. So, if you know me, and I didn’t ask you to consider Jambool, it means you are my friend. :) We also had a similar rule of not taking investments from friends and family either, but that’s another story.By the time the Jambool was acquired in August 2010, we had hired over 30 employees either part-time or full-time over the course of 2 years. Here are some interesting facts about how these 30+ individuals came together to help build Jambool:

  • Only 2 employees came through a recruiter, everyone else was either somebody known personally or referred to us by someone we knew
  • We did not directly advertise any of jobs (see footnote), recruiting was done exclusively through our network
  • It was not until we received our second round of funding did we put up a Jobs page on our corporate website.
  • Only 1 person was hired who applied through the Jobs page, but only because someone we knew told him to do so

I enumerated the variety of different circles that formed the source of all the hires we made. These roughly were

  • Vikas’s work circle
  • Vikas’s professional circle (3 groups)
  • Vikas’s social circle
  • Reza’s work circle (3 groups)
  • Reza’s social circle (2 groups)
  • Interview candidates
  • Reza’s & Vikas’s  entrepreneur circle
  • Employee’s social circle
  • Employee’s college circle
  • Employee’s work circle (3 different employees)
  • Investor (2 investors)
  • Recruiter

It’s obvious that one would expect a recruiter, investors and employees are good sources of hires, but there are some other interesting observations about the above:

1. Social circles are great sources for hires that are not like you

I am often asked by aspiring technical founders, how and where to hire non-engineers. If you’re an engineer  it’s pretty straightforward to make a list of all the smartest engineers you’d like to recruit to your startup. But it gets harder to make a list of non-technical folks you might want to hire. This is when the strength of the connections from your social circle cannot be underestimated. These are folks you meet up in social situations in a group of friends and acquaintances. I remember telling all my non-techie friends, each time I met them about the all the positions that were looking for at that point in time, hoping but not counting on whether they knew anyone.  Once was for a position in customer support, and another time for a graphic designer. We were fortunate and we were able to hire at least 2 employees through referrals from two different social circles.Social circles were also great to tap into totally unexpected pools of talent as well through obscure relationships. One of the more odd chains in relationships was of the graphic designer who designed the initial version of the Social Gold logo; who was a good friend’s husband’s college buddy who lived 2000 miles away in Chicago. Then there was the time we wanted to write a press release and couldn’t afford to pay the steep fee commanded by a professional PR writer. We ended up using the writing services of an employee’s wife’s journalism school classmate to get that job done at a more startup friendly rate.

2. Use your professional relationships to help you recruit

A professional circle includes folks who are service providers whom you work with to support your organization such as lawyer or accountant. Once we achieved a certain level of success, an interesting phenomena happened: a fair number of employees were referred in by professional relationships. For example, we hired a payments expert through the introduction and recommendation of our part-time accountant. An in another example, some other engineers came through the recommendation of a customer who had them do some contract work in the past. In both situations, we had created deep bonds working with them. i.e., through the weekly meetings with the accountant, and through the frequent updates and good impressions left on the customer.

3. Every interview candidate is also a great source too

One particular source that produced a fair number of referrals was candidates themselves who did not, for whatever reason, join Jambool. In one example, we were looking to hire a Customer & Community Manager. We interviewed a candidate and that person did not join and we parted ways amicably. A few weeks later that person introduced us to one of their friends for a similar position and that friend was hired. This was totally unexpected as we were unable to get to an agreement with the first candidate, and yet we left a good enough impression to have that person be a spokesperson for the company. There were also other examples of candidates whom didn’t receive job offers but referred their friends after the fact. If you can leave a lasting impression with everyone you meet, then they become an invaluable source of referrals.

4. Meet your ex-coworkers regularly, they’re the best source of hires

This goes without saying but the best source of hires are people that you’ve worked with in the past. It takes more than just remembering names of awesome coworkers for future reference. Once you leave the your previous company and move on, it’s still very important to keep in contact with folks in these work circles. During the early Jambool years, it was much harder to recruit and the fact that we evolved our hiring into the myriad of social circles is an evolution of the company and not by randomness. We depended heavily on recruiting people whom we’ve worked with in the past. One such work circle generated about 7 potential candidates who entertained the idea of working for Jambool, of which 3 eventually joined full-time. This work circle was from a group that worked together more than 8 years ago and yet kept in touch with semi-annual meet-ups.Invest in time to meet up with ex-coworkers, you’ll never know but they might end up working in your startup in the future.

Good to bootstrap, but is it sustainable?

It’s hard to say, if given the independence of Jambool, how long we would have been able to sustain the rate of growth simply through our network. There have been a lot of factors that have changed since then, including increased competition in the hiring market, higher salaries, tapping out our networks, etc. Towards the tail end of Jambool we started offering referral bonuses (iPads instead of cash) to spur the recruiting. We were acquired before we could evaluate the full results of this incentive. In general, this technique has worked very well in bootstrapping and growing our tech startup.

Footnote: Our recruiter may have anonymously advertised our positions but he did a lot of pre-filtering beforehand before the resumes reached us.

 Posted by at 4:21 pm
Feb 162011

It may seem ironic that the first real post of my blog is about failure, despite the recent success. I think it’s important to recognize where you’ve gone wrong so you don’t repeat it again.

Jambool is most well-known for it’s marquee product –  Social Gold – a virtual currency monetization platform. The Social Gold product launched in 2008. What many people don’t know is that Jambool was founded a few years earlier and had a very different business goal. Social Gold was actually Business Plan #3. The first version of Jambool was a social collaboration site. The idea being that you could share digital content such as snippets of web pages, emails, photos, documents, etc, with a friend or group of friends. In theory, this could have been a very powerful tool. In practice, this was very hard to make into a great product given the reasons below.

Start by failing but don't turn into a wreck. Image courtesy flickr user moo_sa


In retrospect, I can think of a hundred reasons why Business Plan #1 failed with customers & investors. Here are the Top 3 that can be generalized for any tech startup:

  • Lack of clear monetization strategy
  • Not having the right match of skills
  • Building in a vacuum

1. Lack of clear monetization strategy

The monetization story for the collaboration site was to start out with CPM based ads. Everything was based off user-generated content (UGC) and we’d use SEO to optimize the pages and generate revenue through Google AdSense. Maybe, at some later point we’d charge for subscriptions for premium features. Giving away something for free and then figuring out how to monetize was a proven model in the early 2000s and Fred Wilson coined this as Freemium. This monetization strategy was not so hot in 2007. The US economy was on the verge of collapse and VC money was drying up. Investors had also grown tired of startups who’s only monetization strategy was ad based. We also made some wild-ass assumptions about the potential CPM rates that our pages could generate. That was simply impossible simply without targeted high quality ads from an ad partner as well as really, really relevant content. Needless to say, we were turned away from many investor meetings for not having a better story for making money.

2. Not having the right match of skills

Both my co-founder & I had technical backgrounds in building large scale, highly distributed platforms. Our customers of the systems we had built in our prior jobs were other internal teams and hardly the external customer. The systems we built were capable of handling thousands of TPS and ensured the delivery of billions of $ in ecommerce transactions and requirements were driven by how fast you could move bits around. The social collaboration site was furthest from needing to be a scalable platform. It was a consumer oriented site, heavy in Javascript, AJAX, CSS and we had grand plans to make it as Web 2.0 as possible. Neither of us had experience in UX development or design and this was the first real attempt at building a consumer site that had to be usable by the average consumer. Overtime the two of us overcame the technical learning and figured out how to work with Prototype & jquery, but really sucked at UX.

3. Building in a vacuum

We made a product that was kept under wraps all the time. Users who browsed to our home page were required to log in and only we could generate login credentials. Every now and then we’d ask our friends and family to use our product as alpha customers. Being polite, everybody logged in once and gave words of encouragement and feedback and never used it again. Nobody told us we were building a crappy product, they simply stopped using it. Moreover, the set of alpha customers were not the type of target consumer that would use this product on a daily basis.

Needless to say by Labor Day 2007, after a solid 9 months after the first line of code had been written for the social collaboration site, we came to dire conclusion that needed to do something drastic. I’ll blog about what we did in another post.

Lessons Learnt

We could have solved each of the above problems by first proving we could generate a high trafficked website by opening up the site earlier, hiring a technical team that had the proper skills and aggresively partnering with vertical customers that could give feedback quicker. The catch here was the hiring the team. We simply didn’t have the capital to hire folks with the right skills at that time.It turns out that having good answers to address these 3 pitfalls are high on the list of things investors look for as well.As a comparison, when building the Social Gold product we took great pains to learn from our mistakes Here’s a little table that summarizes the differences:

Social Collaboration Site (aka Business Plan #1) Social Gold (aka Business Plan #3)
Monetization Strategy Ad based/Freemium Revenue Share
Skills Limited team, mismatched skills Able to hire incrediibly skilled team
Early Feedback Private Alpha/Beta. Full-on Vacuum mode Incorporated customer feedback from the beginning

Granted, that by the time we built Social Gold, we had received sufficient funding to make most resource constraints go away. But there was a period of time where we were still able to build a sufficiently successful business pre-funding. Check back in the future to read about Business Plan #2.

 Posted by at 5:05 am
Feb 102011

In April 2007 I quit my job at Amazon.com and co-founded Jambool with Vikas Gupta. Along the way I learned what it took to build a product, run a company and eventually see an exit. Google acquired Jambool in August 2010.Starting a startup has been an incredible journey, ever since the first day I’ve been telling everyone  that I would someday write a book about my experiences. I can’t write a book just yet because

  • all the details of  the business are still material
  • I won’t have content to make a very thick book

The next best thing is to start a blog, that I can jot down my learnings in general, and leave out all the juicy stuff that would get me in trouble. Perhaps some day I’ll write a book, by then I hope to have several more interesting entrepreneurial adventures.

 Posted by at 4:44 am