I’ve recently been reading The Goal, a classic business book in novel format (love story included, no extra charge). It explains how traditional cost accounting leads to irrational decision making.
Many of the metrics and decisions of enterprises are creating local optimizations that have the overall effect of moving the organization further away from its one true goal (hence “The Goal”). The goal of a commercial enterprise is to make money for its owners. (My apologies to continental European readers hoping for social solidarity and employee satisfaction — we don’t do that here.) The goal can be expressed in different ways. Traditionally, ROI, cash flow and net profit are the financial measures. But this is not the only way.
The book (and a whole mini-industry of follow-ups and corollaries like this) proposes an alternative approach to cost accounting, namely “throughput accounting”. The core fallacy of cost accounting is to try to assign costs and revenue (value added) to individual steps of a process. This attempts to locally optimize the system, but globally is almost certain to lead to the opposite of what is desired. For example, inventory appears on the balance sheet of a corporation as an asset. But in colloqial terms it is a liability — you want less of it, not more, because it ties up working capital and makes you less lean and agile.
Throughput accounting focuses on value delivered to customers (“throughput”), not cost. Secondary to throughput is inventory, which includes anything you’ve bought in anticipation of turning it into throughput. Raw materials are one such thing. Inventory could include intangibles like consultancy for product development; it could also include the undepreciated value of machinery as well as product materials. Both of these are deviations from cost accounting. Finally comes operational expense, which is the outlay of turning inventory into throughput.
The goal of throughput accounting is to increase throughput whilst simultaneously reducing inventory and operational expense.
The twin brother of throughput accounting is the Theory of Constraints (TOC). In a nutshell, the throughput of a system is always limited by a system constraint somewhere in the delivery chain. This could be a physical constraint of a machine, but could equally be a policy constraint (“you can’t do that because we don’t work that way here”), or simply market demand pulling inventory through the system. Whatever the constraint may be, you need to subordinate decision making to optimize the throughput of that constraint. The throughput of the system is determined by the rhoughput of the constraint.
OK, that’s enough theory. You can go and read the literature yourself. So on to telecom.
I’ve been having an email discussion with David Anderson, who runs a blog on agile software development and the theory of constraints. Some pretty spectacular improvments in productivity and predictability have been shown using TOC for software development. He’s got an article on throughput accounting principles applied to telecom, which I recommend you read right now.
What is the goal of a telco? As I said earlier, this can be expressed in different ways. In product terms, I believe the purpose of a telco is to maximize the number and value of user communications events, whilst simultaneously protecting the user’s attention from unwanted interruption. This only applies for a telco that has ambitions to supply anything more than raw transport.
What are the implications? Well, the purpose of a telco isn’t to build networks. Networks aren’t thoughput. Customers don’t buy networks. Networks are inventory and operational expense (caused by depreciation). Customers buy communications events (“messages”). These may be sold individually, or as a bundle.
In David’s words:
The inventory is the data to be communicated (including voice calls). The raw material is the “inspiration to communicate”. You generate more inventory by stimulating demand to “talk more” or use higher bandwidth services.
The constraint is the network capacity, or in the case of wireless, the available spectrum. Any telco that is proud of being nowhere near using up its spectrum resources is misguided: it is probably carrying unnecessary inventory, because network constraints are easily elevated by adding more routers and switches.
Telcos do a lousy job of managing “inspiration to communicate”. American Express send me an endless stream of offers with inspiration to spend. My nameless telco suppliers never suggest now is a great time to call to my mum or message my boss. So this suggests a change of emphasis. For instance, social software is clearly targeted at inspiring communications events. The phone system’s value needs to be enhanced by integration with social software. I’ll go into this in more depth another day. Funnily enough, BT used to have an advertising campaign based on “It’s Good to Talk”, essentially trying to promote phone calling as a passtime and stimulate demand. I think we can do better!
Users need ongoing inspiration to communicate until the pipes are completely full.
There are some things that make it hard to model telecom using TOC. The inventory is extremely perishable — every moment the fiber isn’t lit or the airwaves excited is a moment lost forever. Furthermore, not every user communication is welcomed.
In normal TOC parlance, a faulty product delivered to a customer simply counts as zero throughput. But a piece of spam, or a misdialled number at 3am cause a qualtifiable loss of utility to the customer. It seems to require a concept of negative throughput. This probably implies more theory is required. Perhaps the inspiration to spam counts as negative raw material.
We do a terrible job of protecting the user from unwanted communications. A smart phone service would intercept the 3am call from a first-time caller, and pass them straight to voicemail with a special greeting: “You have called Martin Geddes on 913-484-6903. The local time is 3am. If your message is urgent and you wish to ring the number you dialled, press 1 now, otherwise please leave a message after the tone.” And I hardly need discuss email spam as a collective failure of the ISP industry to achieve its goals.
Our virtual communications services are certainly well outside the bounds of a management accounting system conceived for manufacturing physical goods. But without a solid idea of what the goal of a telco should be, or a theoretical model for knowing whether you’re getting closer to it, you will never escape the cycle of network overbuilds, capacity glut and service provider disintermediation.
As a foreign body in the USA, when I first came to live here I couldn’t help notice some recurring themes. A lot of places sell a lot of really cheap stuff — cheap food, cheap clothes, cheap cars, cheap houses. And when you buy something, it tends to be physically large. I’m not saying you always get a lot out of it at the end, because sometimes that largeness is filled with air, soy protein, high fructose corn syrup and styrofoam. But always large. And consistently crap. Out of date food on the supermarket shelves. New cars that look like the height of 80’s fashion and with the durability of a wet paper bag. T-shirts that turn to rags on their first wash and dye all the rest of your clothes pink in the process. Enormous houses that have less durability than my parents’ garden shed back in London.
Cheap. Crap. And lots of it. The driving force of the world’s largest economy. Just look at Wal-Mart. “Always low prices”, my arse! “Cheap, crap and lots of it! Come on in and load up in Sam’s place!”. Just look at the prices, never mind the real cost.
Which brings us to our old friend, the Internet. In many ways this is the Wal-Mart of data communications. It’s hideously inefficient at real-time communications. Every penis extending spam packet needs to be delivered at the same speed as my webcam with the oldies back home. You end up using a lot more of it dealing with waste and breakage than if you’d bought something decent in the first place. It’s truly awful in terms of efficiency at managing jitter and routing: look at how Skype sends your data stream over multiple simultaneous routes in the hope of something getting there in a timely fashion. We end up building endless patches to deal with the lack of hooks into the analog world and things like identity infrastructures. Every spammer and DDoS fiend is anonymous because nobody thought of creating a governance mechanism to deal with abuse of an IP address.
But it’s really simple. And thus has a very low cost to turn standards into hardware and software that universally interoperates. Despite the technical horror of the whole edifice, it does deliver real-time video and voice in a reasonably reliable fashion. And it’s economic efficiency has led it to fill every data delivery nook and cranny around.
It’s easy to get on a high horse and say that we should have a “better” open network, with channel reservations, jitter and latency guarantees, etc. You may also think that shoppers at Wal-Mart would be “better off” not buying supersize packs of Doritos Cardiac Arrest Flavor and unstimulating plastic kiddies toys. Yet the economic reality is that Wal-Mart is the ultimate shopping monster and the Internet is the ultimate data monster. Cheap. Crap. And lots of it. It’s the American way.
End-to-end isn’t just a technical phenomenon of the network and application architecture. It is a philosophy about value creation and efficient allocation of economic resources to meet user needs. It is also deeply political. It effectively states that the users are in control, even if the original paper on end-to-end arguments casts it entirely in technical terms:
End-to-end arguments are a kind of “Occam’s razor” when it comes to choosing the functions to be provided in a communication subsystem.
Indeed the original paper fails to point out an important fact: the reason all the examples given support keeping functionality in the upper layers of the application stack at the end points isn’t because it makes those applications more technically efficient per se. Instead, it is because the user places value on certain features (reliable delivery of voicemail, secure file transfer). It is the economic desire that then drives the need to create an efficient architecture for their delivery.
There is a subtle difference between this and a pure technical efficiency argument. It says that the user should decide what the valued functionality of the system is, not the product development department of a telco or network vendor. Hence the user needs to express those desires in the only way they can: by control of the handware and software at the end points. Hence the success of companies like Nokia and Microsoft, and the relative decline of the Lucents and Suns (sorry, “the computer is still the just the computer”).
The effect of an “intelligent network” above the IP layer to constrain, inhibit or degrade application functionality is secondary to removing friction in turning user need into usable application. The Internet is successful because of its economic attributes, which it inherits from its technical design. As many a dead dotcom attests, technical efficiency at delivering something customers don’t value does not equate to long-term success.
A recent thread between some of the bloggerati highlights the political nature of the argument:
… I would love to see their [US presidential] candidates make an impassioned plea to keep the Internet free of interference from the entertainment industry. I would welcome this for two reasons.
1. First, I’m part of a constituency, like many others, who are looking for a candidate to vote for who supports our primary issue. Nothing unusual about that, easy to understand.
2. But as important, it would signal that the candidate is not beholden to the media companies. I would happily give money to candidates for ads that warn that the media industry is trying to rob us of our future, and explains how important it is to protect the independence of the Internet. Use the media industry channels to undermine their efforts to the control channels they don’t own, yet.
Via Doc Searls comes this partial retort from Mitch Ratcliffe
My take on this issue of making the Net my primary issue is that this would be both counter-productive and destructive to the Net as a thing, since it puts the definition of the Net into the political domain when what is really at stake is a series of procedural decisions about the flow of information.
Well, not really. Soon, all your television (or whatever comes after TV is Nettified) will flow over the Internet. All your radio. All your music. Every magazine you read will be downloaded, customized and printed locally. Every snippet of news. Every photo you share.
Ultimately, if you can’t be there in person, or whistle really well, then every interaction you have with the outside world will flow over the Internet. Every idea you send or receive.
And you are part of this world. As they say, You Can’t Be Neutral on a Moving Train. You are part of history; there is no opt-out. Inaction is not an absence of action, merely an abdication of your responsibility: where were you Daddy when the Internet was no longer free?
Statements about who controls what economic resources are political. Statements about who controls the flow of ideas and criticism are political. End-to-end is a political philosophy. Don’t fool yourselves that it is otherwise.
Farewell, PSTN — we loved you so. But via Gizmodo today comes a story on end-to-end encrypted cellphones calls. Just $3500 for a pair of devices, but this will certainly become standard-issue to corporations within a few years. Read their FAQ, it’s fascinating.
They admit to some latency in the voice service (it runs over the GSM data channel), and that it only protects against man-in-the-middle attacks on the content of the voice call. The addressing data (i.e. who calls whom on what devices) is in the open. And you’re still vulnerable to more “traditional” attacks like bugs. The source is freely published for your inspection against backdoors. And there’s a Windows client for those who don’t have a few thousand dollars for a cell phone. All in all, an attractive deal, although the economics have some way to go for mass-market appeal.
Telepocalypse pundit prediction alert: I may come to regret this, but I believe that encrypted telephony being demanded by enterprises will be the bullet in the PSTN’s head. At present, an insider can easily tap into the corporate LAN and see all voice traffic on a converged network. The change will start with board rooms, contract negotiators, business development, legal, financial and HR departments. Then a general uptake of encrypted telephony internally. Finally, the islands will become joined up. And the PSTN will be a bedtime story for your kids, and how things were in the old days.
Via Arts and Letters Daily comes a useful reminder on how the intrinsic mobility of cell phones disconnects them from geography, as well as disconnecting the user from their public surroundings.
It’s kind of obvious, but you have to point of to telco-heads that IP addresses don’t begin with a country code. This is a political statement, that the network world is flat and borderless. The consequence is you can’t charge for crossing borders. It’s interesting to see this brief twilight world of mobility, but with the pain of international call charges and roaming fees. The stories we will be able to tell our grandchildren!
If Larry Lessig thinks that code is law, then the choice of namespace is part of the constitution. Anything that touches the user’s identity must also touch your business relationship. I’ve seen things like the choice and structure of user identifiers delegated to technical design fora, with executives totally unaware of the long-term business implications of the decisions being made.
I was amused to see how Cingular are trying to distance themselves from any services that might actually make money. The dirty secret of the ISP industry is that the profit is all in illicit chat, dirty pictures and finding ways to get high or get laid (ideally both, simultaneously).
The irony is that the telcos with their closed-network closed-system mindset have fallen into a trap. A true end-to-end network operator doesn’t get their brand tarnished by the data that flows over their network. When compelling content equals sex, gambling, porn and general vice, you can’t have it both ways.
Perhaps the solution is to have a tiered portal model; the Yahoos, AOLs and MSNs of the world carry the nudge-and-wink businesses within the framework of the carrier portal. (I’m using “portal” in its loosest sense here, as a collection of voice and data services.) The carriers then get referral kickbacks for delivering customers to the services that their own brands can’t be associated with.
What a strange idea. The poor family in the rotting urban core of a city have to pay a tax on their phone line to keep a wealthy family’s ski mansion in the hills connected. Anyway, today’s sermon is not about social justice, just about the logic of universal service funds.
The USA has long been concerned with rural telecom, with its huge distances and relatively recent intense urbanisation. These pressures are also showing up now in densely populated western Europe, where universal voice service was achieved long ago, but broadband access remains an issue.
As always, the end-to-end issue is affecting all forms of regulation, and exposing assumptions from long-ago that may no longer hold. There are two problems with universal service funds: they aren’t anything to do with service, and are only universal in a very limited sense.
The cost of stringing a wire, glass fiber or microwave link across the nation is roughly proportional to the distance covered. That means network access in rural areas is considerably more expensive. In the past, you also needed more switches and repeaters out in the field too. But those days are gone: the distance to your application server is more constrained by the speed of light than network geography. So the cost of service is uncorrelated with the location of the subscriber. We want universal access, not universal service.
I would therefore propose that any “universal service” fund that persists in future be re-structured, and it be hypothecated for unbundled access. Indeed, I would go one stage further, and insist it only be used to provide Internet access. Why have a public subsidy for a dead-end network technology like the PSTN?
Now, you might argue that people in rural areas with low incomes need a subsidy for service too. Perhaps they can’t afford Vonage. But that’s a general social engineering problem, not a telecom problem. If you want so subsidise habitation in remote or sparsely populated areas, there are much better ways of doing it than taxing urban telephony.
Indeed, hypothecated taxes for telecom subsidy create a dead loss on the recipients. Rural subscribers may control the spending of that money indirectly through the ballot box, but ideally they would be able to individually chose how to spend it. Some people might prefer a better private track to their mountain retreat than a new phone line.
The second problem is that the resulting service is not very universal. Companies like Western Wireless are happy to dip into my tax dollars to build wireless networks on the borders of Indian reservations to cover nearby freeways and towns. But I get zero benefit when I’m out in those rural areas. It’s only universal in terms of geography, not people. I want a universal service to provide total access to everyone, everywhere; not just some people in some places.
My second proposal, therefore, is that a condition of dipping into a universal access fund is a forced contribution back to the access commons. All subsidised access points would be free to re-distribute their access. Violation of end-to-endism would also be prohibited; no NAT, filters, discriminatory QoS, etc. The countryside would have no lack of open WiFi networks!
Finally, there is no reason for one communications application (voice telephony) to be singled out for taxation. Otherwise, you should also be taxing email, IM, peer-to-peer, etc. — and we all know that has zero feasibility. I suggest the obvious answer (if you insist on cross-subsidy) is to tax access, not service.
The PSTN is part access, part service. It isn’t hard to do an economic analysis that decomposes its value into each bucket. Only tax the access part. Similarly, cable, satellite and DSL access would get taxed. The tricky part is sizing the tax in a way that is neutral between access technologies, and is fair and minimally distorting to the market. It should not discriminate against adoption of broadband. These seem to me to be mutually possible objectives.
This will also have an unexpected benefit. Regulatory fees have driven people off the PSTN towards VoIP. Likewise, access taxes will help to encourge the use of mesh networking and access sharing. For example, if I have a cable modem, my neighbour has a DSL line, and we both have a smart wireless router, then we should between us get fantastic uptime.
The revenue base of PSTN universal access funds is going to wither as quickly as PSTN users disappear. If we wait too long, only the weak and stupid will be left on the PSTN, and the political pressure will be enormous. Wrong choices then become inevitible. The time to change is now, while we can experiment and learn.
There’s been a lot of noise recently about the upcoming FCC meeting on VoIP. Many are concerned that the US government will unnecessarily encumber VoIP providers with irrelevant regulations conceived for a different world.
I believe that VoIP should be regulated, but not in the way that you might think.
Firstly, you have to understand that what is really being discussed is not VoIP per se, but rather interconnection with the PSTN. Stand-alone VoIP is for all intents and purposes beyond regulation. It’s too easy to move your client software downloads, proxies and directory servers off-shore, and encrypt traffic from prying eyes. Just ask the folks at Skype. The only alternative is to bring on the full violence of the state and try to outlaw VoIP entirely, as South Africa and Bangladesh have tried to do, with little success.
My position is that the consumer should make up their minds what they want to pay for, as long as there are no externalities or market failures involved. This means that consumers need the facts of what they are being offered clearly presented to them.
The Vonage terms and conditions are a 5800 word opus. Compare this to the dainty 1400 words of a Sprint PCS plan, or humungous 7700 epic from Nextel. Who really reads all this stuff?
The problem with VoIP operators is that the public may be expecting both the product features and regulatory baggage of traditional PSTN, but don’t understand that unregulated VoIP doesn’t guarantee these.
The Vonage Ts&Cs remind you that their system is unpowered (no electric, no calls); doesn’t guarantee 911 service; and outbound number portability is solely at Vonage’s discretion. Would the average consumer looking to slice a few dollars off their phone bill understand the significance of this? Even know that they should be looking out for this? Probably not.
What we need is a disclosure form similar to that for credit cards. The FCC’s job would be to regulate the form of how the offer is presented to the user — not the function, and definitely not the price through the imposition of random fees. Regulate market conversations, not markets.
I almost split my skull as my head crashed into my keyboard after reading this article on how WiFi operators are treading the well-worn path to failure by trying to offer unique content.
Even if you were stupid enough to try this, at the very least you would be thinking “location, location, location” — what to do near the WiFi access point you’re currently stuck at.
Does anyone really think that the decision to buy WiFi access is based on them offering a jazzed-up version of Business Week? Perhaps they do. I want to meet these people. I’ve got some great card games we can play involving a rolled up $50 bill and three upturned cups…
Over at Circle ID today is an article that reflects what I was saying recently on how branding is the solution to spam. In essence, someone else needs to assure you that the sender is not a spammer; the sender needs to have placed some sort of collateral at risk if they do spam; the collateral needs to be sufficiently large to make spam uneconomic; and the number of “someone elses” doing the assurance needs to be modest (otherwise we haven’t solved the problem if every message requires me to authorize the sender).
My proposal was that your ISP is the natural assurer that the sender is kosher. You (or a delegate) maintain a list of whitelisted ISPs. CircleID’s proposal is that the mail server is the source of assurance. Every mail server comes with a unique key; every mail server is paid for; the supplier of the mailserver maintains the whitelist of kosher mailservers.
This alternative system delivers the trust assurance into the hands of Microsoft and IBM. You trust the source of an email to be correct and non-spamming because the Microsoft brand makes you believe in their whitelist. Your ISP isn’t adding value over and above operating the system on behalf of Microsoft. Free open source software is neutered, because the very act of not having to pay eliminates any associated economic incentive not to abuse the product — there’s no excommunication event for spammers using Sendmail.
CircleID goes one step further and suggests that only recipients with a mail server from the same vendor will be able to use the filter, delivering a near-instant monopoly to Microsoft because of the network effect.
The problem with making the mailserver supplier the center of trust is that it is a one-size-fits-all solution. What may be spam to you might be perfectly legitimate to me. It is also not an exclusionary business in the same way that operating systems are. Hosting multiple OS’s and learning how to use them is a significant barrier to end users, hence the Windows monopoly. Adding multiple trusted sender assertions to an email is not an issue — the receiving mail server and client deal with it on behalf of the user. So maybe some hybrid approach will emerge.
My suspicion is that telcos have never even thought of decomposing where value comes from in communications, and which parts of the value chain they want to be in. They aren’t even aware that third parties are about to nibble at their lunch. They don’t know what their brands are supposed to be asserting to the user. Am I promising you clear unsullied personal connectivity at the IP layor or the application layer?
There is no reason your ISP has to be your access provider. My email server and hosted web space can be bought from anyone. (This subtle distinction is often lost — ISP is often used synonymously with access provider.) Access is slowly becoming cheap, fast and easily substituted. But the trust you place in a third party to filter your communictions to only things of interest is not easily substituted. It’s like wanting to marry a second wife on a trial basis before you divorce the first one. The world just doesn’t work that way.
Whether it is ISPs or software platform vendors that capture the trusted intermediary role, the telcos lose either way.
Have you ever thought about what happens behind the scenes when a telco decides to build a new network?
It runs something like this. Their market research folks go out and ask customers what sort of communications they will be performing, and how often. They come back with these vague labels (“surf the web”, “chat”), and the network folks guess how many bytes or minutes each interaction uses, and how many interactions there are. They then go off and scout around for a network technology that maximizes the overall throughput (bytes per hertz, switches per second, etc.). The network gets built, marketing set a price, and the punters roll in.
Now, there could be more or fewer customers than the business plan anticipated. They could be more or less active than planned. Their communications could be lighter or heavier than planned. If there are more bytes and minutes than the system capacity, then a whole new truck roll needs to happen to expand the capacity of the network.
So in short, the price gets fixed, and the capacity then varies according to demand.
But this isn’t the only way to approach the problem. What would happen if we turned the whole thing inside out? What if we fixed the capacity, and then varied the price? Whooaaa! What would that even look like?
In one sense, we’re already there. Your Internet access provider has a big pipe upstream. You can buy yourself a local pipe, with a nominal maximum speed. A menu of speeds and prices are normally on offer. But all it really gives you are more chances to sling a packet in the direction of the upstream pipe. Statistically, you do better than your cheap neighbor who bought a thinner pipe. You aren’t buying bandwidth, but rather priority as a emergent property of a probability distribution curve. Of course, this faux priority rationing is just a short-term effect: if the upstream service is lousy enough, and you have a choice, you may leave for another access provider.
Another way of doing it is through traditional QoS. Now, what you don’t need is to fiddle with packets on a per-service basis. This approach has been thoroughly debunked, at least with respect to wireline networks. It’s so easy to add more capacity to wireline services, it just isn’t worth building more silicon to ration stuff out. Repeat after me: optimizing for one application de-optimizes for all others (Om Mane Padme Hum, etc.). Only your end point should decide the priority of packets between services.
On wireless networks, the world of abundance has yet to arrive. Notwithstanding the recent theoretical research that wireless networks have infinite capacity, we’re stuck today with wireless technologies that simulate wireline channels. And the pipes are really thin. Demand can easily exceed supply. Adding capacity is slow and expensive. Even with wide-area OFDM “4G” networks, this isn’t going to change soon.
Who decides which people get priority access? Delegating priority to the end points only works to the extent that each end point can express relative priorities for its own communications. But everyone wants to be first in the queue when the collective bandwidth needs to be divided up! Thus priority of access is a different concept to priority of service. The difference is important.
Imagine a future where your laptop doesn’t just come with spiffy new Centrino wi-fi hardware. Instead, if embeds a wide area wireless network card, along with pre-provisioned nationwide low-priority access. The cost of lifetime service is built into the device. The huge volumes keep incremental prices low. Of course everyone knows that accident, obsolescence and changing fashion will guarantee a steady stream of replacement hardware business.
Perhaps the network capacity is gently scaled to ensure the great majority of users enjoy a dial-up level of connectivity in the great majority of cases. As appetites are whetted for always-on wireless, you up-sell higher access speed (or lower latency — a discussion for another day…). Sometimes at quiet periods even the cheapskates will get to experience great throughput. You under-promise and over-deliver.
So you don’t need crazy Enronesque dynamic pricing to ration fluctuating network demand. Static priority pricing schemes like those above are also an option. This fix-capacity-vary-price makes the economics align more with the reality of the physical network. It makes for more stable business cases. It expands the potential market to new low-end applications. It enables a crude form of price disrimination, because people wanting to run low-latency high-bandwidth applications with great reliability will need to outbid others riding on the same pipe.
In many ways it closely resembles the way airlines work: they don’t lay on an extra plane just because the flight has sold out. Every seat that flies unfilled is a lost revenue opportunity. On a network, every wasted hertz, electon or photon represents wasted capacity that can never be re-used. You can’t store bandwidth. It’s a perisable commodity with a lifespan of zero. Price it to sell!
Could it work? Time will tell…
The copyright industries have gone to war with their customers is the face of business model change. Will the same happen to telecom? How long will it be before you can expect a lawsuit from the CTIA for unauthorized use of your cell phone?
There are some obvious differences, and some less obvious similarities.
The telco content (for want of a better world) is owned by the users, not the telco industry. Apart from a few trivial sideshows like ringtones, communications is about user-generated content. People buy cell phones to say they’ll be late home for dinner, not to watch mobile TV. They buy AOL to find a mistress to make them late home for dinner, not to listen to music. They buy fax machines to file the divorce papers, not to receive comics.
This isn’t a question of copyright, though. It is an issue of control: who is allowed to get value from the system, and whether that value can be monitored and a toll exacted from it. For that, you need to understand where the peaks and valleys are of the value landscape. Which mountain passes will be guarded with regulatory and legal forts?
There are two places where the value gained is disproportionate to the number of bits transferred.
The first of these is the ability to locate someone else on the network: a directory. In traditional circuit-switched telecom, the routing tables that support the SS7 signalling are invisible to the end users. To participate, you had to be a member of the telco club. On the Internet, the DNS layer is much more exposed. Anyone can access the directory. Indeed, anyone can set up their own directory for services like dynamic DNS that were not envisioned by the original spec for DNS.
The arrival of ENUM will potentially enable the Napsterisation of telecom. ENUM makes public previously private routing data. The telcos just don’t see it in these terms yet. Napster enabled users to locate other users with certain properties with whom they wanted to exchange messages. Prior to Napster, I had no chance of finding anyone else with whom to exchange obscure Finnish folk rock. Suddenly, I can connect to fellow weirdonauts. The Internet rocketed in value, but the access service providers were not the ones making rent from that value (and neither was Napster).
Now, the telcos will try to turn ENUM into a private monopoly for regulated, approved, offical carriers. It will be painted as a cost saving mechanism — a mere technological efficiency. The problem is that this requires some wriggling around anti-trust rules to keep out third parties from the database. This can be done — there is no shortage of examples in other industries such as healthcare. Anticipate machine guns to be guarding the ENUM data.
You can imagine users attempting an end-run around even ENUM. As a thought experiment, what if there was a mass defection from the PSTN to Skype? Suppose Microsoft bought Skype, and the next thing from their hardware division was a Skype phone that you could give to granny. You register your regular PSTN telephone number in Microsoft’s Skype directory. A callback system calls your PSTN phone with an authentication code, so you prove you own that number. The Skype directory uses that number for routing Skype calls. ENUM isn’t even touched.
At which point does the telco dinosaur twitch and forsee its own extinction, and try some armed regulatory resistance? Is there a copyright in the matching of users’ phone numbers to their (telco-issued) IP addresses? The experience of others is that such associations can be proprietary property. But following number portability, who owns the phone number? Is it even “property” to be owned?
Expect directory access to be the Normandy landing of the IP revolution.
The second place where the battle will be fought is caused by a strange asymmetry in communications systems. If I’m receiving a signal from a central server, the most value that can be extracted from my access is the information in that single message. But if I’m transmitting, the same message is multiplied in value by the number of recipients. It’s more valuable to be a server than a client on the Internet. Even better, I can be both at the same time — everyone is a peer.
So it’s all fine and dandy to be able to get the IP address of someone through a directory, but if they can’t accept incoming requests for connections, their location has no value. Services like Skype spend a lot of effort doing an end-run around NAT and similar filters. They create outbound connections to intermediary servers, and then fake inbound connections. The terms and conditions of many ISPs explicitly forbid the deployment of servers.
This could quickly put the telcos at war with their customers. Both traditional copper line phone companies as well as cable companies want that rent from voice service. Municipal and mesh networks are rare, so you may have a choice of providers and prices, but no choice on terms and conditions. The final march to Berlin will be around whether you or your telco controls inbound connection requests to your own devices.
This isn’t a question of setting up a web server at home, attracting too much attention, and running up a bandwidth bill for your service provider. It is about them extracting economic rent from you, and price discriminating against activities that derive excess utility to the customer. Will they dare to seriously enforce their own contracts?
Expect a bloody fight and a lot of innocent casualties.
I was just doing some computations today on my fingers and toes on the scale of the problem of replacing voice revenue at a hypothetical telco totally unrelated to my employer:
Assume you retain 100% of access fees, lose 100% of service fees. Assume wireless and wireline go 100% IP-based.
Some numeric prestidigitation means that on average you need to find $600m of new business EVERY YEAR for the next 17 years. This needs to be conjured up from completely new intermediary roles and new services: access revenue is already factored in. And this is just to stand still. Oh, and it’s front-loaded since most of the people who are motivated to leave because they’re paying a lot will leave in the early years. The last person with a pure circuit PSTN line will be a sad case indeed. So you’re probably looking at having to invent a new billion dollar business every year. Without fail.
Holy schmoly.
Our friends at The Economist have a survey of telecoms (subscribers only) in their October 11 issue. They suggest that the current approach to regulation of legacy telcos is the worst possible method, except for all the others (with apologies to Winston Churchill). The argument they put forward is that in the absence of meaningful established competition, deregulation just creates virulent monopolies that can’t be displaced. The alternative of horizonal breakup into network assets and service operation is also dismissed. This is because you either have to create a nationalised network infrastructure operator (the equivalent of kiddie porn to the Economistas), or equally belligerent private access monopolies.
I beg to differ. Local and district/state government could be the custodian of such assets.
When national PTTs are privatised, it is reasonable to suggest that the local populace have already paid for the infrastructure, and already carry the debt burden of what is not yet paid for, so they might as well carry on owning it. Of course, operation of that infrastructure is clearly better done by a private company.
Localities still have to compete against each other for populace and business. Places that underinvest or overprice their access infrastructure will fall behind. The physical building and operation of the infrastructure can still be sub-contracted to private entities. Monopoly rent-seeking will cease to be the driving force. Monopolies deliberately under-supply the market. This has external effects on local business, education and employment that telcos don’t need to take into account. A more holistic economic model would probably suggest a much higher (and cheaper) level of pervasive service.
There is also a strong precedent in the roads system. Few would argue that we should have a pervasive metered private road system, depite that system having to reach every door, much like Internet access needs to be.
The usual argument against commie concepts like communal ownership is that such endeavours are captured by political interests, fail to respond to market needs, and are grossly inefficient in implementation.
In this case, these are red herrings.
The local government will only control wholesale rates, with free market entry to any retailer. Political favors just won’t be possible. Ideally the national law should mandate universal service for any municipal solution, so nobody is left with the cost of service without getting the benefits. Corruption in public life in public contract tenders is not a major issue in most developed democracies, at least when measured against private corruption.
Few disagree with asphalt as the target technology for road building. Likewise, in urban areas, fiber is clearly the winning technology. In rural areas, wireless is often more attractive. Multiple technologies can still compete, just in a mosaic of municipalities. There is no need for interoperability at layers 1 and 2 of the network stack.
Keeping retail separate also means that innovative pricing, bundling and marketing schemes are still possible and likely.
Of course, in the USA municipal broadband efforts have encountered fierce lobbying resistance from private telcos that would lose their cosy monopoly or dupololy rents. So you need to choose your timing, or have friends in high places with a lot of political will.
And while I’m picking bones with The Economist, I’d also disagree with the following statement:
As this survey has argued, the three trends currently driving the telecoms industry are wireless, broadband, and the convergence of voice and data, telecoms and information technology.
Maybe it’s just a matter of wording and emphasis, but there is one, giant, over-riding trend: end-to-end networks and the separation of connectivity and service. Everything else is a corollary of this. I’m surprised that they can have a telecom supplement on structural change in the industry without once mentioning this driving concept.
PS — If anyone from the Economist is reading this, send job offers for a new improved telco correspondent to the address on the right…
The Register has an article on Google from the barbed quill of Andrew Orlowski. It correctly points out that Google is really two businesses: a media buying agency, and billboards that attract eyeballs into the vicinity of adverts. Andrew thinks that Blogger is their sole billboard: I would say that the search site is their main billboard. I would also personally say that Andrew is wrong in stating there is no such thing as a search engine business — clearly, the search billboard has real stand-alone value. No matter. The important bit is this quote:
So apart from Blogger.com, Google doesn’t own its own billboards. It can pursue either one of two strategies, or more likely a combination of both, to remedy this. Firstly, it can buy or develop more of its own billboards: features which make Google’s own pages more ‘sticky’. By definition, Google’s search results aren’t sticky: you’re on your way somewhere else. Email, a dating service (Google reportedly bid the bubblicious Friendster.com the San Jose Mercury reports today) and beta services such as Froogle could all come into play, turning Google into a mini-portal.
I would be tempted to argue that the flaw in Google’s seach engine billboard is their failure to develop stickiness. Search results are not tailored to me. Indeed, there is no feedback loop at all. Recently I have been searching for hotels in Helsinki (don’t ask) and car rentals in Dodge City (I really mean it!). Googlespam link farms have made it very hard to find what I want. If Google started to ignore domains that people don’t click on in the results page, the search would get better over time, not worse.
This is probably a limitation of technology at the moment. Google is a marvel of the information age as it is. But they seem to have lost their way about two years ago when the main service stopped improving, and all their energy was diverted into side services that the customers rarely use or don’t value much. (Anyone having images of voice command systems here? Screen savers? Celebrity voicemail greetings?)
Ultimately, this is a question about their business model, and it has a very close parallel with the telco industry. Google faces pressures to disaggregate the components of its business. Their very own advert syndication system (AdSense) separates the service from the underlying transport of the search engine. One component involves staccato user interactions (think: telephone calls, search results). The other is a long-running relationship with advertisers.
It is possible to build a customer intimacy business model even in the space of transient search results and phone calls. Once the system learns about you (e.g. I never answer calls from strangers between midnight and 6am, I never click on the domain helsinkihotelstoday.com), it can adapt and intelligently act as your agent. Leaving the service is like getting a divorce: it’s painful to train your mistress to adapt to all the quirks your wife got used to decades ago.
I wouldn’t buy Google shares when they IPO, because their business is too unstable. But then again, I’m up to my neck in telco stock, so you should discount my advice accordingly.
Via boingboing and smartmobs comes this gem on bluejacking. In a nutshell, this is a perversion of Bluetooth, spamming people in your physical proximity with random unsolicited messages. Read it. It’s quite funny, in a juvenile way.
This shows a lack of forethought and insight by the developers of Bluetooth. As I wrote last week, identity collateral is an important part of any open communication system that is going to socially scale. You need to be at risk of losing something if you abuse the system. Bluetooth lacks that.
I can imagine an evil business opportunity. I set up my PC with a Bluetooth card in an office above the entrance to a heavily-trafficed Tube station in London, for example. Every passer-by with a Bluetooth phone gets a spam - thousands of them. Might just be a simple message, might have a URL of a website to visit. As Bluetooth becomes pervasive, so does my spam - and you can’t avoid it without turning your phone off.
How could you have built identity collateral into Bluetooth? Well, it’s tricky, because you have no guarantee that the device is attached to any sort of service contract or provider — so there’s nobody to complain to to de-provision the device if it is abused. Blacklisting the device after being spammed isn’t going to help you, particularly if there’s no trusted serial number to rely on (just a user-assigned device name) and no means of sharing your blocking preferences with the as-yet unspammed.
(I’ve skimmed the Bluetooth spec and there’s nothing I can see in the link layer spec that helps; the Universally Unique Identifiers don’t cut it either. If I’m committing a gross libel of the Bluetooth authors, please let me know.)
One approach would be the “end-to-end” distributed way: just like with email, put a Bayesian filter into every handset to guess whether to accept incoming messages. Not the obvious way of best using limited memory capacity on a handset. Some form of identity collateral by proxy is possible: the sender must send a digitally signed (by a service provider) phone number or handset ID in the message. If you get spam, then someone’s phone service gets cut off. (Without the certification, you could give anyone’s number.)
In fact, this highlights a non-obvious feature of telephone numbers. Unlike domain names, only registered service providers can ask for one. The artificial scarcity means that they are only rented out as part of a commercial relationship. The minimum cost of getting such an identifier is buying a pre-paid handset. (Post-paid handsets require a credit check and thus verified personal identity, limiting abuse to one per person per service provider.) Abuse isn’t economically worth it.
The telco business model lesson? There is potentially money in making communications systems socially scale. Handset vendors don’t even think about the operational abuses that their technology can be put to, whereas for telcos dealing with these customer care issues is (or ought to be) second nature. As with the Liberty Alliance spec, the operational issues of trust and identity have turned out to be much more complex than the technology issues. There’s a danger of throwing the baby out with the bath water if we try to build naive replacements for the PSTN on open networks - be they the Internet or the Sidewalknet.
Appeal to readership: I know the concept of identity collateral is not an original Martin thought — but I can’t remember whose blog I found the idea on, and can’t discover it in Google. Post me the reference, I’ll update the entry.
Om Malik has some commentary on the rise of the cablecos as powers in the voice telephony space. Of course, they need briding back to the legacy PSTN world, hence the need for alliances with the likes of IDT.
The real cableco play here is to create an ever-increasing pool of on-net users (i.e. calls that you route between your own customers or those of cableco partners). The revenue associated with bridging to the incumbent telcos will slowly dry up, and these “partners” will find themselves left high and dry.
Unfortunately for the cablecos, this won’t make them rich. The cost of telephony will continue to fall, because most of the traditional copper and TDM equipment was laid and paid for long ago. Any cost savings associated with VoIP won’t give cablecos an edge - most of the cost issues now come from customer acquisition, billing and care. Prices will just drop further in response to increased competition. The profit pool will just get re-allocated onwards to the customers.
Long-term defensible value doesn’t usually come from arbitrage in the relatively static world of connectivity. Arbitrage only works when there is change happening faster than information about the change can be disseminated and reacted to in the marketplace. For the slow-moving, you need a proprietary technology (e.g. Windows), a proprietary name space (e.g. AIM) or even better some proprietary data (e.g. credit rating agencies). For the fast-moving, you can create temporary advantage through innovation in devices and services - think of iPods or Google.
“Fast-moving” is not something that should appear
in the same paragraph as “telco”.
So what should the cablecos be doing? It’s that same old story: you need to move on, create something that the PSTN can’t do. For example, notwithstanding the high-pass filters built into many landline handsets, you could squeeze out better sound quality with a VoIP adapter. If you’re Time Warner, partner with device makers to offer the AOL phone, compete with presence indicators of all your buddies. If you’re Microsoft and want to make your Comcast investment work, turn the Windows Media PC into a home media and telecom server - and be able to access your address book, notes, files, calendar, voicemail, etc. from anywhere on the Net.
Someone soon is going to puncture the public’s perception that telephony = PSTN, and at that point there’s going to be a lot of trouble. I don’t know about you, but I’m looking forward to the show.