Now I get a chance to see what it is like on the other side of the table, or class room.  I am currently a Teaching Assistant for MIT’s introduction to circuits class, fondly named 6.002.

Better the Second Time Around

As I am sitting in on the class and listening to the material a second time, it dawns on me that these lectures are quite insightful.  Only after having “learned” this material before can I actually appreciate some of more subtle things in lecture.  

Incidentally, by “learn” I mean have gone through the Jedi trials of doing the home works, designing/building real life circuits, and then questioning all the basic assumptions in an effort to push the state of the art in my research.  Despite what the previous sentence my imply, I am actually not trying to boast that I know circuits.  The opposite is true I now know that I don’t know circuits…

Drink from a Fire Hose

I remember as an undergrad I would just sit in class and try to absorb as much information as possible.  However at MIT they have a saying that classes are like drinking from a fire hose.  So it was more like trying not to drown in information rather then absorbing.

Professors Do Care

I guess as a bitter grad student I was under the impression that most profs only care about research results.  I am pleasantly surprised that the professors actually do care about the quality of the lectures.  They aren’t merely just “phoning it in”.

I mean you have these four brilliant profs who has spent more time pondering circuits than most people ponder anything (me included) and I feel that it is a bit of a waste for their insights to be lost.   …So I decided to jot them down.  If you want to take a look just follow the link below. 

Please bear in mind…these are graceful and brilliant insights filtered down and passed through a lowly TA…so I would use these notes to judge the quality of the profs or MIT.  :)

Noteworthy from MIT’s Intro to Circuits Class 6.002

“Crowds are smart and most of the people in a crowd are good and are willing to help. Airline security agencies can harness this wisdom and this willingness to help by first acknowledging that it exists. Then implementing a few simple policy changes.

1. Give passengers a way to quickly and discreetly report suspicious activity to officials, both before boarding and during the flight. A button that simply reports suspicious activity within a 3 passenger radius may be all that’s needed.
2. Give crew a way to constantly monitor an aggregation of these reports in real-time and develop response policies depending on the severity or intensity of the reports about an individual.
3. Encourage passengers to engage their fellow passengers in conversation, but not in the context of preventing terrorism. For example: Include a simple statement during the pre-flight briefing of: “Northwest airlines passengers are known for their friendliness in the skies and many of us traveling to and from far off destinations. We encourage our passengers to compare travel notes and to get to know each other….etc..etc…”

With these simple changes, you have now created an on-board distributed intelligence gathering network that is reporting back to a central authority in real-time.”

Crowd Sourcing or Mob Panic

Great idea! (sarcastically) Hey, maybe we can take it a step further and implement a anti-terrorism prediction market. Oh wait, the Pentagon tried that already…

Do You Understand Crowd Sourcing?

1. It’s not crowd sourcing!  The sample size is too small. In your scenario, you are only really going to get about 3-4 people watching any one person.  In other words, the people directly sitting next to them.  Even if you assume the whole plan is watching, several hundred people is still too small a sample.
2. Inherent positive feedback loop. As soon as someone starts getting attention they will continue to get more attention.
3. People are notoriously bad at judging things they have no experience with. What may be perfectly normal behavior for one person is counted as suspicious to another.  This fact mainly comes from their social, economic, and cultural background.  For example, one person may view wearing a mask to cover your face as untrustworthy another may interpret it as being modest.

How It Will Play Out

Let’s say I am non-white and my seat neighbor tries to strike up a conversation and I seem strange.  This maybe due to motion sickness, language barriers, I think they are flirting with me, or maybe I just don’t want to talk.  Next thing I know, someone hits a button and now more random people are trying to talk to me, who I also try to ignore.  Which further confirms that I am different.  Quickly after that, I’m being lead to the bathroom for a strip search.

This mob panic scheme will lead to a bunch of random people being picked out because they seem strange to the people next to them. More importantly, it will not lead to safer skies, just more panic.

A BETTER SOLUTION?

Possibly have more flight attendants serving less people.  Train the flight attendants to identify warning signs without drawing any particular attention. Trained flight attendants are much better at judging what is cause for concern because they fly with people everyday.  Of course, this leads to another problem: what to do once a possible threat is identified.  I would suggest shifting them to a different seat for closer monitoring.

Of course, all these things cost money.  Having a flight attendant/cop onboard and empty seats is probably going to increase ticket prices.

Additional Note – Crowd sourcing works for detecting spam and porn mainly because they both have a “I know it when I see it” characteristic.

Mind your viral coefficient!!

The number of people an average user can “infect” is the viral coefficient, which is the only thing that matters.  Simply, if each recent customer can bring in at least one new customer then your product is viral.  In other words, you want a coefficient greater than 1.

The viral coefficient is solely dependent on the product and customer. No amount of marketing can change this.

What makes a product infectious?

1. Give users a product they can talk about.
2. Make it easy for prospective users to listen.

Last summer I met a team that was working on a new iPhone music app.  I thought that it was absolutely great, but unfortunately I didn’t have an iPhone. I recommended it to my friend who I thought would love it.  She was their ideal customer. She loved karaoke and facebook, plus she had a ton of friends that did the same.  After a bit of prodding, she bought the app, so I performed my +1 viral duties.   Unfortunately, she played with it a couple of times found it underwhelming and confusing.  This promptly killed the viral chain.  Once this happens I am guessing there won’t be a second chance..

Essential Characteristics for Viral Growth

Product Must Be Remarkable!

Greatness means different things for different users.  In the example above, I found the technology amazing and by that alone I was willing to rave about it.  On the other hand, my friend could care less about how it works.  She was more concerned about the user experience, which was not great.

Don’t get me wrong. I am a firm believer of releasing software products early and often but until you are 90% complete, your product is not ready for viral growth.  I guess, that means you shouldn’t be disappointed if your initial public version doesn’t go viral immediately.

Know Your Market

Are you the key master?  I am the gate-keeper (sorry Ghost Busters reference) to my personal social network.   Even if I do mention it in passing to several friends, I only follow-up with the people who I think will get the most benefit.

As such, your product must either have:   (having both is the best)

1. Super Large Potential Market
   If your product has mass appeal then there is a greater likelihood I can find someone to tell. These are usually consumer facing products.  To give a reference about what is mass appeal, there are 600 Million people in the US and about 9 Million or them have an iPhone
2. Built in Community
   If you are in a niche market you need to have a built-in community.  Let’s say you have a super niche product that is useful to only people in long distance relationships between the USA and Canada.  Yes…it is super niche.  Even if I love the product, I pretty much have no one to tell.  B2B (Business to Business) products can be thought of a niche products.

“Power of Weak Ties” (term from Tipping Point)

A community can be thought of as a formal way to foster weak ties.  In terms of sociology, weak ties are acquaintances, people you know, or people you hang out with.  In contrast, strong ties are people you whom you share an emotional connection, such as family, close friends, and lovers.

Research has found that new and different information tends to travel through weak ties.  The people in a close network already have the same information.  Therefore, new information only comes from connections beyond the close network.  More so, people tend to have more weak ties than close ties, which adds to the diversity.

Weak ties are essential to viral growth.  Therefore, having a product that is conducive to being spread through acquaintances  is more important than one that spreads well through close networks. But that is not to say that close networks aren’t important.

This seems counter-intuitive but consider Facebook.  Facebook’s great for staying in contact with my friends but not close friends.  In fact most of my close friends aren’t my Facebook friends.  I do not need to see their latest party pics because I was probably there.  I don’t need to read on their wall that they broke up with their boyfriend because they probably called me already.  Weak ties get friend requests.  Strong ties get phone calls and text messages.

Consider another example.  Let’s say I put up a respectful and serious website devoted to my late grandmother.  If there is nothing unintentionally funny about it, the site really serves no benefit to anyone outside her strong ties and won’t be very viral.

Network Effect

Network Effect is probably one the most important aspects of a product to make it viral.   As such, I plan to devote an  entire post on this subject.

Viral marketing is the holy grail for companies, especially small startups.  It is like getting something for almost nothing.  The proverbial Free Lunch if you will.

From the marketing experts I spoke to and various things I have read, Tipping Point and Influencer Marketing, it seems that all I have to do is get a few Trend Setters/Social Connectors/Influencers on board and my product will be more viral than H1N1.

However, after digging deeper, I realized that it’s not quite so easy.  Contrary to popular belief, these elusive and elite Trend Setters/Social Connectors/Influencers have no influence on viral growth.

I guess my dad was right!!  There is no free lunch.

Before you spend time and money to court the elite influencers, such as major blogs, I would like to share some of my findings about viral marketing and growth.

What is Viral Growth?

Aside from being like a virus, viral growth is represented in math as a geometric series that goes to infinity.  In other words, it looks like a hockey stick, which most entrepreneurs know all too well.

picture I got from http://mac.honan.net/

How Did Marketers Get It Wrong?

Lets start by looking at the analogy.

Word-of-mouth or viral marketing spreads from person to person like a virus.  More so,  if each customer spreads the product to multiple new customer, then we can get epidemic-like growth.  All that is needed is to find an initial customer (or customers) who has contact with a large numbers of prospective users.

Unfortunately, the analogy itself is wrong.  As illustrated by David Skok, only recent customers find new users.  This fact becomes obvious when we consider the nature of personal social networks.  Essentially, once we go through our friends, family, and acquaintances we run out of people to “infect”.

Consequently, this diminishes the effect of the elite Influencers because the next wave of users will come from the average customer alone.  In fact, the recommendations of the average customer will determine if a product will go viral.

To prove this to myself I built a little model in Excel based on the work of David Skok.  The graph below shows that no amount of influencing can help an inherently non-viral product.  

The virility of a product is determined by its viral coefficient.  This coefficient is defined by how many new users a recent user can bring in.  If each recent user can bring at least one new users then the product is viral.

The graphs shows a product with a viral coefficient of 0.9 for the average user.  The top graph, red, shows the product growth with an Influencer that knows 10 Million (yes 1e7) times more people than the average person.  This represents an astounding 1 Million initial viral coefficient.  Despite the initial growth, we do not see a hockey stick curve and it quickly saturates.

Although, one may argue who needs viral growth if you can initially get over 90 Million (and grow to 90 Billion) users from elite Influencers.  That is true, but its going to be an expensive lunch.  Influencers that can deliver 150 Million prospective users usually would cost a considerable amount of money.  To give you an idea, the Super Bowl, which draws 95 Million viewers (not necessarily prospective users), costs $3 Million to place a 30-second ad on the show.  Therefore the large initial viral constant used in the example above is not achievable for most startups.  The graph illustrates that no matter how big the influencer they cannot make a product grow virally.

Who cares! Growth is Growth..right?

Yes, in that growth implies that things are going up.  But at the same time absolutely not!!  Viral growth refers to something very specific, which has a definition based in math.  Essentially, it means that you grow at an increasing rate.  This is more than just an academic difference.  In the real world, it implies a way to achieve an enormous amount of growth in a relatively short period of time and with limited resources.  Virialness can make or break a company.

Therefore, unless you land a free superbowl Ad, the shape of the growth curve is more important than the absolute value.  Once you achieve a capital efficient strategy for viral growth, scaling the rate is easier.  At that time, one might say,  “the product sells itself”.

More to come:

• How to increase viral coefficient?
• Network effects on product growth?
• When influencers matter?
• What is the model I am using?

Until we figure something out, we are just going to use it as a blog for our own personal musings and thoughts on entrepreneurship,  mobile,  EventHive (our current statup), and other random things.