Monday, February 24, 2014

Do we know what a Product Manager does?

As a consensus, we don't. From Wikipedia:
A product manager investigates, selects, and develops products for an organization, performing the activities of product management.
Quite an oxymoron, although the Wikipedia entry concedes:
The product manager title is often used in many ways to describe drastically different duties and responsibilities. Even within the high-tech industry where product management is better defined, the product manager's job description varies widely among companies. This is due to tradition and intuitive interpretations by different individuals.
The role of a  Product Manager was first invented in 1931 Proctor and Gamble by Neil H. McElroy, a Harvard graduate and President of P and G. He noticed that selling a new soap named Camay was in competition more with another brand of P and G, the Ivory soap.

He is credited with the invention of Brand Management and nominated a Brand  Manager for each brand and later for each product marketed within the brand, a product manager.

This simple invention skyrocketed McElroy career and he become the Secretary of Defense, reorganizing the armed forces under President Eisenhower.

As stated in the P and G archives:

Brand management as a business technique was one of the signal innovations in American marketing during the twentieth century. It epitomized the persistent theme of balancing centralized oversight with decentralized decision making based on who in the company had the best information about the decision at hand.

High Tech Product Management

This is a different perspective than corporate PMs, where actually the PM name was invented to begin with as part of a bureaucratic hierarchy. Like P and G, like Oracle today, or Sun Micro yesterday PMs were typical for  large or extremely large organizations, not run by entrepreneurs, but by hired top schools executives. They have generous budgets for priority products.

In contrasts to these multinationals, a high-tech startup is small enough. The" persistent theme of balancing centralized oversight with decentralized decision making based on who in the company had the best information about the decision at hand" is NOT an issue. There is no gap between centralized and decentralized decision making, Startup are flat organizations and often maintain this agility as they grow up

Most of the times, the founders themselves are the initial and most important Product Managers of a start up. The bootstrap success sits on their shoulders, and by comparison to the multinationals inventors of the Brand Management, have very small budgets

As shown in the illustration, entrepreneur's skill is to start  with borrowed or investor venture money a negative cash flow free fall, reverse it and bring it to break even (stages 1 and 2). After the break even, the hired managers can take over
From Entrepreneur-ese versus MBA-ese

The product world according to Des Traynor

Startups have a different breed of Product Managers. Last week at Product Strategy for Start-ups with Des Traynor + Unconference I saw a memorable Product Management Strategy presentation by the co-founder of intercom.io

Des is an Irishman, without the heavy accent and a Dubliner, like James Joyce.  His slide shows are entertaining and many slides are fascinating. So  to get the essence we need to deconstruct.

You must have a Vision and not dabble in the dark. Are you marketing? Then everything shifts to the personal, clicks are less important than conversions, data and numbers are key and ads are a conversation starters.

Vision = Your domain + How software looks like in your domain on today's technology everyone uses

Des is obviously referring to his own company. His vision theory come from his experience
Similar with P. and G. Brand Manager from the 30's, he wants to make people want what we make. Here is a list of wisdom sayings:
  • Customers rarely buy what the company thinks it sells
  • Customers don't buy categories
  • Customers are not demographics when they choose to buy
  • Don't confuse correlation with causation
  • so on
One does not have a customer until she gives us a check.

Des Traynor is now in a circuit giving this presentation, so I am going to stop here.

intercom user cases

Have a look at this page . Everything Des teaches, intercom.io delivers

  • Understand how your product is being used. 
  • Increase conversions by automatically reaching out to users after they sign up
  • Treat customers like humans
The last bullet seems obvious, yet reality proves otherwise. For these reasons, intercom sells what it preaches.

Wednesday, February 19, 2014

From Big Data to Big Decisions

In my blog entry from February 12, I quoted the opinion of Larry Fink,  the CEO and Chairman of the largest asset management firm in the word, Blackrock (three and half trillions dollars under their administration)
  • The markets in America reflect a lot of fear. We are bombarded with information that we can not make sense out of it. We can't decide whether this  Big Data information  is good or is bad.
  • As a result every CEO, government decision maker handling money only invests short term.  Some CEO's buy back shares to invest the cash in government bonds at 2% interest rate. Companies sit on lots of cash, too afraid to spend it. 
  • Banks have tons of money, they sit in the vaults without making loans. The banks want to make loans, but say someone who qualified for a loan in 2007, does no longer qualify. The companies who qualify, they do not need any loans as they have tons of cash already.
The main inhibitor, it is worth repeating, is not Big Data itself. It is the fear of it. The top decision makers in our society, the CEOs and the Government economists, are unsure how this big data can be used.

As seen at Strata 2014, people are talking about Predictive Analysis, creating the popular confusion that now, with lots of data, and fancy statistical and analysis, we can predict the future from past data. This is not true! See what Benoit Mandelstam , the creator of fractals, clearly states.

So what can we do to make Big Data useful for our top decision makers who have the power to create jobs and economic growth? How can we win over Larry Fink and his clients to make longer term investment using Big Data input? How do we transform their skepticism in belief?

The partial answer to these questions I saw  in Matei Zaharia from Databricks keynote at Strata 2014 conference two weeks ago.

Here is the story.  Commodity clusters like those Hadoop based has made up to 20x cheaper to store datasets


By 2020 we will have 44 ZB (one zettabyte equals one billion terabytes), but the cost of storing this enormous data will be about the same as in  2013.  What we can do with all this data? We can do two things:

  1. Define speed as "How quickly we can go from data to decisions?"
  2. Define sophistication as "Can you run the best algorithms on the data?" These are NOT off-the-shelf-algorithms.
Bullet #1 drew my attention. The purpose of any analysis of big data is not to deliver solutions. The purpose of of all  big data is to convince the top decision makers to take decisions.

The future can not be predicted, but human can make informed decisions with minimized risks. Risks can never be eliminated. The technology is just an enabler for humans.

To use some metaphors, the world information is reported by the media. Wall Street Journal and New York Times, The Economist, etc are extremely reputable that the information is real, and inaccuracies are rare. The National Enquirer is another publication that no one or - very few people, perhaps from tabloid media business -  will use to make investment decisions.

Even when reading Wall Street Journal, the decision makers have different criteria. George Soros, for example, in one interview said that each morning, he reads the papers to discover an error humans always do. We are not perfect. But such mistakes can make hundreds of millions for Soros, an investor who bets on stocks and currencies going down, not up.

Big Data is information that each one of us assesses whether it is reliable and usable in different ways.

The best tools in Big Data then, are the ones that generate decisions. Lutz Finger in his article STRATA 2014 - Data Demystified quotes James Burke:
Information is causing change… if it is not causing change, it is no information.
If this is true, then we may have some sort of metrics on how to judge any big data tool. Something based on how popular and effective the tool is for decision makers.

Media, newspapers, books, movies have circulation numbers, Nielsen ratings, revenues generated.

Why not Big Data? Sure we heard the women sleep more than man, or customers in a bar ordering Corona beer spend more money.

But this somewhat useful information is trivia in comparison to Blackrock. If Larry Fink can use Big Data to revitalize the economy, and his success will propagate virally worldwide this will be by far the greatest breakthrough Big Data achieves.

Friday, February 14, 2014

Tweets from Strata Conference 2014

Here is the power of Twitter . It saves saying more. My impressions from Strata Conference 2014.











Kudos to  @annalisclint and @kathykmy

 


Wednesday, February 12, 2014

Big Data: How It Affects Our Livelihood

Charlie Rose interviewed last night Larry Fink, the Chairman and CEO of Blackrock, the largest asset management firm in the world which administrates 3.5 trillions dollars.Yes, trillions.

Larry Fink, CEO and Chairman of Blackrock and one of the great minds from Wall Street

Mr. Fink, an UCLA alumni, is one of the most lucid voices I have ever heard.  I share with you what  I humbly learned.
Big Companies executives fear Big Data more than cyber-attacks
  • The markets in America reflect a lot of fear. We are bombarded with information that we can not make sense out of it. We can't decide whether this  Big Data information  is good or is bad.
  • As a result every CEO, government decision maker handling money only invests short term.  Some CEO's buy back shares to invest the cash in government bonds at 2% interest rate. Companies sit on lots of cash, too afraid to spend it. 
  • Banks have tons of money, they sit in the vaults without making loans. The banks want to make loans, but say someone who qualified for a loan in 2007, does no longer qualify. The companies who qualify, they do not need any loans as they have tons of cash already.
  • There is no growth. There are no jobs, companies re-organize, because by saving money they can live with smaller revenues and still show a profit. As there is no growth, the government deficit increases because we pay no taxes on the growth that did not happen.
  • This is vicious circle that we in North America must break, Mr. Fink says we are already energy self-sufficient in oil and gas, if we combine how technology dramatically increased the output in Mexico, US and Canada.

Mr Fink talked about many other topics, but I will stop here.

Based on his diagnostic of the situation - and he is the most qualified in the world to unlock where the illness is hiding - we must eliminate the fear of Big Data. Here are some quotes:
Big data is NOT about watching employees if they perform or not. or a Big Brother terrorizing  tool  This is a waste of time and a narrow vision.

What Big Data ,Machine Learning, data scientists and mathematicians must reach is a sufficient transparency, that will give any CEO and government official the confidence to make a decision.Big data is for the CEO  like a pair of prescription glasses just invented, empowering them to really see how make big money longer term,

I looked at Google Flu Trends Can I know whether I will get the flu or not, if I have had a flu shot? Do I know the probability in my area? Depending on what I need, this ingenious tool does not work of the shelf.

The way we process, promote and offer big data today is not useful enough. We are drown in oceans of big data,. We need to sail on top of it..

Tuesday, February 11, 2014

Netflix: The Gift of Prophecy

On September 11, 2011 - just as Netflix shares dropped from $250 to $80 in two months - I wrote my blog post What Happens with Netflix? I said
Success is not rooted in the technology. Success is rooted in the people desires to pay for entertainment delivered in their homes. Once the streaming technology has been developed, and here Netflix has a big merit, it becomes like generic drug. Anybody can use it at lower and lower costs. It is no longer a differentiator. The content and pricing of the content are the differentiators now.
I know my inspiration come from above, because I am human. I like to point this out as many people on the Valley seem to have forgotten this. What I like about Netflix, is that they are different.
As Netflix likes to quote, there is no certitude, there are only opportunities Netflix will react. I hope with humility, as without humility, one loses the gift of prophecy
If we have the The House of Cards, season 2 starting  February 14, is because the Netflix founder, Reed Hastings who was humbled by that September 2011 stock free fall. He hired Kelly Bennett as CMO from Warner Bros in 2012.

Neil Hunt, Chief Product Officer changed strategy. After developing their streaming technology, they de-emphasized Netflix as leading engineering company.

This is what Fortune magazine wrote in January 2013:
 It's only logical that Netflix will lose its dominance over the video-streaming business. It's been slowly happening for a couple of years. Competitors have been piling in. Netflix is losing some big providers of movies and TV and paying more to the ones it lands or retains. That situation is likely to continue as more licenses come up for renewal.
Contradicting Fortune prediction, Netflix is now a leading original content provider, because this what their paying customers want. They don't care about the technology. The stock price of Netflix is today over $430.

Monday, February 10, 2014

What is happening with Cloudera?

In March 2009, about five years ago, New York Times published this photo
Christophe Bisciglia, Amr Awadallah, Jeff Hammerbacher and Mike Olson
 started their company, Cloudera, around Hadoop.
Mr. Hammerbacher - ex Facebook and Mr. Bisciglia - ex Google were joined by Amr Awadallah,  a former Yahoo engineer, and Michael Olson. Mike Olson is the one who sold his SleepyCat database to Oracle, has some money and good reputation.

They were the media darlings of the moment. They were to stir a revolution:
By mapping information spread across thousands of cheap computers and by creating an easier means for writing analytical queries, engineers no longer have to solve a grand computer science challenge every time they want to dig into data. Instead, they simply ask a question.
Simply asking a question  as New York Times puts it, was a bit of simplification. Cloudera and Hadoop are not for everyone. The first thing a customer must do is to go to Cloudera University, graduate and then , maybe, she starts using Hadoop. Maybe because people don't buy on features.

They buy on sentiment, feelings. "The instinct cannot be expressed in terms of intelligence, nor, consequently, can it be analyzed" said a famous French philosopher.

Competition started cropping up from everywhere, because Map Reduce and Hadoop are open source

According to Scott Denne, M&A analyst with 451 Research, in Quora here is a list


Today GigaOm reports that
Splice Machine, a startup promising a SQL-on-Hadoop database that can handle both transactional and analytic workloads, has closed a $15 million series B round of venture capital from InterWest Partners, along with Mohr Davidow Ventures. 
Cloudera raised a series E capital infusion of $65M in December 2012 (14 month ago) that nearly doubled their funding  to $141 millions.

But didn't they loose their soul, a little bit? I played by scribbling a piece of paper and I came out with this:

Miha's scribblings
The base box originally included the founders, I still see the original 2009 photo  from NYT, everyone looking like young Beatles. But looking  at Cloudera management page I see a lot execs, fifteen people managing this company. The five years of expected success plus the tons of money they got from venture firms, elevated their name to one of the High Priest temples on Silicon Valley. Ph.D tries to recruit another Ph.D's. a Stanford graduate another Stanford graduate, transforming all other people in second rate candidates with a second rate know how.

Cloudera has become, in my humble opinion, an ivory tower. I wonder how many times someone studies what a Joe IT worker says and what he feels when a Cloudera salesperson calls to offer him Hadoop.

John Gourville from Harvard Business School says "Many innovations fail because consumers irrationally overvalue the old, while companies irrationaly overvalue the new"

Even if "the benefits of using a new product are clear and substantial", if the use of this products require a high degree of behavior change, they are doomed to fail.

The above quotes are from How to make products with a happy ending and from here, where I refer to Fogg's Behavior Model. It shows  a behavior, any behavior, can change. A decision to loose weight or buying a Hadoop product has the same components.

As Mike Olson sold his previous company to Oracle, we saw consistent rumors in the media that Larry Ellison will acquire Cloudera. However Oracle culture does mix with Hadoop culture like oil and water. The litmus test for Cloudera is to deliver products and services Oracle sales people can handle, without attending Cloudera University.

As time passes by, each day Cloudera value for Oracle decreases. With so many alternatives for Hadoop companies, it is a buyer market.


Friday, February 07, 2014

Dr. David Ungar about “anti-lock,” “race-and-repair,” or “end-to-end nondeterministic” computing

Two of the blog entries here had 15,000 unique visitors
  1. Many Core processors: Everything You Know (about Parallel Programming) Is Wrong!
  2. An Interview with David Ungar, IBM Research
The most startling discovery of Dr. Ungar is the postulation of a new type of computing in many core processors (and by analogy in distributed computing). He calls it
  “anti-lock,” “race-and-repair,” or “end-to-end nondeterministic” computing....When we give up synchronization, we of necessity give up determinism. There seems to be a fundamental tradeoff between determinism and performance.The obstacle we shall have to overcome, if we are to successfully program manycore systems, is our cherished assumption that we write programs that always get the exactly right answers.
Here is the video with David's  original talk at Splash-2011  Everything You Know (About Parallel Programming) Is Wrong!: A Wild Screed About the Future 

Warning: this is a 66 minutes video, but it is worth seeing as one day will part of the history of Computer Science.  Benoit Mandelstam first book on Fractals sold only 19 copies. Real great theoretical discoveries are not popular when they emerge first time


This is the reason I placed David's video here, for convenience


Thursday, February 06, 2014

The Mathematics of Miracles

What is a miracle?

The word נֵס (nes),   נִסִּים (nee-SEEM) - plural of נס - is a miracle in Hebrew.

For observant people, a "nes" as an action, from G-d, that is supernatural. The word "nes" also means the sail of a boat. The sail enables the boat to move along, to travel in a path.  One can see that there is a force controlling which way the boat is traveling. When a person experiences a miracle, it becomes clear that there is a force controlling the direction of his life. Quoting Chabad web site resources:
What we refer to as nature is actually miraculous and “unnatural.” It is only because “natural” events happen all the time that we take them for granted.
In the words of the Talmud, “The one to whom the miracle is happening, does not recognize the miracle.”  Extraordinary miracles wake us up to the fact that all of life, down to the minute details, is one big miracle.

How do miracles work?

Maimonides, in his Guide for the Perplexed, (his philosophy treatise) writes that all supernatural events were “programmed” into the world at the time of creation. Thus, the Talmud says that G‑d “made a condition with creation” that when the Jews would arrive at the Red Sea on their way from Exodus from Egypt, it would split. 

Others see open miracles as G‑d “stepping in” and shattering the law of nature to change and defy it.

The latest and most unexpected explanation for miracles comes from David J. Hand, - a mathematician  and renowned statistician  - the author of a book "The Improbability Principle: Why Coincidences, Miracles and Rare Events Happen Every Day" . You can  pre-order the book from Amazon.

David J. Hand argues that extraordinarily rare events are anything but. In fact, they are commonplace. We should all expect to experience a miracle roughly once every month.

The Improbability Principle breaks this circular reasoning against
miracles,dispelling the myths that miracles are false and do not occur

Never Say Never

This is the title of an article written by David Hand for Scientific American on January 21, 2014

He defined a set of mathematical laws, - The Improbability Principle - that tells us we should not be surprised by coincidences.

The birthday problem

For example, how many people must be in a room to make it more likely that two people have exactly the same birthday?. More likely means there is 51% probability. At first , trying to answer the questions, one thinks of herself. How many people have the same birthday as I? But the question involves any two people who are in the room.

The answer is surprisingly low, 23. Only 23? Yes, only 23 and this completely unexpected for the majority of people

First we calculate the probability the none, zero people have the same birthday like me. For  1 person the probability  is 364/365 If there are "n" people in room, there are "(n-1)" people who must have a different birthday than mine. For 23 people, the probability that no one has the same birthday as me is ([364/365) *(23-1)]. The result is 0.94.

If there is a 94% probability that none share my birthday, then there is a 6% probability that someone is born on the same day and month as me.

This is very small and this is the wrong answer. We want to know if any pair among the 23 people in the room share the same birthday. We have in the room n*(n-1)/2 pairs of people. For 23 people we have 253 pair. The total probability that none of the pair share the same birthday will be:

(364/365)*(363/365)*(362/365).....*(343/365) =0.49

As the probability that none of the people in the room share the same birthday is 49%, then the probability that one pair shares the same birthday is (100%-49%) = 51%

Isn't this elegant?

Calculated miracles no one thought possible

On September 6 and September 10, 2009, four days apart, the Bulgarian lottery had exactly the same winning numbers; 4, 15, 23, 24, 35, 42 . The authorities ordered a fraud investigation and they found nothing. The probability for this event to happen, as calculated with the Improbability Principle is 1 in 13,983,816.  

Compared to nature,  - a 1  in 14 millions for the same consecutive winning numbers lottery - is a high probability. The probability for one sperm to fertilize the egg in humans is 1 in 280 millions, 20 times lower. Every person alive today, including of course everyone reading this blog, is the winner of a seemingly impossible to win lottery.

We are all a miracle and we do not know it

The World Wide Web

There are 2.5 billion users on WWW. This means there are  3*1018   pairs of people and 10750,000,000  possible groups of interacting members

Now this number is big, but quoting David Hand, this is good thing.
Even events with very small probabilities become almost certain if you give them so many opportunities to happen

Statistical mathematicians versus Data Scientists

Data Scientists

In  What a Data Scientist Does?  I wrote:
There is a big debate about who the Data Scientists are. Are they statisticians or are they computer science?
In another blog entry, GigaOm founder Om Malik says
The problem with data is that the way it is used today, it lacks empathy and emotion. Data is used like a blunt instrument..
The idea of combining data, emotion and empathy as part of a narrative is something every company — old, new, young and mature — has to internalize. If they don’t, they will find themselves on the wrong side of history.
The computer scientists in business  may use advanced the statistical methods where the math as a shield  tactic is used so "fewer critics are available to be properly skeptical." Their ideas of emotion is executing orders, in which they prove whatever the boss wants them prove.

The Improbability Principle 

The Improbability Principle has an elegance, depth and powerful metaphors to discover the world around us. The maths are simple to follow. Mathematicians don't analyze just data, that analyze the world we are in and, there a spirituality in their work. Their mere existence is נֵס (nes) by itself and it has a very profound human touch. From torah.org
The letter "nun,"  represents downfall, suffering, and misfortune. The letter "samech," which in the alphabet, and in the word "nes," follows "nun," represents uplifting, salvation, and redemption. A miracle, a "nes," is the combination of these two elements: we are faced with trials and tribulations, and our situation is perilous. Yet, through divine providence, a supernatural occurrence rescues us and provides us with salvation. The word "nes" is a reminder of the ups and downs in life. 

Monday, February 03, 2014

November 19, 2001. Defiant Sun fighting the wrong enemy

Cover page Businessweek November 19, 2001
This is the cover of Businessweek from November 19, 2001, thirteen years ago from today.

BusinessWeek suffered a decline during the late-2000s recession as advertising revenues fell one-third by the start of 2009 and the magazine's circulation fell to 936,000.  In late 2009, Bloomberg L.P. bought the magazine—for a reported $2 million to $5 million plus assumption of liabilities—and renamed it Bloomberg Businessweek

Sun Microsystems On January 27, 2010, Sun was acquired by Oracle Corporation for US$7.4 billion, based on an agreement signed on April 20, 2009. The following month, Sun Microsystems, Inc. was merged with Oracle USA, Inc. to become Oracle America, Inc.

Microsoft On July 20, 2012, Microsoft posted its first quarterly loss ever, despite earning record revenues for the quarter and fiscal year, with a net loss of $492 million due to a writedown related to the advertising company aQuantive, which had been acquired for $6.2 billion back in 2007.

Interesting to note that Microsoft write down derives from aQuantive's acquisition where they paid a price nearly as high as the price Oracle paid for Sun.

As of January 2014, Microsoft's market capitalization stands at $314B, making it the 8th largest company in the world by market capitalization

Microsoft was not the enemy. Microsoft was the big brother of Sun.

Google market cap today is $395B, larger than Microsoft. In 1991, they were a 3 year old private company, no one took seriously, except the funding in August 1998  of $100,000 from Andy Bechtolsheim, co-founder of Sun Microsystems, given  before Google was incorporated in 2004.

Facebook market cap today is $154B.  Mark Zuckerberg was 16 year old  and still in high school in November 2001. He founded FaceBook in 2004.

No more comments, ladies and gentlemen Lets just quote our friend Reuven Cohen  who quotes in turn  Allen Kay in Forbes Magazine:
“Don’t worry about what anybody else is going to do… The best way to predict the future is to invent it. Really smart people with reasonable funding can do just about anything."

Blog Archive

About Me

My photo

AI and ML for Conversational Economy