Small Ideas

Finding the next big idea by exploring small ones – by Rishi Khaitan

Virtual to Reality

without comments

One of the big stories out of this year’s 24 Hours of Lemans race was a 21 year-old driver by the name of Jann Mardenborough. What’s notable about Jann is that two years ago he was pretty much a regular teenager who had no racing experience. Most racing drivers that rise to the Jann’s level at his age start developing their race craft at a very young age. For example, Fernando Alonso, a top F1 driver, started his kart racing career at 3 years old! So how did Jann do it?

It is true that two years ago Jann was a regular teenager. However, he did have a unique skill. Jann was extremely talented at Gran Turismo 5, the most popular racing simulator video game on the PS3. Jann used that talent to win the 2011 Nissan GT Academy in Europe. What’s GT Academy? Here’s the official blurb:

GT Academy seeks to find the fastest Gran Turismo 5 gamers on PlayStation 3 and turn them into real racecar drivers. For the winner of GT Academy, racing becomes his job. His life. Within months, he’ll be competing on the international circuit.

GT Academy is an online tournament in which Gran Turismo gamers compete in a series of challenges. The winners have the highest rankings (the fastest times) across these challenges. Six months after winning the tournament, Jann earned his international racing license and by January was on the winner podium of the 24 hours of Dubai. Jann was not an isolated data point either. Lucas Ordonez, the first ever winner of the GT Academy back in 2008, is also well on his way to a successful racing career.

To understand how fascinating this is, remember that professional race drivers at the upper echelons of the sport, begin their career in most cases before they can even ride a bike. The notion that a gamer can successfully transition from simulation to reality and and do so successfully is pretty revolutionary. Simulators are becoming so good at re-creating the real world experience that the skill developed in the simulator actually does translate well to real-life.

Racing is an expensive sport to participate in. Many talented drivers are unable to pursue a career in the sport due to lack of financial resources. Amassing 10,000 hours of real life seat time in a race car is extremely expensive (think low 7 digits in dollars) and requires a lot of support. It’s a huge barrier. However, the capital required to earn 10,000 hours of Gran Turismo 5 is maybe a thousand dollars (a PS3 and a nice wheel/pedal setup) and you can do it in your living room. In a sense, this sort of democratizes the sport. Now, of course, this is not going to happen overnight. There’s plenty of people/establishments involved in the sport today that don’t welcome this change. But as long as Polyphony Digital (the makers of Gran Turismo) continue to close the gap between simulator and reality and as long as companies like Nissan are willing to source their race drivers via a video game, this trend will continue.

Remember, though, that Gran Turismo is an amazingly sophisticated simulator. It has an outstanding physics model, life-like graphics and sound, and fleixble input. All this comes at a price. There’s been hundreds of millions of dollars that have gone into the development of hte Gran Turismo franchise. The cost is justified by the fact a driving simulator is a form of entertainment that a hundred million people around the world are willing to pay $50 for. Other simulators are not so lucky. So then it becomes a question of can you build a sufficiently realistic simulator that the market for that simulator can bear?

An exciting innovation that should dramatically help this problem is low-cost motion-based sensors such as the device by Leap Motion. These devices should make simulators both more realistic and less expensive to develop. Hence, I think we’ll start to say a lot more simulators that otherwise would not have been either feasible or viable or both. Newly feasible simulators may be things like games for other sports in which the interaction mimics playing the sport. Such simulators already exist for golf, but they’re cost prohibitive for mass adoption. Performing surgeries or any sort of skilled work such as assembly line tasks may all be trainable at low-cost via a simulator with motion sensors. And it may not be viable today to create a simulator for each of a thousand tasks for assembling a Honda Accord, but that may change if the cost to develop a simulator for each task decreases by an order of magnitude. I think there’s a lot to be excited about here and I will definitely be keeping my eye on startups that emerge from the $25 million Leap Fund that Highland Capital just announced.

Written by Rishi Khaitan

July 5th, 2013 at 6:40 am

Hero Products

without comments

iPhone 1%I love my iPhone 5. Its small enough to use with one hand, its screen has amazing color accuracy and it’s damn fast. But that’s not what makes the iPhone 5 so endearing for me. Out of all the phones I’ve owned (including every other iPhone I’ve owned), the iPhone 5 has an astonishing ability to cling to its last 1% of battery life.

Last Saturday I was headed out to a party, but I needed to coordinate with friends. After I had already hopped in a cab, I realized that my phone was at 1%. I said a quick prayer to the Apple Gods and proceeded to make a 5 minute phone call, use Google Maps to get directions to where I was going, sent a couple of texts and waited another 10 minutes to receive an important text reply. After all that, the phone still proudly indicated 1% battery remaining. My iPhone had my back when I needed it most. I was beyond impressed. In fact I was so impressed that when I finally met up with my friends that night, I couldn’t stop talking about how amazing a product the iPhone 5 is.

I greatly suspect that this behavior is the result of clever battery tuning on the part of some brilliant engineers in Cupertino. Apparently, I’m not the only one to wonder about this. If Apple did indeed artificially inflate the length of 1%, they didn’t invent this concept. Auto makers have long designed fuel gauges such that there is about a gallon of “reserve” gas left in your tank even when the car’s computer reports that you have 0 miles to empty. And similar to the 1% battery effect, surely some of the most happy/satisfying moments that a car owner has is pulling into a gas station after having convinced themselves that they weren’t going to make it.

It occurs to me though that the products in my life that I have the most emotional connection to are those that come through for me in critical situations such as the ones I mention above. So is this phenomenon something that only physical projects can embrace? I don’t think so. Software can too. A great example of this is the Undo Send feature in Gmail. It’s saved my butt at least once a week for as long as I’ve been using Gmail. And every time I click Undo, I fall in love with Gmail all over again. Another example that comes to mind was Microsoft Office’s document recovery functionality that could recover most of your document after a system crash. That was outstanding software design when that feature was first introduced in the early 1990s and it without a doubt saved me hours of agony in high school alone.

If you’re an engineer or a PM that’s reading this, I’d urge you to take a few minutes and think about an answer to this question: When and how can your product be your user’s hero?

Written by Rishi Khaitan

July 3rd, 2013 at 5:14 am

Quick Thoughts on Apple’s Crazy Week

without comments

Wow, what a week for Apple in the news.

  • First, the iPad launches.  The media embargo lifts.  Favorable reviews (would you really expect anything less from Walt Mossberg?) pour in, but some negative reviews surface as well.  Turns out, Apple sells 450,000 units in the first week.  Success?  Too early to tell, but those sales numbers a bit north of where I thought they would be.
  • iPhone OS 4.0 is unveiled on Thursday.  Lots and lots of good stuff in here.  Apple seems to have done a great job thinking thru the multitasking issue.  Background GPS + local (local as in local to the phone..not notification system) notifications is going to be huge.  And it’s just plain hard not to be giddy about  background voip and music.  As many of my Android friends have reminded me, yes Android has supported this since ’09, but that’s not going to kill the glee that thousands of iPhone devs and tens of millions of iPhone (err make that iPhone 3GS +) users will feel come the official launch.
  • The launch event also confirmed iAds, that is Apple’s very own ad network for iPhone apps.  We all knew this was going to happen once Apple bought Quattro.  Some of my friends in the mobile ad world are calling anti-trust on this.  One of Steve’s key points about iAds is that the in-app ads will actually expand into an HTML5 window that floats above the app and allows the user to interact with the ad without leaving the app.  Now, if this functionality is not made available via the SDK, then yes I do believe this is anti-competitive.  If they don’t then there’s really nothing that stops any other ad network (Google/AdMob, etc..) from implementing the same.  However, Steve claims that iAds will strive to make mobile ads more engaging than current mobile ads.  I think this goal + the official Apple association + the non-disruptive-to-the-app user experience will be very appealing to big brands.  Steve estimates that Apple could be pushing 1B ads daily given current iPhone app usage.  Obviously that number is sort of like a TAM measurement at this point.  But with this announcement Apple is arguably the new #1 player in mobile advertising…for now.
  • Finally, the terms for iPhone 4.0 SDK imply that developing an app using a non-Apple approved languages is prohibited.  What this means is that Adobe’s Flash-based iPhone developer tools are banned.  In his (personal) rant today, Lee Brimelow, evangelist for Adobe, makes it clear that this has nothing to do with Flash player support for the iPhone (which I think it’s pretty fair to assert now is not going to happen anytime soon if ever because this effectively would allow apps to be delivered to the iPhone without going through the iTunes App Store).  But, banning developer tools?  Just doesn’t make any sense to me.  Why is Steve so anti-Adobe?  I guess he forgot the 1990s when Adobe was arguably a key reason for the survival of the Macintosh (and hence Apple itself).

Finally though, one can’t help but wonder if this closed-off approach to the iPhone is going to eventually do Apple in once again.  I tell a lot of friends that Steve Jobs is a visionary and innovator for the ages..one of the greatest ever.  However, he also has a huge ego.  I argue that Steve represents both Apple’s ability to create new paradigms and Apple’s inability to capture, in the long-term, dominant share of the market that paradigm shift creates.  Of course, the clearest example of that is the Mac and the PC market.  Will this be repeated again in mobile?  Well, Apple once again is alienating hardware vendors (by not opening the iPhone OS), alienating developers (limiting funtion and distribution of apps), alienating many users (for some of the above reasons as well as pricing and carrier limitations), and is now potentially alienating the ad/monetization world.  We’re already seeing numbers that suggest Android’s smartphone market share is catching up real fast on the iPhone.  My guess is by the end of 2011, we’ll start to know the answer to this question.

To end this post on a positive and inspiring note, if you haven’t watched this video of a 2 year old girl’s first experience with the iPad, you must watch this video.  The iPad may not change the world but man you can’t help but watch this and believe Steve when he said the iPad is “the most important thing I’ve ever done”.

Written by Rishi Khaitan

April 9th, 2010 at 9:22 pm

Tagged with ,

Quick intro to Hadoop

without comments

After hearing words like “Big Data”, “Hadoop”, “MapReduce” and “NoSQL” more and more recently, many friends and colleagues have decided that they need more than just a vague idea about what this stuff is all about.  A couple of years ago when I started on this same discovery path, info on this topic area was far more fragmented. Recently, though, with an explosion of “big data” interest and companies, there are many great resources. In fact, Cloudera is a great starting point.  First, is an interview of Cloudera (essentially they are a RedHat for Hadoop) CEO Mike Olsen by Robert Scoble. Mike nicely pulls back the onion step-by-step, first by walking through the computing/data trends that have driven this “big data” movement, which leads into who the key developers/contributors have been, and finally what the ecosystem looks like today.

The key takeaways that I like to focus on are:

  • Data sets have gotten MUCH bigger.  It used to be that data was human-scale.  In other words, data was created by  human-driven activities.  Things like credit card transactions, bank accounts, static html pages, and e-mail.  This type of data consumed GB’s maybe even  some TB’s of data.  Now we live in a world of machine-scale data.  Machines with code that are producing new data every  millisecond.  We’re talking about things like sensors and data-driven web sites that generate, repurpose or syndicate data.  The Web is described today in PBs (petabytes = 1,000 TB).  Many large-scale services deal with datasets in the 10s-100s of TB or larger and process TBs of data daily.
  • Physical I/O throughput is the bottleneck – The amount of data we can store on a disk is on the order of  1,000x what we can store in RAM as a working set.  Commodity hardware can handle 10′s of TB of hard disk storage but only 10′s of GB of RAM.  So what does that mean?  Well, in order to process any arbitrary subset of data within a massive dataset, this data needs to be loaded into RAM.  Even fast HDs today can only read at a theoretical maximum of not even 100MB/sec.  And even at this rate, reading just 1TB of data into RAM will take nearly 3 hours.  100TB would take more than a week.  When I worked at Ariba from 2001-2004, we basically told our clients to go out and buy the biggest Sun box they could afford and throw Oracle on it.  Whatever the maximum througput of the resulting DB server was essentially set an upper limit of overall application throughput.  This was in a 1TB world.  That’s just not good enough today for many applications.  The Hadoop/MapReduce concept breaks this paradigm by saying instead of having 1 uber expensive DB box, let’s put small chunks of data into large clusters of cheap, commodity hardware.  Hadoop achieves massive parallellization and thus fights orders of magnitude larger datasets with orders of magnitude higher I/O.
  • Finally, Hadoop is not the “replacement” for the RDBMS.   solves new kinds of data problems that SQL databases were never designed for.  For financial systems, transactional data needs to have perfect integrity.  If data is updated then the very next query that references that data must see the updated data.  On the other hand, if you add a friend on Facebook, it’s not the end of the world if a split-second later your # of friends stat still shows 467 versus 468.  Further, data like webpages, profiles, and tweets are not highly structured.  Such data should be stored differently because it is accessed and processed differently.  Furthermore, such data tends to be read orders of magnitude more often than it is written.  (Think about a typical webpage and heck a tweet for example can never be updated).  So instead of making reads expensive, let’s make writes expensive for such applications.

If this kinds of stuff sounds interesting, take a few minutes and watch the video:

After watching this, if you’re ready to dive deeper, watch Clouder’s Hadoop training video series starting with this video on Thinking at Scale

Written by Rishi Khaitan

April 5th, 2010 at 4:12 pm

Thoughts on Great Companies

without comments

A couple of weeks ago, someone asked me: What makes a company great?

I thought about it for a minute, but couldn’t at the time come up with an answer that I was satisfied with.  Since that conversation, I have continued to think about this question and have arrived at a definition that I actually like quite a bit:

A great company is a company that not only makes a great product, but innovates on the business model to make the company’s product accessible.

Google web search is obviously a great product.  But what if you had to pay for each search?  Amazon’s mission has always been to offer the best possible customer experience.  But what if Amazon’s prices were higher than those of other retailers?  Thankfully, we don’t have to answer such questions because both Google and Amazon innovated on the business side of their respective companies.  Advertising allows Google to offer free search and superior business operations enables Amazon to offer extremely competitive prices.

Intuitively, this definition makes a lot of sense.  “Great” companies tend to be those that change the world in some way.  Well, the way you change the world is generally not by delighting only a small segment of customers.  Instead, it is by putting your products/services into the hands of a large many customers that real impact is made.  From Ford to Walmart to Intel, these companies are considered “great” because they led paradigm shifts in their respective product/service categories.

But, what then, about a company like Ferrari?  Is Ferrari not a great company?  Since the company’s start it has sold essentially hand-built exotic sports cars that are far out of reach for the vast majority of the population.  They could have gone the direction of the Corvette and tried to innovate in terms of design, production methods, etc. to build a more affordable, high-volume, yet still potent sports car.  But they never have.  Yet Ferrari’s contribution to motorsports is staggering and undeniable.

Taking a step back, does this then mean that a company that chooses to stay focused on a niche will never be eligible for “great” status?  I guess it depends on what one really implies by the word “great”… Hmm, I guess I need to do some more thinking about this.

Written by Rishi Khaitan

March 22nd, 2010 at 1:31 am

Great Online Advertising Blog

with one comment

I was very excited to stumble upon Mike On Ads, a fantastic blog authored by Mike Nolet who was at Right Media before co-founding AppNexus and is currently CTO.  I ended up spending a couple of hours today reading through many of his posts and learned a lot.

I really like how Mike dives into topics such as RTB from the various stakeholder’s perspectives and goes deep enough on the tech details to add color to his points.  For example, when talking about the cookie-matching challenge between bidder and exchange, he first explains at a high-level why it all matters for RTB and then draws the following diagram to explain exactly how this issue is being addressed:

Some of his recent gems are his series on Real Time Bidding, this post that does a great job of disambiguating ad-exchange/network lingo into functional meaning, and posts like this on industry trends.  So many other great posts on ad exchange vs. network, behavioral ad, cookies, etc.

Mike On Ads is an exciting addition to my blogroll and will now be the site I recommend to others who are looking for a primer on online advertising!

Written by Rishi Khaitan

March 14th, 2010 at 10:41 pm

Tagged with

OnLive Launching June 17th

without comments

OnLive is one of the most exciting startup launches of 2010. Why? Well, OnLive is trying to disrupt the video game industry. Here’s a description of OnLive from their recent blog post announcing the launch:

OnLive fundamentally transforms the way users experience games and interact with each other, and in time, will transform the way games are developed and marketed. By distilling specialized game hardware out of the equation, OnLive will allow games to be played as a pure media experience on virtually any device, with the same flexibility and instant-play experience that we’ve come to expect from online video and music.

You might be thinking that any company description which has the word “transform” twice in the first sentence is likely BS. So, for a more in-depth, tangible description and demo of OnLIve, I would HIGHLY recommend you watch this presentation by CEO Steve Perlman (Quicktime, WebTV) from the 2010 Dice Summit:

Game TrailersE3 2010

Even if you aren’t interested in the video gaming industry nor OnLIve, the first few minutes of Steve’s talk is very interesting. He basically runs through the evolution of media consumption. One interesting stat that’s really stuck in my head is that torrent traffic, while still representing a huge (think 50%) chunk of total internet bandwidth worldwide, actually peaked last year and is now on the decline! Steve says it’s largely because people are growing impatient. They want their media now and thus are preferring streaming/on-demand experiences.

Since their first emergence out of stealth mode at GDC ’09, the chatter about OnLive seems to be slowly shifting past “wait, does this thing actually work?” to “great, so how much will this cost?”. And it is indeed the pricing issue that is one of the most intriguing aspects about OnLive. Well, as of last week’s announcement at GDC ’10, we have the answer..at least part of it.

OnLive will charge a monthly fee of $14.95 for access to the service. This does include access to game demos and other media bits (think PSN or XBox Live), however, this does not include any actual games. Access to games will be available for rent or for sale on an a la carte basis.

So, let’s run some quick numbers:

OnLive costs 14.95/month x 12 months/year – 10% multi-month discount = $160/year

Gamefly costs $16/mo (1 game at a time) x 12 months/year = $200/year

A PS3 costs $300 and let’s say at this point has a 4 year lifetime so amortized is $75/year

So how much will games be on OnLive?  If you take a look at about the 34:30 minute mark in the above video, Perlman shows a graph of which players in the value chain get what margin of a $60 video game. In that graph, the publisher gets around $27. Perlman contrasts this with the OnLive model where his graph appears to show that OnLive is offering about a 70% publisher /30% OnLive split to developers. At that split, OnLive needs to charge about $40 per game for publishers to make at least $27/game sold. (It’s probably also worth noting that publishers also win because that game, because it’s digitally delivered, cannot be resold on the used game market nor can be pirated.)

So, let’s see..

Scenario A: Consumer owns PS3 and rents 1 game/month. Using PS3 + Gamefly that will cost 75+200=275/year. Using OnLive that would be 160+rentalpricex12. Do some simple math and rentalprice needs to be about $10 to equal the PS3+Gamefly scenario.

Scenario B: Consumer owns PS3 and buys 4 games/year (assume $60/ea). Cost of that is 75+60×4= $315. Using OnLive that in my estimate would be 160+45×4 = $340. To make the prices comparable the game purchase price on OnLive would need to be under $40.

Of course I haven’t discussed other advantages to the consumer such as being able to play any game regardless of platform. On the other hand, for gamers that already own a current-gen console, the cost of the console is a sunk cost and thus the argument for OnLive becomes more difficult. A pricing scheme that is interesting to think about is usage-based. In other words, gamers are charged per hour that they play any given game. However, I can imagine that might be a tough sell to publishers because, let’s face it, they are capturing a lot of consumer surplus from gamers who buy a game and never end up playing it much or at all.

Finally, though, OnLive does more than just enable streaming gaming. It adds a social dimension to gameplay that PSN/XboxLive haven’t fully been able to do. One example is brag clips (highlighted in the video above). I look forward to checking out the OnLive SDK to see what types of functionality game developers may be able to leverage.

To signup for a free 3-month launch membership to OnLive, signup here.  Whether OnLive is a success or not, the OnLive service represents a major milestone for all media.  OnLive proves that even the most complex media experiences (HD online multiplayer video gaming) can be delivered via a stream.  We all knew that the the disc was dead.  Well now it looks like the download is also dead.

Written by Rishi Khaitan

March 13th, 2010 at 8:46 pm

Tagged with ,

Lifehack: Pandora on the PS3!

without comments

Okay that pic doesn’t look like much so let me explain.  Last night had some friends over and for living room music, instead of clumsily trying to connect my iPhone or streaming music from my laptop, I thought hmm let’s try Pandora on the PS3′s web browser.  Turns it out works flawlessly.  Quick tip though, in the browser menu, under the View option, turn the Resolution setting down to -2.  That will essentially zoom in on the Pandora flash UI to make it much more legible on your TV.

As usual Pandora served up some awesome music and it was fun to glance over and look at the track info.  Pandora would do well to make a PS3 specific UI.  Maybe even distribute it in the form of a cheap PSN download game?  That would enable Pandora to build a UI tailored to the PS3 controller.  Now that would be sweet…

Written by Rishi Khaitan

March 13th, 2010 at 3:19 pm

Tagged with , , ,

Facebook Ads: “Likes and Interests”, Suggest Tool?

without comments

For those of you that aren’t familiar, Facebook’s Ads platform allows advertisers to target ads towards users whose profiles include certain keywords.  Today, AllFacebook is reporting that facebook will now refer to this keyword targeting not as “keyword” targeting but as “likes and interests”.  While a subtle change, this, from a marketing perspective, may actually be pretty brilliant.  “Target ads to users who like the Boston Red Sox” sounds a lot sexier than “Target ads to users who have the words boston red sox somewhere in their profile information”.  Moreover, the “likes and interests” language is clearly distinct from the “keywords” language that just about every other ad platform on the Internet uses.  Furthermore, it’s a language that Google and the other top ad networks can’t speak that fluently.

However, what is most intriguing to me is if and when Facebook will be launching a likes and interests Suggest tool similar to that of Google’s Keyword Suggest Tool.  Assuming Facebook is tracking the necessary ad click data+metadata (which I’m all but sure they are given that they seem to track everything), such a tool would be a boon to all advertisers, but especially those (majority) advertisers who are unfamiliar with social advertising.

Written by Rishi Khaitan

March 8th, 2010 at 3:35 am

Tagged with ,

Great Ideas

without comments

I firmly believe that a person will have at most one truly great idea in a year.  People who expose themselves to a wide range of information and experiences will probably be closer to 1 and people that don’t.. well, they’ll be closer to 0.  But how much are these “great” ideas really worth?

I argue that ideas, even the best of them (assuming the greatness of an idea is actually measurable), are worth essentially nothing.  Ideas may in fact lead to invention or innovation, but the idea in and of itself is simply a starting point.  The value is in the execution that comes after that starting point.  And the reality is that startups — whether a high-tech venture or a restaurant — fail because they fail to execute in one way or another.  Furthermore, startups that are successful are very often, at best, tangentially related to the original t=0 idea.  All this is not to say that there aren’t ideas that are better than others, but smart execution is easily an order of magnitude more important.

What is interesting is that empirical research shows that there is definitely an idea premium when startup founders allocate founder equity amongst themselves.  In other words, the person with the idea often receives a bit more equity than the other founding team members.  Moreover, that idea person is the CEO of the startup 56% of the time, but CTO only 12% of the time.  One might quickly conclude from this that the market does value ideas.  My hunch, however, is that because the idea person was, by definition, the first to start developing the idea and building the team, he has often put in the most sweat equity and the extra equity is not compensation for the idea itself.

So am I arguing then that that one great idea you have each year should just be quickly dismissed as having no value?  No, absolutely not!  The ideas that I have actually pursued are those that keep nagging me.  But the thing to remember is that even if you vet the idea and it still has you excited, remember again though that it is the execution that will actually make your idea worth even $1.  If execution is all that matters, the important question then is: what does the right team look like to execute on the idea and would you belong on that team?

Written by Rishi Khaitan

March 8th, 2010 at 2:58 am