Sunday, November 22, 2015

Some Valuable Learnings from Amazon Fire Phone failure


This is more of a recent failure. For more reasons than one, i was keen to explore this case further to figure out the possible reasons why Fire Phone failed (Significant number of Fire Phone remained unsold) to cause an impact. Amazon, under Jeff Bezos, have really made some big, bold bets in the past and atleast from outside, it appears to have a culture in which failing is not treated as a sin. This case is even more interesting considering that Amazon has had a reasonable run of success with Amazon Fire Tab. Agreed, a phone is a different animal than a tab when it comes to nuances related to its development, usage and adoption but it (and Amazon's past success with Kindle and its variants) does tell that Amazon was no novice in the area of building hardware product. Read on to know more-

Customer focus gone wrong:

If i try and think about why an organization like Amazon would want to venture into building a phone, i can certainly think of Amazon want to own the starting point of customer experience to access its online retail services. Apple, for instance, takes 30 percent of all revenue generated through apps, and 70 percent goes to the app's publisher. It means that even though Amazon may be selling e-books using Kindle app in iOS, it won't get its full share. Controlling the starting point of buying experience (i.e. owing a smart phone) can have other benefits too where Amazon could have had its own services that could have helped them sell more stuff. So, venturing into phone wasn't a bad idea at the first place. Some analysis that led into analyzing the failure of Fire Phone also led to the fact that it lost out on customer focus. This finding is a bit intriguing given the fact that Amazon is known for its customer centricity. Below are some useful insights from a Fast Company

 "Bezos’s guiding principle for Amazon has always been to start with the needs and desires of the customer and work backward. But when it came to the Fire Phone, that customer apparently became Jeff Bezos. He envisioned a list of whiz-bang features,....Perhaps most compelling was Dynamic Perspective, which uses cameras to track a user’s head and adjust the display to his or her vantage point, making the on-screen image appear three-dimensional."

 And team members simply could not imagine truly useful applications for Dynamic Perspective. As far as anyone could tell, Bezos was in search of the Fire Phone’s version of Siri, a signature feature that could make the device a blockbuster. But what was the point, they wondered, beyond some fun gaming interactions and flashy 3-D lock screens. "In meetings, all Jeff talked about was, ‘3-D, 3-D, 3-D!’ He had this childlike excitement about the feature and no one could understand why," recalls a former engineering head who worked solely on Dynamic Perspective for years. "We poured surreal amounts of money into it, yet we all thought it had no value for the customer, which was the biggest irony. Whenever anyone asked why we were doing this, the answer was, ‘Because Jeff wants it.’ No one thought the feature justified the cost to the project. No one. Absolutely no one."

If this is to be believed, Jeff Bezos vigorously backed some features which didn't make much logical sense to the team. Apparent lesson is that CEO shouldn't necessarily become the customer himself or herself but CEO should rather take the role of the best representative of the business.

Eric Schmidt, the former CEO of Google, brings in an interesting perspective in his book- "How Google Works". He doesn't eat his words when he says "Don't listen to HIPPO". Who is an HIPPO? Read on the excerpt from his book to appreciate this point more-
From the book- "How Google Works"

Lack of App ecosystem:

Keeping platform style thinking at the forefront, it is imperative that a product should be able to build an ecosystem of developers (and consumers), who can help extend the product's capabilities further and make it more useful for the consumers. As this Time article points out, 

Amazon’s app store has about 240,000 apps, compared to more than 1 million in the Google Play store.

As Apple's success had proven, to be successful- it is imperative for the Smart Phone companies to have an app ecosystem that helps its users. On the contrary, as the not-so-wide adoption of Windows phone have proven too, the apps or lack of it, really is one of the key factors. Windows, since its leadership change, has started to take the corrective actions in this regard by announcing the iOS and Android bridge projects for Windows OS.  These projects, if successful, would make it easy for developers to port Android and iOS apps to Windows. To me, releasing a smart phone without adequate app support is a fundamental flaw. It would have been accepted in 2007 (when iPhone was launched), not now when already precedent has been set with so many players already in the market.

Micro-management- the root of all evils? Nah.

There are also some apparent references to Bezos' micromanaging the entire project, which might or might not have been true. Even if it were true, i would rather look into this with a balanced mind. What i mean by this is, if we choose to blame Bezos' micro-management as a reason of failure of Fire phone, we should also be open to appreciate where his apparent micro-management worked and delivered a blockbuster product. My take on micro-management is that it is not as bad as it is often made out to be. Micro-management, like empowerment is a management strategy, which is necessary in some situation where there is a need to monitor the situation closely. It is probably bad if it's an inherent trait of a person. It isn't bad if it is used only in situations where it's used is justified.

Begin with the end in mind:

Ok, this one is not a negative learning. One of the things that i read about the how Amazon Fire phone development effort was started was particularly interesting. It said-

Like every product created at Amazon, the Fire Phone began on a piece of paper. Or rather, several typed, single-spaced pieces of paper that contained a mock-up of a press release for the product that the company hoped to launch some day. Bezos requires employees to write these pretend press releases before work begins on a new initiative. The point is to help them refine their ideas and distill their goals with the customer in mind.

Visualizing the impact that the product is bound to make even before its development has commenced didn't seem like an everyday idea to me. This approach has certain freshness in the times when many product organizations start with a formal product requirements document. With this end clear, reviewed and approved- it has the capability to become a sort of oracle against which product specific decisions could be made. Another notable aspect is the fact that Jeff Bezos' act of involving his employees to write the same. With all the flak that Bezos have been getting regarding Amazon's work culture of late, this seem like a perfect counter-argument.

This point also proves that when a product is a failure, it doesn't mean that everything that happened during its inception was wrong.

What's your take on these points? Please do share in the comments.

 

 

Monday, November 16, 2015

Does Intel Pentium Bug of 1990s Still Holds Any Lessons for us?


Information Technology, at its core, is a forward oriented profession. What i mean by this assertion is that, as a general observation, the rate of change that this profession deals with is unprecedented. In my career time span so far, i have seen many a paradigm shifts including rise and fall and re-rise of Microsoft, birth and dominance of Google, a gigantic comeback of Apple, and all these eventually impacting our lives as professionals and as consumers of technology. In dealing with such acute dynamism, in my belief, it is very easy to lose the sense of history of our profession. I personally feel that having a good sense of history for a chosen profession often helps us connect the dots better and better fathom the current events that we experience. History helps connect things through time and I do consider knowledge of the history of our profession important in shaping its future. Most of the today's methodologies and good practices are evolved by bettering what didn't work in the past. To say the least, sense of history also gives as a sense of connection with the past which we should look to not lose.



I was recently reading the book- "Only the Paranoid Survive", the first person account of Andy Grove (former CEO of Intel) on how he dealt with strategic inflection points i.e. the time in the life of a business when its fundamentals are about to change. One of the narration in the book talks about the Pentium chip bug and it goes like as follows (written as is it appears in the book)-

"Several weeks earlier, some of our employees had found a string of comments on the Internet forum where people interested in Intel products congregate. The comments were under the headings like, "Bug in the Pentium FPU." (FPU stands for floating point unit, the part of the chip that does heavy-duty math.) They were triggered by the observation of a math professor that something wasn't quite right with the mathematical capabilities of the pentium chip. The professor reported that he had encountered a division error while studying some complex math problem.
We were already familiar with this problem, having encountered it several months earlier. It was due to a minor design error on the chip, which caused a rounding error in division once every nine billion times. At first, we were very concerned about this, so we mounted a major study to try to understand what once every nine billion divisions would mean. We found the results reassuring. For instance, they meant that an average spreadsheet user would run into the problem only once every 27,000 years of spreadsheet use.

Andy spends a quite a few pages later in the book to tell why this bug was critical and how it turned his thinking around some peculiar was happening in the world around him. Let me summarize that point of view in next few points and also explain its relevance in today's world-

1. The beginning of social media as a force to reckon with:
Internet Forum in 1990s
     We pretty much take social media for granted these days. It generates a lot of data and opinions every passing second, which is very valuable to those who see the need to seek information out of it. This is especially true for anyone seeking feedback for a 
newly launched product or a service. Consumers, on the other hand, provide feedback often without being asked on social media. It more often turns out to a medium for venting out imperfections and bad experiences. This is now. But when we talk about 1990s, when the Pentium FPU bug occurred, things were still in infancy w.r.t Internet and people had started to use Internet forums to share opinions. Intel, then was not in the business of selling the computer chips directly to consumers. It used to sell via PC manufacturers like IBM. Intel's emergence was at the cusp of PC industry turning more horizontal oriented than vertically oriented meaning that earlier one manufacturer like Digital used to manufacture/assemble all parts of a computer (vertical orientation), later each key component became individual business (horizontal orientation) serving the PC assembler like IBM, Dell etc. Andy mentions in his book that with this bug, he smelled something unusual happen in the field. And it was that though he was not selling directly to consumers, he was getting feedback from them directly. He inferred that if this situation wasn't handled with proactive stance, then he could receive a lot of negative backlash. Mind you, this was 1990s, when it was hard to imagine the power of social media. Andy took the corrective actions quickly and even justified the huge cost of this bug- around USD 450 million (mammoth amount now but more so 20 years back).


It stands the lessons for today's times too. Proactively dealing with feedback received on social media is the order of the day. It is easy to manufacture negativity even by bad intentions of the competitors. The birth of techniques such as Sentiment Analysis that help to proactive assess positive and negative sentiments around the events like product releases further help to deal with negative perceptions well. In my recent memory, i am reminded on the social buzz that was created by the security vulnerability in SSL- Heartbleed bug and the negative response generated in the social media when the news about their (hidden) social experiment A/B test leaked out publically where they subjected a certain percentage of their consumers to negative news deliberately. Even though social media as a channel is quite useful to generate feedback but it also makes companies vulnerable to negative publicity in the event of bugs that catch public attention.

2. Handling strategic inflection points need different skills
   In the wake of negative press and crisis-like situation that the Pentium FPU bug generated for Intel, Andy made a very interesting observation in his book. He says-
 "A lot of people involved in handling this stuff had only joined Intel in the last ten years or so, during which time our business had grown steadily. Their experience had been that working hard, putting one foot in front of other, was what it took to get good outcome. Now, all of a sudden, instead of predictable success, nothing was predictable. Our people, while they were busting their butts, were also perturbed and even scared."

In short, the skills needed to handle peace time in business are quite different from the ones needed during war time. People often come to work believing the workplaces to be fair i.e. if i do "X" amount of work, i will get equivalent of "X" credit. While there is nothing wrong in this assumption generally but such thinking (from employee's perspective) do not take into account changing business situations. The reality of today's times is that an effort that would have resulted in a great output (for company and personally) in a certain business situation would not just be enough in a very different business situation. This often happens because of no fault of employee, who did his best given the current situation but probably lacked situational awareness to alter the nature of efforts. To quickly explain this perspective, Nokia's example comes to mind. The story of rise and further decline of Nokia is widely written about. During good times (till atleast 2007), the company made a big fortunes with its existing model (with its phones based on Symbian OS). But when the time came to change to more modern mobile OS like Android, they just failed to move swiftly. I can imagine the employees in this situation would have put in great efforts with their key skills around Symbian OS but due to situational change, the same efforts which bore huge fruits earlier were just not enough to reap similar or greater rewards.

3. Lessons in Defect Advocacy
    To me, the most interesting part of the narration regarding Pentium FPU bug was this- "an average spreadsheet user would run into the problem only once every 27,000 years of spreadsheet use"
This was actually a known problem before the Pentium chip was released. What might have happened is, following the usual defect prioritization principles, it would have been given acknowledged but given less priority as the frequency of this bug happen was staggering 27000 years of spreadsheet use. Now, one may question this data's accuracy, which is probably a fair question but larger point that this case teaches is that the usual defect prioritization approach usually fail to consider the macro aspects impacting the product. Let me explain this point a little bit-
Pentium chip was released at the backdrop of the legendary "Intel Inside" marketing campaign. The extent of popularity (due to marketing efforts) of this campaign was so huge that Intel almost became a household brand. When people started seeing the effects of the error related to this, they put the blame squarely on Intel and not the computer manufacturer. The early social media in the form of Internet forums gave voice to their concerns. Had the defect prioritization decision, take into account the macro environment that the product will operate under, it would probably have been chosen to be fixed. 
One of the key learnings here that is still relevant in today's times is to have a holistic approach towards defect advocacy. A tester advocating the defect should relate the bug information with the macro environment happenings like business situation, popularity of the bug, users impacted and much more. For a tester to be playing the role of the headlights of the product, he/she should not just think about internals of the bugs but also associate it with the necessary business information and related factors.

What else do you learn from this case? Please do share your thoughts in the comments.


Images source:
http://www.amazon.in/Only-Paranoid-Survive-Exploit-Challenge/dp/0385483821
http://copybrighter.com/history-of-social-media
http://www.pitstopmedia.com/sem/minimize-the-impact-of-website-testing-on-seo

http://www.slideshare.net/thoodcpa/strategic-inflection-points-einstein-learning-in-the-cpa-profession

Friday, November 13, 2015

Are the Lessons from Google Wave Debacle Still Relevant ?

In this blog (or hopefully a series of them), i intend to write about the key lessons to be learned not only from the tech products that failed in the past but also i intend to write a bit about why products succeed as well. The motivation to write this comes directly from my experience. I started my career with working on a large ecommerce product, which was (at that time) supposed to be the biggest technology application built entirely on Microsoft technology. Ok, the word "biggest" in the past sentence has more or less an anecdotal reference but the key point is that this project really struggled to find its feet due to various reasons. I will talk more about my own experience in the coming posts but to start off, i wanted to put a microscope on some of the public products that bit the dust. Failures, like success, leaves a lot to be learned from. My personal hypothesis is that most failures share similar reasons that led to eventual results. As a series of this blog, i want to test this hypothesis as well. I will try and bind my narrative to around 500 words so (covering top 3-4 reasons) that it remains within the readability limits.

The first product that pick for analysis is Google Wave. The reason I pick up this product is simply that i got to know a bit about this product while reading the book "How Google works" and it is quite fresh in my mind. As the book says- Google Wave as the creation of
a small team of engineers who took their 20 percent time to explore the question "What would email be like if it were invented today?" Google Wave was said to be a technological marvel, but it proved to be a major flop. Its user base never grew as expected (said to be close to 1 million or so) and Google eventually cancelled this project within one year of its launch in around 2010. Google Wave was said to be "collaboration and communication tool consolidating] core online features from e-mail, instant messaging, blogging, wikis, multimedia management, and document sharing,”

Some reasons for its failure as I researched as below-

Complicated User Experience:
As this Quora article suggests,
In retrospect, the lack of success of Google Wave was attributed among other things to its complicated user interface resulting in a product that was a bit like email, a bit like an instant messenger and a bit like a wiki but ultimately couldn't do any of the things really better than the existing solutions.

One of the studies even pointed that a celebrated Tech journalist even wrote a 195 page manual on how to use Wave.
This fact that someone need to write a hefty manual explaining a product alone would testify that Wave probably missed a trick or two in designing a simpler application. It appeared that it was not only hard to use but it was also perceived as hard to explain.

If we fast-forward to 2015, companies are adopting unbundling as a product strategy, which means that they are decoupling the features to make it more usable or focused e.g. Facebook unbundling messenger from the core app. It means that the world gets to value simpler, one feature app more than an app doing too many things, which Google Wave tried to do.

Launched with Lofty Expectations:
As this Mashable article suggests, The first lesson that Google or any web application developer can learn from Google Wave is the importance of managing expectations. Because the hype window started four months before Wave actually launched, the idea of what Wave was easily exceeded the reality. Phrases like "radically different approach to communication" and "e-mail 2.0" were bandied around, along with buzz-word laden phrases like "paradigm-shifting game-changer."

As I have experienced, when a product is launched with much fanfare, it always runs the risk of being subdued under its own expectations. This is something Intel observed when they launched Pentium chip (more about it in later blogs) when one edge case bug resulted in a loss of close to a half billion dollars majorly because of the marketing hype preceding it.

Lack of Extensions:
As one of the reasons cited in eweek- Google Wave was open sourced and yet failed to catch on with developers. While SAP, Novell and Salesforce.com all vowed to work with Wave, and there were a number of extensions created, the support didn't match that of other Google projects, such as Chrome, for which there are thousands of browser extensions. That's a big killer.

As I explained in one of my earlier blogs, in today’s tech era the successful products are more defined with a platform style architecture where building a successful ecosystem of developers is the key. The absence of incentives attracting enough developers timely impacted the speed at which extensions were created and hence resultant user adoption.

No Integration with Google Apps

Again a reason cited in one of the analysis- Google proudly displayed Wave as its own entity. It would have been better served attached to Google Apps similar to the way Google Buzz was tied to Gmail, with Google suggesting users try it out for certain collaboration functions in Google Docs or Sites.
Integration between products is one of the key problems faced with most big sized tech. companies that typically have multiple products in its portfolio. Big companies usually expands their portfolios by acquiring other companies. Acquisitions usually have a negative engineering impact when it becomes to integration because of conflicting architectures.
The book that I referred earlier- “How Google Works” described Google Wave as an ahead of its time product. I politely disagree to this given the fact that now, 5 years later, the world still doesn’t see a compelling reason to have a product like this. To be fair to Google Wave and its superior technology, Google did use the pieces of Wave platform in Google+ and Gmail. But hearing Eric Schmidt say that Google liked the Wave UI represents a sort of disconnect between what users felt and what management saw the product. At the core, this aspect is something that’s common across most product failures.

Please do share your thoughts, ideas around this blog.