In the Web Scale Era, Size Matters

Many now seem to believe that bigness is itself a problem to be solved by antitrust action (breaking firms up into smaller entities). That recently has manifested itself in antitrust actions taken even when a particular merger does not remove any competitors from a market.

Another way of stating the concern about “bigness” or “size” is to note that the need for scale itself drives the bigness trend.  In the era of “web scale” computing, scale matters.

In other words, business success now requires a certain amount of scale, since network effects exist (value is created as the size of the community or network grows). Another way of describing the importance of scale in any endeavor with network effects is to say that value is correlated with size and that, therefore, profits are correlated with size.

source: Dailyhunt

source: Ahmed Banafa

The point is that, in any business with network effects, scale matters. And that means “bigness” matters.

When AI is Commoditized, Where Will Value be Generated?

In principle, one might argue that open source AI and machine learning capabilities will allow small teams of people, using relatively small amounts of capital, to rapidly mimic the features and functions of any product.

However, in principle, one might also argue that gleaning insight, even when such small teams can use AI, will remain difficult, as the best insight will be generated by the smaller number of firms with huge amounts of aggregated past behavior.

To be sure, the ability to rapidly copy features and functions of new and existing products, using machine learning and artificial intelligence, will be widely possible.

In many ways, that is simply an intensification of current trends, where cheap and plentiful computing resources allow smaller teams to move fast, consuming less capital, than would have been required a few decades ago.

Open source helps, too. WhatsApp was able to build a global messaging system that served 900M users with just 50 engineers, compared to the thousands of engineers that were needed for prior generations of messaging systems, says Gianluca Mauro, AI Academy founder.

The same thing will happen as AI tools likewise are open sourced, allowing small teams of engineers to build state-of-the-art AI systems, he argues. That is going to mean that relatively-routine features and functions of any product will rapidly be copied by competitors.

So is that a source of value? Maybe, but maybe not as much as you might think.

The corollary is perhaps that value will be created not by firms able to apply AI to business processes, but to entities that own huge data stores. If the ability to create code, and therefore create features and functions, essentially is commoditized, unique value will then be possible when some entities are able to mine bigger and different stores of past behavior.

The reason is simple. Creation of  predictive models will be more accurate when using huge stores of highly-granular past history. If the value of AI is the ability to create insights, then the owners of the most relevant data should be able to glean better insights.

Many important product features and functions, even those dependent on AI, will be capable of being reverse engineered at relatively low cost. As always with intellectual property concepts, the key is that it will be lawful to create a particular implementation of an idea or function, if unlawful to implement an idea, function or process in exactly the same way as another provider.

Almost by definition, then, the biggest entities, with the greatest number of interactions, sessions, transactions or queries, will have the biggest opportunities to create insights. Almost by definition, small firms will possess little internally-owned data of this type.

The great potential “leveler” then becomes the ability to lawfully use data stores owned by the biggest owners of data. Privacy laws might prevent that, however. In the coming AI era, there might still be trend toward bigness, for that reason.

How Much Would You Pay for an Ad-Free Facebook or Google Sites?

How much would you pay for a subscription-based Facebook or Google? The question now is more than academic as there are strains on the ad revenue model.

Facebook average revenue per user in the U.S. market was a bit more than $6 a month in the fourth quarter of 2017. So the question is what Facebook would have to do to create a revenue-neutral subscription product.

Using only the precedent of converting ad-supported, bundled TV channels into over-the-top streaming alternatives provides some guidance, if only rough guidance. Surveys have routinely shown that consumers deem the expected price for a single TV channel to be less than $2 each, per month.  

Suppliers have priced single-channel subscriptions in the $6 to $15 a month range, though.

If you apply the same ratio to Facebook, that implies a retail subscription price for a “no ads” service of perhaps $18 a month up to $32 a month. That is likely too high a replacement price, though, since Facebook cost of goods is far lower than is the case for TV content providers who have to buy rights to their content. So cut Facebook’s cost in half.

That implies monthly charges between $9 a month and $16 a month. How many people do you believe will pay that much?

Source: Facebook

Google ARPU in early 2016 was in the range of $13 per quarter. By some estimates, Google (Alphabet) reached  $7 average monthly revenue per user by the end of 2016.

Using the same methodology, a subscription-based set of Google products (all Google sites, for example) could cost $14 a month or more.

As always, there are a couple really-important questions: how much will consumers pay, and how many consumers will pay, instead of using ad-supported versions of these products?

Asking consumers whether they will pay for ad-free products is one thing. Actual behavior tends not to track such preferences in a linear way, and almost alw

End of the Ad-Driven Internet Business Model?

There is no inherent contradiction between any particular business model and obtaining scale or market share leadership, but there are imperatives to scale (network effects) in any ad-supported business model that might tell us something about the next wave of internet application development.

Put simply, will the next wave be driven by business models less dependent on network effects? If so, then the business need to drive scale by any means necessary (driving addictive or abusive behavior) will be less.

It remains impossible to pinpoint what the next period of app leadership will look like, except to say that observers expect a “next wave” to develop. The “post-Google, post-Facebook” era is hard to imagine, but  subscription-based business models, with firms operating for long periods of time at lower volume (less scale) seem almost inevitable, given the building backlash against advertising-supported models.

Should privacy become a big-enough end user concern, then alternate revenue models based on subscriptions are nearly inevitable. Assuming one rules out donations as an attractive long-term model, there are few other options.

If one removes all revenue models dependent on monetizing user information, that mostly leaves subscriptions, commerce or device purchase (lease and rental being variants) as the main possibilities.

source: 1000ventures.com

In principle, an appliance model (like purchasing a phone) is a conceivable, if far-fetched way to sell most applications. It is not without precedent, though. TiVo sold its time-shifted TV capability using both a subscription and device purchase model.

To the extent one conceives of in-home Wi-Fi as an application, there is a “rent or buy modem” model. In the business customer segment, there are subscription plans for Boingo venue Wi-Fi, if no device purchase model.

source: Prof. Michael Rappa

Still, in principle, a few highly-valued apps conceivably could be bundled with a device purchase. Think of a smartphone purchase that includes a two-year or three-year subscription to Netflix. It is a stretch, but is possible, if not terribly likely.

So if advertising models and device purchase models and donations are mostly infeasible, what remains? Mostly versions of the subscription model, or pay-per-use models.

And Facebook already has hinted that if the ad model is compromised, a subscription-based model that does not require intrusive profiling and behavior tracking could be created. That has important implications related to scale.

Most ad-supported consumer revenue models work best at scale, since most advertisers are used to buying on a “cost per thousand” basis. So the more “eyeballs” a venue commands, the more revenue it can earn.

source: Prof. Michael Rappa

Subscription models also are scale-dependent, up to a point, since but any single consumer is willing to spend relatively limited amounts of money on subscriptions. That obviously means there is a minimum number of “units sold” required to support firm operations.

The main point is that subscription models do not have the same “cost per thousand” business dynamics. Scale might be desired, but is not required, to the same extent.

Pay-per-use models also are conceivable, if logistically cumbersome. We are familiar with pay-per-use video or music models (buying songs or renting movies and shows). The cost of billing instances of usage that individually have small costs is one issue, but conceptually, micro-payments are possible.

The point is that, in the next era, if advertising does not drive revenue models, then something else will. Those other models arguably are less dependent on scale (cost per thousand) than advertising.

To be sure, there are many possible ways to implement a revenue model based on commerce, for example, though the challenge of creating an app model using commerce means the supplier has to avoid the “intrusiveness” and privacy issues that make “targeting” so useful for merchants.

And if opposition to retail data mining becomes vigorous enough, even the commerce model might not be so workable.

And that leaves one with subscriptions as the easiest alternative model.

Possibly significantly, that could mean a new era that initially begins again with lots of smaller providers. When consumers must buy apps on a subscription basis, they generally are picky about what they buy. So scale (volume) will be an issue for most app suppliers.

And that, in turn, answers the concerns about “bigness” and “lack of innovation” in the present era. If the advertising business model becomes toxic, subscriptions will be the main alternative for most potential apps. And that will mean, at least for a time, that bigness is not so much an issue, as demand will be fragmented.

The point is that markets tend to work, when customers are able to make choices with low friction. A big shift away from ad-supported apps is going to mean some present problems (privacy, perceived barriers to innovation) will tend to fix themselves.

There is no easy way to solve the problem of bad manners, aggressive behavior and courtesy, though. But peer pressure seems to work well enough, in other areas of life. What we now need, to fix the “bad manners” problem, is enough peer pressure from people to shame abusers into exhibiting good manners.

Creative Destruction in Telecom Industry Requires Creative Construction

In 1996, with the passage of the Telecommunications Act, regulators thought they would unleash a huge new wave of competition and innovation in the telecom business. So they unleashed competition in voice services, about four to five years before the voice business reached an absolute peak, and began a sustained decline.

Innovation did come, as did competition that sliced prices. But the innovation came from suppliers of internet-based services. In fact, the whole global telecom industry stopped trying to create the “next generation network” platform, and instead adopted TCP/IP, the internet model, completely and irrevocably.

"Creative destruction" has been at work, where the new displaces the old. It is not pleasant for "losers," as profitable as it is for "winners."

Looking back a few decades now, it is simply historical fact that as profit was driven out of the old telecom business, so was investment ability and relevance. Where telcos used to create and own customer premises gear, switches, operating systems and so forth, today a shrinking roster of suppliers exists to supply such products and platforms.

Less investment in basic or applied research now is conducted by service providers, as they simply cannot afford to do so. The converse happens at the fastest-growing and largest internet app provider and platform firms. There, internally-owned development is the norm.

source: Third Way

There are some other “scary” implications for capital-intensive access providers. App development these days is lead by asset-light firms (software-based firms). So access to capital--vital for access providers building expensive networks--is less a requirement for app development, where, arguably, the greatest value now is being created.

As Professor Jerry Davis notes, Blockbuster, which was delisted in 2010 and liquidated, had 84,000 employees and more than 9,000 stores at its peak. Netflix, which arguably drove Blockbuster out of business, has 3,700 employees and rents server space from Amazon, even as it expands globally.”

“Zillow has 2,200 employees; Yelp has 3,800; and Facebook, with a market capitalization of more than $350 billion, has just 13,000 employees globally,” says Davis.

Any plier of any trade, all companies selling a product, every non-profit organization and every governmental agency can continue to sell products--or avoid bankruptcy--only so long as there is demand for the products.

That is why people sometimes move from one industry to another (some are shrinking, some are growing), why companies must change the products they sell, convince buyers a product is needed  and why non-profits often must find some new problem to solve (the original problem actually is solved).

Almost perversely, curing a disease means any company solving that problem must find new diseases to eliminate. As the old adage goes, there is not much of a market for buggy whips these days. A better current example is demand for dial-up internet access, phones unable to use the internet or landline telephones and services.

There sometimes are instances where “fake problems” are said to exist, so entities can stay in operation (and continue to get donations, volunteers or revenue). Political parties and politicians arguably do that all the time.

In the telecom industry, we have seen many example of this sort of creative destruction. There no longer are switchboard operators. Many people and firms that once made a living selling long distance service are doing something else.

Specialists selling services to small businesses discovered they no longer can compete with the older products when up against cable TV providers of those products. Likewise, many competitive local exchange carriers find there is a non-existent consumer services business model and a slender opportunity in the small business segment, with the greatest opportunities (perhaps targeted opportunities) in the enterprise segment of the market.

The point is that every industry affected by competent new competitors and the internet is going to have the profit margin wrung out of it. Big rearrangements must necessarily be made, and regulators must be wise enough to let them do so.

Creative destruction is destruction, nevertheless. Survivors will have to creatively construct their new roles. Regulators need to allow that to happen.

Freedom is the better framework for all in the internet ecosystem. Keeping industry segments in boxes does not make sense when the walls between industries keep coming down.

There are, to be sure, issues around privacy, trust and so forth that are going to be addressed. That is a different matter from impeding participant search for, and construction of, new roles, to supply new value.

Firms and industries are shifting roles and value propositions. Let it happen.

Do You "Lack" a Tesla?

Do you “lack” a Tesla or a yacht? If so, should the government take action to allow you to fix that “lack?” That’s a silly analogy, but makes a point.

When a government report says 31 percent of residents lack a fixed network internet access service, all we know is that consumers do not buy a service. We do not necessarily know why that behavior exists, and there are many possible explanations.

When consumers do not buy a product they actually can buy, even an arguably valuable and, in some cases, essential service, there are potential problems.

Affordability is among the most-obvious concerns. Were the networks literally not built, that would be a problem, as that would prevent customers from buying. If potential customers did not know how to use the internet, that would be another type of issue.

But if consumers make rational choices not to buy a particular product, that actually is not an addressable “problem,” but an expression of consumer choice and freedom. There are many issues we ought to be concerned about. People making rational choices not to buy the particular expression of a product is not among those problems.

According to the report by the Mayor’s Office of New York City, about 12 percent of homes have no computing devices of any type in them. That would explain the “do not buy” (“lack”) situation. About 10 percent of homes report using mobile data, about one percent say they use dial-up access, while about three percent of homes say they could buy, but do not wish to do so.

With the caveat that use of broadband internet does correlate with income, education and use of computing devices, possibly 26 percent of homes that do not buy broadband have some plausible explanation. They do not wish to buy, use an alternative form of access.

As always, single-person households are least likely to buy broadband internet access services, suggesting the value-price relationship is most problematic in single-person households. Fully 42 percent of single-person households do not buy a fixed network broadband connection.

For households with at least two people, “not buying” happens in about 25 percent of households.

That is an illustration, in many cases, of wireless substitution, where customers make choices to rely on mobile service for voice, messaging and internet access.

Still, poverty and income are correlated with buying of broadband internet access. Using several indicators of “poverty,” between 20 percent to 28 of New York City households are defined to be “in poverty.”

Subsidized usage is among the tools to deal with the “cannot afford it” problem. But so are free Wi-Fi hotspots, mobile substitution and use of public library facilities.

Still, there are several issues, even there. Some households really do not wish to use the internet, and therefore will not wish to buy broadband, at any price. Others will rely on a mix of public resources, Wi-Fi hotspots, “use at school” and mobile substitution to meet their needs.

The point is that there are many reasons people “lack” fixed network broadband, and some of those problems cannot be easily fixed. Choice plays a part. Other resources play a part, as do subsidies. Lower prices, caused by competition, also play some part.

The point is that there are lots of reasons consumers do not buy particular communications products, and perceptions of value, compared to price, always matter. A majority of U.S. households no longer buy fixed network voice services. That generally is by choice. The product simply is not viewed as desirable, and there are other substitutes.

In the coming 5G era, the range and value proposition of new choices is likely to grow substantially.

What is Status of Internet Free Speech?

Ideas and content we find objectionable often are the basis for tests of freedom of speech and press in the United States, and that has not failed to be the case, so far in the internet era.

Among the early First Amendment cases to address internet app freedom, broadcast-style  restrictions on “indecent” communications were part of the Telecommunications Act of 1996. As often is the case, freedom of speech often involves speech (ideas, content) we do not agree with, or are otherwise troublesome.

In the case of  Reno v. ACLU, in 1997, the U.S. Supreme Court held that the Telecom Act restrictions on both the “display” and “transmission” of indecent communications online violate the First Amendment.

“Through the use of chat rooms, any person with a phone line can become a town crier with a voice that resonates farther than it could from any soapbox,” the court said. “Through the use of Web pages, mail exploders and newsgroups, the same individual can become a pamphleteer.”

There is “no basis for qualifying the level of First Amendment scrutiny that should be applied to this medium.”

So far, other cases have dealt with protections for child pornography, again illustrating the principle that tests of freedom of speech often come in troublesome ways apparently unrelated to the arguably central principles of freedom of political speech.

In Ashcroft v. Free Speech Coalition (2002), the Court struck down a federal ban on “virtual” child pornography in the 1996 Child Pornography Prevention Act (CPPA). That case found that a prohibition of images that “appear to be a child” engaging in sexual conduct where no actual children were involved prohibited a substantial amount of protected expression and violated the First Amendment.

In Ashcroft v. ACLU (2002), the Supreme Court reversed a decision of the 3rd U.S. Circuit Court of Appeals to enjoin enforcement of the Child Online Protection Act (COPA), successor to the Communications Decency Act.

Once again, obscenity was the direct issue at hand, not political speech.  

The Supreme Court ruled in June 2003 that a law requiring libraries to filter pornographic internet content was lawful. In the United States v. American Library Association case, the Supreme Court ruled that filtering software does not violate the First Amendment, even though it blocks some lawful web sites.

There are other potential examples of unappetizing areas where “free speech” rights will be hard to evaluate and protect. Do terrorist organizations have such rights?

U.S. network neutrality rules recently came into play, as some might argue regulators used common carrier regulation to take away First Amendment free speech rights. All things internet traditionally have been unregulated data services, not a highly-regulated industry using the common carrier framework.

Utility regulation provides far-less freedom than do laws regulating broadcast TV and radio, cable TV, newspapers, magazines and internet media.

Also, much network neutrality arguments premised on promoting freedom actually do not have much--if anything-- to do with actual blocking of content, and quite a lot to do with the business models of various participants in the internet value chain. Nor is it ever easy to separate the permissible management of congestion on communication networks (lawful) and the separate business practices relating to how networks price and package their access services.

The new wave of threats actually do not come from the government restricting freedom of speech, but in other ways. In other words, it arguably is no longer enough to insist that the government be barred from restricting speech.

In many cases, other actors (universities or public schools) ban speech to protect the sensitivities of people who do not wish to hear ideas they disagree with. That is the problem, some might argue, with “hate speech” codes. In another context, “hate speech” often is political speech, and cannot be blocked by the government, though often such speech is blocked by non-government actors.

That might be the big new development: the First Amendment prevents the government from infringing freedom. The First Amendment does not protect freedom when other actors (firms, people, organizations) are the sources of restrictions.