Saturday, November 24, 2012

Why Wi-Fi does not have a future in the long run - an argument based on domestic power generators in India

I'll give the same argument here that I gave during a guest lecture by Ashish Wadhwani, Partner at IvyCap Ventures, in our Corporate Strategy class on 16-Nov-12.

Wi-Fi [including hotspots and tethering] is a temporary solution to the core problem of having a pervasive, low-cost, high-speed, always-available wireless network. Sooner or later, the world will have such networks, and these will most likely be provided by telecom service providers. Just like today we have nearly-universal, always-available, highly-reliable, and low-cost wireless networks for phone calls and SMSes [particularly in developed countries], we will eventually have networks with these desirable attributes for wireless data needs too.

Think about diesel- or kerosene-powered domestic power generators, widely used in homes in India. Why do people install these generators at their homes? It's because the electricity supply in India isn't reliable. Power outages are common and frequent, often unpredictably. Hence people purchase and install their own power generators to compensate for the lack of pervasive, nonstop, reliable power supply. Do people in Singapore also own and use domestic power generators? No, because the power supply there is clean, nonstop and overall reliable.

And sooner or later, this will be the case with India too. India will eventually have a power supply that will be clean, nonstop and overall reliable. You can bet that Indians will then stop feeling the need of having their own generators, and hence will stop buying them. A domestic power generator is a temporary solution to a bigger problem.

You can begin to see the parallels. For similar reasons, as wireless networks keep improving in every dimension, the stopgap solution that Wi-Fi fundamentally is will be needed less and less, as people will increasingly rely on seamless wireless networks provided by telecom companies. So instead of localized pockets of high-speed wireless provided by conversion of wireline broadband into Wi-Fi, there will be all-encompassing wireless networks. Whether 4G does this or 5G or 6G is not the question. The real issue is that this change will happen in the next couple of years.

Monday, November 12, 2012

Could Wikipedia 'affect', rather than 'reflect' the information present in reliable sources, and create a feedback loop?

Consider the term planned obsolescence. The New York Times, in May'11, said this of the term in the context of Brooks Stevens:

"A speech he gave at an advertising conference in 1954 was titled “planned obsolescence,” and while he didn’t coin the phrase, he is said to have popularized it. More significant, he had faith in the concept; for that he was reviled by some." - From the Pen of a Giant of Industrial Design, NYT, May'11

The current version of the Wikipedia article on this term does not cite any source for this claim [that the term was first popularized by Brooks Stevens]. Consider for a moment that the author of the NYT article, in May'11, was looking at this Wikipedia article before he wrote his story on Brooks Stevens, in order to both get a feel of the topic and also to smell test his article. And assume that based on his reading of the Wikipedia article of that time, he included the rather vague statement "he is said to have popularized it" in his article.

Because the current version of the Wikipedia article includes a notice at the top saying "This article needs additional citations for verification", it means that it is possible that someone will eventually pick up the May'11 NYT story as a reliable source, and will cite that story as a published evidence for the claim - in the Wikipedia article - that Stevens first popularized the term.

The set of events described above is entirely possible in real life. And if this happens, it will mean that unverified stuff on Wikipedia going into so-called reliable sources can eventually become verified, just because the so-called reliable source happens to publish it now.

Thursday, November 8, 2012

Google Maps for free is actually a good example of predatory pricing

Predatory pricing refers to pricing a product so low [or free] that you drive out existing competitors [who cannot bear the losses that you can], and also deter potential competitors from entering [again, because they cannot or do not want to bear losses].

France recently convicted Google Maps for unfair competition. I believe that Google giving away Google Maps for free is a perfect example of predatory pricing. If predatory pricing is illegal, then Google Maps cannot be allowed to be free. Note that I am not voicing an opinion on whether predatory pricing itself should be legal or illegal - I am only saying that assuming predatory pricing is illegal, Google Maps for free is illegal.

Why? Google most likely runs its Maps product at a significant loss - a comprehensive mapping service requires truckloads of software programming, energy, data centers, data collection, personnel, and so on. Naturally, this can only be proved if Google is forced to release its internal numbers about the cost and revenue of this product.

Sure, Google can argue that Maps provides Google with valuable data about its users, and that it is able to convert that data into some incremental revenue [and incremental profit] at its AdSense and AdWords products, but this conversion does not mean that the Maps product is itself profitable. Google probably cannot offer Google Maps profitably for free [currently].

Further, Google naturally can continue to bear losses at its Maps product because of its highly-profitable other businesses. Few companies have this luxury. In particular, smaller competitors who cannot provide a mapping product for free will naturally go out of business, because from a customer's perspective, a good product that is free is better than a good product that is paid.

Note that the inability of smaller mapping providers to offer a free mapping service does not result from any inefficiency in the way they run their businesses. Not does Google have a magic wand that allows it to sell Google Maps for free, and yet make profit, while others are unable to do so. Google is purposely making losses on Maps and it is doing so to attract customers away from other services.

Perfect predatory pricing.

Sunday, October 14, 2012

Is a person's IQ fixed? Does "workout" help to temporarily "boost" it? Why can GMAT score be improved by prior practice?

These are some very important questions, I believe. Ideally, you should receive the same IQ score, irrespective of whether you take the test with prior practice or without. After all, the objective is to measure your intelligence, and presumably, intelligence cannot be altered through practice.

In practice, however, one's IQ score is positively affected by prior practice. It's the same with GMAT. I scored 760 on the GMAT in Sep'10. Would I score the same if I were given the test right now? I am sure that I would not score even 730. Any other student will tell you the same thing. I would need at least a few weeks of self-study in order to get back my accuracy and speed on the GMAT.

Which makes me wonder. Are IQ tests really a measure of one's raw intelligence? More fundamentally, is intelligence fixed? I am doubtful that intelligence is fixed. The human brain is a learning computer, and intelligence is measured by not only the raw hardware capabilities of one's brain [say MIPS], but also by the quality of the software programming fed into it. And the latter part can surely be improved with practice. Science might prove that even the former can be improved using the right techniques.

Tuesday, August 28, 2012

Wikipedia and other non-profit websites are taking away millions of page-views from for-profit websites

It's quite natural actually. Any organization that tries to provide a product/service on a non-profit basis, when there exist alternatives from for-profit firms, is a competition for the for-profit firm. A good example is Mozilla. Is Mozilla in competition with Opera Software? Sure. Is Apache a competitor for Microsoft and others? Most definitely. Is Wikinews a competitor to commercial news websites? Theoretically yes, but since Wikinews is so small and undeveloped, practically no.

That last statement is important. Thousands of free, non-profit alternatives to commercial products exist [especially in information and software industries], but companies consider these serious competition only if these get meaningful traction in the marketplace. By this definition, Wikipedia is already a serious competition to many for-profit websites that aim to provide information.

I wrote back in May this year that "The millions of views that Wikipedia pages get each day "eat up" page views (and reduce ad income) on other websites ("

I now feel that the problem is even more severe that I thought back then. Look at the page view statistics for the article on Neil Armstrong on Wikipedia, following his death.

The page views following his death run into several million per day [even before his death, these ran into several thousand per day]. And all of these page views are those that could've been on other, for-profit websites, had Wikipedia not existed [for example, on biography or obituary pages on other websites]. Should Google, Bing, etc., be blamed for blessing Wikipedia results all they way to the top?

Update: This page lists the most viewed articles on English Wikipedia. Looking at the page views, it's clear that Wikipedia causes significant damage to other, commercial websites.

Update [Sep'12, Oct'12]: Cumulative page views for the Wikipedia article on Gangnam Style clearly show the extent of the lost business for other, commercial websites.

Update [27-Feb-18]: Sridevi's page on Wikipedia alone has had several million views in the last 2-3 days. All of this is ad-business lost for commercial entertainment/news websites. Essentially, the charity/donation funds that Wikipedia gets from folks all over the world can be thought of as pre-payment for such articles by a small set of people for a large set of readers.

Thursday, August 9, 2012

Apple Maps, Siri, and a new way to look at cost of goods sold (COGS)

In the age of free online services that accompany hardware devices, the cost of a device cannot be and should not be calculated using only the hardware cost of the device [materials, assembly, etc.]. Look at Apple - it recently launched Siri and Maps as free services for its iOS 6 devices. Does Apple make [or plan to make] money on these services? If not, the cost to build and run these services must be included in the cost of the devices that Apple sells [including the expected future costs], in order to correctly calculate the profitability of these devices. An iPhone or an iPad isn't merely the hardware, it's also the services that come with the device - App Store, iTunes, Maps, Siri, etc.

Sunday, July 1, 2012

Plurality based crowdsourcing will probably lead to mediocrity

Since the IQ of a populace follows the normal distribution [like below], it follows that if plurality is used as the method to implement crowdsourcing, sub-optimal decisions will most likely emerge. This is because the number of votes will be dominated by the votes from individuals of average intelligence level. The number of votes by above-average individuals will not be enough to outnumber the votes by individuals with average intelligence.

Notes on the above thought:
  1. Giving extra weight to votes by individuals with above-average intelligence [using weighted mean] can be explored as a solution to the problem presented above.
  2. Any plurality-based result [non-weighted] from a population with a normal distribution [like above] should be interpreted as the choice of the largest number of people, and not necessarily the "best" choice. For example, which stock should we invest in? A plurality-based result should, in my opinion, almost always result in an inferior yield compared to choices made by exceptional individuals.

Sunday, April 22, 2012

What happens when a strong retailer decides that it's really in the products business

That's what Amazon, traditionally a retailer, has been deciding in recent years. That it's much more profitable to make and sell products, versus merely selling others' goods. While Amazon's private label products are already well known [AmazonBasics, Pinzon and others], and these compete with products from other sellers on Amazon, it's Kindle that has really pitched Amazon against some of the very sellers that Amazon's business depends upon.

Look at the screenshots below [these and more documents are here on SkyDrive]. Amazon is actively trying to steer people looking for an iPad away from the iPad and towards its own Kindle. Now, Apple isn't too dependent on, so it shouldn't feel a lot of negative impact, but here's a worthy caution to those companies who are too dependent upon a single distributor/retailer - the channel partner could one day decide to sell its own products, effectively putting you in competition with it.

Amazon's homepage heavily promotes the Kindle - Apple, et al., can do nothing about it.

Amazon encourages searchers to look at the Kindle.

Update [23-Oct-19]: Just bought a Fire TV Stick 4K product from Amazon's website. Unlike pages of other manufacturers' products, where when you scroll down, you see sections such as "Sponsored products related to this item", "Customers who viewed this item also viewed", "Customers who bought this item also bought", etc., for Amazon's own hardware product, you do not see any of these sections. They don't want to show you any alternatives/competitors. They're not a neutral seller, and they're not one bit ashamed to admit this. Full resolution screenshot in RB 69.

Friday, April 20, 2012

Types of widgets needed on every content publishing website to optimally expose past and present content

The keyword here is expose [both past and present content]. Why? Because there is hidden value in both past and present content that can be harvested by optimally exposing the content. By effectively exposing the existing content, it is possible to achieve higher overall returns on investments already made [by way of reduction in the amount of fresh investment required to continually create new content].

Most Popular box on HBR, Apr'12. The sub-sections here are Most Commented and Most Read. The scopes are Last 24 Hours, Last 7 Days, and Last 30 Days.

In my opinion, the following types of widgets should be implemented on every content publishing platform [such as a blog, a photo or video sharing website, etc., and even for reader comments, since the comments too are a form of content]. Presently, websites implemente a subset of these widgets, resulting in sub-optimal exposure of existing content. Websites should implement all of these widgets.
  1. By publishing date: There should be two sub-sections in this box - newest and oldest. These sections should have list of articles/content in chronological order. Also, a calendar box should allow the reader to pull a list of articles published by choosing a specific date [day, week, or month].
  2. Featured: A list of articles that the website owner wants to highlight/push [although these articles might not necessarily have the highest rating by users]. The owner should be able to select, say 50 articles, and the widget should randomly display a subset, say 10, of those in this box.
  3. Recently edited/updated: Content that has been recently revised. This widget will allow readers to take note of content that has changed.
  4. Random: A randomly generated list of content. A button called "Randomize" will randomly rebuild this list. The idea behind this box is to expose past content on a random basis, to give past content visibility. Scopes: Recently, Today, Last Week, Last Month, Last Year, Ever. Default scope is Ever.
  5. Related: A list of articles that are related to the current article, based on similar keywords/phrases. This widget will require Cloud-based processing of all the content on the website.
  6. Suggested: A list of articles from the current website suggested specifically to you, based on your past content consumption, your demographics/profile, etc. This widget is also intensive, in that it requires significant Cloud-based information and processing.
  7. Popular: Sub sections in this widget will be Most Read/Viewed, Most Shared, Most Commented, Most Cited/Linked. Most Shared includes sharing both via email and social networks. Most Cited/Linked means that the articles in this list have the highest citations [in documents, emails, etc.] and also the most inward links from content out there on the Web. Scopes: Recently, Today, Last Week, Last Month, Last Year, Ever.
  8. By rating: Best rated. Worst rated. What makes this box different from the Popular box is that the sheer number of views, comments, etc., don't matter here. Scopes: Recently, Today, Last Week, Last Month, Last Year, Ever.
  9. Tag cloud: A tag cloud box should display the phrases/words used the most in all of the content on the website. Again, this box will require Cloud-based processing of website data. Once again, there should be scopes available for this box, to allow a reader to see the topics that a website has focused upon in the last day, week, month, etc. Automated construction of a tag cloud could use Amazon's SIP idea.
  10. Tags: The website owner should specify tags to every piece of content, and a user should be able to pull content corresponding to one or more tags [return articles using tag A; return articles using tag A AND B; return articles using tag A OR B].
  11. Search box: Of course, a search box is still required to allow readers to search for content using date, keywords, sentences, etc.

Thursday, April 19, 2012

Why Samsung has an upper hand over Apple with regards to allegations of "slavish" copying of design, etc.

Two primary reasons:
  1. Difficulty/ease of copying: What Samsung has [manufacturing expertise and plants for chips, displays, DRAM, flash memory, etc.], Apple can't cheaply or easily copy. What Apple has [product design, software features, user interface, etc.], Samsung can cheaply and easily copy. What is noteworthy here is that Apple might have spent millions to come up with a particular design, and it's as easy for Samsung [and anyone else] to copy it as it is to say "copy it".
  2. Difficulty of proving infringement: Whether or not two designs are similar depends on human judgment. We do not have an objective way to evaluate the degree of similarity of two designs. So even if it is proved that Samsung did copy Apple's designs, it is easy for Samsung to modify them only so much that they're sufficiently different to address legal hurdles, although they can still be similar enough to provide Apple-like user experience [to Apple's competitive disadvantage].
Based on the equally important reasons above, I predict that in these ongoing design/interface infringement battles, Apple is likely to have a lower hand against Samsung.

Tuesday, April 17, 2012

A predominance of American cases, examples and perspectives is a major shortcoming of today's MBA programs

Most of the cases, examples, laws, perspectives, and stories we study in MBA are related to the US. There is relatively less focus on rest of the world - the Middle East, Asia, Africa, South America, Australia, Russia/CIS, and Europe. In my opinion, this disproportionate focus on America is a major limitation of the top MBA programs today. While it cannot be argued that easy availability of detailed information related to the US is significantly higher compared to regions such as Africa or the Middle East, it is naive to assume that innovation is taking place only in the US, or that interesting cases exist only for US companies.

From a global perspective standpoint, it will be so much better for the students if top business schools set a mandatory 50% quota for using non-US cases, examples, laws/regulations, perspectives, etc., in their curricula. Such a limit will not only give a truly global view to MBA students, it will also teach MBA students the relevance and significance of other countries in the world.

Update [23-Nov-12]: I've also noticed that most of the cases we study are HBS cases [note that this in itself does not automatically imply that the case itself will be focused on America, although in practice that is usually the case]. While these cases are no doubt good, it would be better if a quota is also put on the proportion of HBS cases that can be used, so that students are also exposed to cases from other top schools such as LBS, Wharton, CEIBS, Stanford, Oxford, ISB, etc.

And one more thing. There is an overuse of the example of Apple in most classes, to the point of making students bored. There is no doubt that it is easy to connect with a high-profile example such as Apple, but an over-coverage of Apple leads to shrinking marginal learning [to students] coming from the repeated references to this company's moves.

Finally, whenever innovation is discussed, it is predominantly in the context of Internet, smartphones, tablets, and Web 2.0 startups. There's a world of innovation outside of IT too, folks. Let's not ignore it.

Business schools which adopt the above suggestions [either as laid out above or in modified forms] can use these quotas as a potent marketing weapon, to show to prospective and existing students [and also to rankings bodies] that they are focused on giving their students a truly global view, written by global scholars, and focused on a diverse range of industries.

Saturday, April 14, 2012

Perfectly personalized news service could pose a grave danger

As things stand today, some in the public are able to see a bias in news reporting. Some among the public know what Western media hides and what it distorts, although most of the public is oblivious to these facts. That's because news isn't too personalized today [at least in terms of news sources]. For example, a non-customized Google News shows the same story from a variety of publishers [NYT, CNN, RT, etc.], so a user is able to see different interpretations, different perspectives [and different sides] for the same story.

Imagine a world in which Google News is perfectly personalized. And imagine that a particular user John has told Google News that his preferred news sources are only these: CNN, BBC, NYT, WSJ, FT, Economist, ABC, CBS, Fox, NBC, AP, AFP, and Reuters [basically all Western news companies].

Imagine that Google News, which has perfectly personalized his news experience, shows him stories only from these sources. What will happen? John will see only those stories which are reported by his chosen outlets. Worse, he will only see their perspective of the stories. John won't get to know that there are other, non-reported stories, hidden facts, etc., because he only consumes Google News and it's perfectly personalized. What is troubling is that John might not even be aware that by personalizing his Google News experience, he has actually hidden from his view the other side of the story.

Do we expect John to be smart enough to know in the first place that he should be including channels such as RT, Press TV, CCTV, etc., in his basket? We should not expect this, because we know well how mired in Western media the Western public is.

What if Google, algorithmically or secretly, chooses to automatically bias its Google News service to exclusively or predominantly promote stories from Western news outlets, in the name of personalization? We don't want such a over-personalized future, when the relative freedom of today is already so costly to nations that aren't allies of the West.

Thursday, April 12, 2012

Factors crucial to the success of rechargeable batteries in vehicles

I can think of the following factors, and having constructed this list, it is a lot easier to analyze the issue of the future of car batteries.
  1. Cost [of the battery, of course].
  2. Capacity [to store energy]. Capacity is inverse of size, since you can always increase capacity by combining battery packs. So capacity here refers to capacity while keeping size constant.
  3. Charging time
  4. Power capacity - how fast can a battery deliver power.
  5. Commercializability - the ease and practicality with which a battery can be manufactured at scale.
  6. Energy efficiency - what percentage of input energy is retained [and delivered] by battery.
  7. Life - in terms of both number of charging/discharging cycles and time [years].
  8. Safety - what happens if the vehicle meets with an accident? Toxic? Fire hazard?
  9. Self-discharging - how quickly does the stored energy deplete when the vehicle isn't being used.
  10. Weight - closely related to capacity and size, but size and weight are two different things, and we want both to be small.
  11. Recyclability - can the battery be recycled easily, fully and profitably? What environmental risk do the non-recyclable components pose, and what is their recommended disposal method?

Wednesday, April 11, 2012

Methods by which companies are trying to make virtual keyboards feel like real ones

Five methods come to mind, which can be grouped by three fundamental human senses:
  1. Visual/Color: Change the color of the key which is pressed/touched (momentarily).
  2. Visual/Size: Increase the size of the key (again, momentarily).
  3. Visual/Animation: Animate the movement from the original size to the larger size.
  4. Audio/Sound: Make a clicking sound for each keystroke.
  5. Touch/Vibration: Vibrate the device to indicate that the key has been pressed.