Full Article: The Long Tail, Debunked?

The fifth Telco 2.0 Executive Brainstorm November 2008 continued its theme of business model innovation at the intersection of telecoms, media and technology by welcoming back Will Page, Chief Economist at the MCPS PRS Alliance, a copyright collection society that represents over 50,000 songwriters and 5,000 publishers.


Will took the opportunity to present, exclusively to Telco 2.0, new research – based on an unprecedented analysis of digital music sales data gathered over a year – that opens to question the recieved wisdom around the ‘Long Tail’ theory, and helps to re-define what it actually means and for whom. The presentation created quite a stir at the event, in the media and blogosphere. Here, Telco 2.0 discusses at length the presentation and the reaction to it with Will Page.

Telco 2.0: At previous Telco 2.0 executive brainstoms you’ve covered file sharing, the economics of two-sided markets and now the long tail. Coming from outside the Telco world, how useful do you find the events?

Will: Very! It’s ironic that we live in a world of convergence yet too many of the disparate camps like to remain in their pigeon holes and preach to the converted. What Telco 2.0 events do, by allowing someone from a copyright collecting society to speak openly to an audience predominately made up of ISPs and Telco operators, is invaluable to both content and connectivity businesses.

You can see the importance of this more and more now. We have an truly awesome CEO of UK Music (the newly formed music industry trade body) in place now – Feargal Sharkey – and he’s spending an increasing amount of time at OfCom. Two years ago, that simply would not be happening. So, in many ways, what the Telco 2.0 Initiative does in terms of brining different industries together is ahead of the curve.

Last week’s Long Tail presentation was a good example of this. I first met Andrew Bud, Executive Chairman of MBlox (and now Chair of the Mobile Entertainment Forum) and a key collaborator on my analysis, at a Telco 2.0 event back in October 2007. I can’t think of another ‘platform’ where our paths would have crossed, even though both our businesses share surprisingly similar characteristics.

Telco 2.0: For those who follow Telco 2.0 but missed the presentation, could you bring them up to speed on what exactly you, your colleague Gary Eggleton and Andrew Bud have been working on over the past few months?

Will: Sure, it’s worth going back to the beginning as it’s been an interesting journey. Firstly, like so many others, I read the original Wired article on the Long Tail in December 2004 and was genuinely inspired by it. For the next two years I was active in the blog-to-a-book website and have to credit the concept as being one of the principal reasons behind moving to London to work in the music industry in the Summer of 2006, ironically when the book came out.

I guess the presentation I gave last week reflects what I’ve learnt in the two years since from working in a collecting society, an organisation which by default is in the long tail business. Indeed, the Performing Right Society (PRS) has been dealing with long tail markets since 1914. The whole purpose of constructing and offering a collective licence is that it doesn’t matter if it’s a song is a hit or a niche, all the tracks have been licensed under a blanket agreement.

Given the clear relevance of collecting societies to the ‘long tail’ debate, I was surprised to see so little mention of them in Chris’s book – or the blogs that followed. For example, our US equivalents’ ASCAP and BMI don’t appear once in the book’s index.

For those who weren’t there, let me break the presentation down into three parts. It began by looking at the evidence in terms of actual historical data. I drew upon a great expression that I learned whilst in the Government Economic Service, which is to always strive for ‘evidence-based policy making’ and resist the temptation of ‘policy-based evidence making’. Increasingly, when I hear those words “here’s another great example of the long tail at work��?, I’m inclined to expect that claim to lean towards the latter of the two.

I made the point that looking at volume-based Rhapsody data, which much of the long tail application to music has been ‘built’ around, is like a glass half empty – at best. We need to also consider value, and by that I mean not just retail spend, but marginal profitability in terms of what gets back to the artist and songwriter, and also ‘displacement’.

One achievement of my two years at the MCPS PRS Alliance is to get ‘displacement’ into the everyday lexicon of the UK music industry. Is every digital track sold to be celebrated (a P2P user now gone legitimate)? Or regretted (a £9.95 album sale lost)? The reality of the the long tail is now being uncovered by many stakeholders in the music industry’s head (hitmakers) and tail (poor sellers).

The second part revisited ‘histograms’ as a way of plotting the long tail. Andrew Bud, who’s been like a Professor to me throughout this project, put me onto a fantastic book published in 1956 by Brown, entitled ‘Statistical Forecasting for Inventory Control’, and which described the importance of log normal distribution as an analysis method.

This concept is not radically new and is still discussed today (for example, Chris Anderson refers to log normal distributions in his speeches and blog too). But for me this book was fifty years ahead of its time.

Using this approach our team constructed log normal intervals and plotted an unprecedented amount of digital music data over a significant time period. The basic shape of consumer demand for digital music clearly fits the log normal distribution…��?with eye-watering accuracy��?. It was really striking. There are many new schools of thought, but the old rules seem to hold truest.

The difference between a Pareto-style distribution and a log normal is neatly summarised by Chris Anderson in his recent response to my analysis, below:

��? …The two distributions look similar at first glance, and you have to plot them log-log (or fit them with a statistical package) to tell the difference. Long Tails are “heavy-tailed��? distributions, where a lot of the total volume is the tail, while lognormals are more like the classic top-heavy hit distribution…��?

The third part of my presentation at Telco 2.0 last week concluded with two important slides. The first one plotted two heads. The first ‘head’ was the concentration of tracks which sold very little or none at all. It questioned whether the net revenue generated from these tracks cover the real sunk and commission-based costs for a.) getting the song there and b.) getting money back out of the system.

The second ‘head’ was to show the effective average revenue per track in each ‘bin’ (or statistical grouping of data). This was a crude averaging method but it proved highly illustrative. The inequality in revenue between hits and niches was jaw-droppingly stark, justifying Andrew’s observation that “in this tail, you starve��?.

For example, we found that only 20% of tracks in our sample were ‘active’, that is to say they sold at least one copy, and hence, 80% of the tracks sold nothing at all. Moreover, approximately 80% of sales revenue came from around 3% of the active tracks. Factor in the dormant tail and you’re looking at a 80/0.38% rule for all the inventory on the digital shelf.

Finally, only 40 tracks sold more than 100,000 copies, accounting for 8% of the business. Think about that – back in the physical world, forty tracks could be just 4 albums, or the top slice of the best-selling “Now That’s What I Call Music, Volume 70��? which bundles up 43 ‘hits’ into one perennially popular customer offering!

This chart really drove home the theme of the presentation: what does the ‘long tail’ actually mean, and for whom?

If you’re a for-profit aggregator, it means one thing, if you’re an individual copyright holder, it means another. Again, this is something the debate has largely overlooked to date, yet everyone ‘down here on the ground’ increasingly recognises it.

As a not-for-profit membership governed collecting society, I’m extremely fortunate to be in unique position to make a balanced interpretation about the facts, and what they mean to both sides – individuals and firms. My interpretation is in no way gospel, but I can at least build an argument that’s based on evidence from the coal face.

My argument, in summary, was that the future of business is definitely not selling ‘less of more’. Scale matters. To tee up the interactive debate with the brainstorm participants I concluded with a final slide posing the dilemma facing firms in the content value chain as regards their investment strategies. Thanks to the way Telco 2.0 allows the participants to ‘blog live’ with the presenters using lap tops and special software [Ed. – we call the format ‘Mindshare’], a mountain of comments and questions flooded in.

Telco 2.0: Of the (literally) hundreds of questions the audience threw at yourself, your colleague Gary Eggleton and Andrew Bud, which would like to answer in more detail here?

Will: Firstly, I’m genuinely grateful for that ‘blogging facility’ you have at Telco 2.0. The day after the presentation, your CEO emailed me every question, idea and challenge your audience threw at me – that’s a fantastic facility, a “free lunch��? of excellent feedback and advice. My thanks to everyone who posted.

Now, I think there are three themes which I can draw from your excellent interactive facility at the conference and expand on here: (i) the black market, (ii) digital inventory costs and (iii) scarcity.

1.) The Black Market (P2P)
There were lots of really insightful questions on this topic which can be summarised by this one: ‘is the P2P market more or less concentrated around hits than the legal one?’

To help answer it I would direct readers to a now infamous paper I published with Eric Garland, CEO of Big Champagne, titled ‘In Rainbows, On Torrents’ (pdf).

My hunch, based on the evidence we presented in that paper (pointing out the 2.3 million illegal downloads of Raidohead’s new album when it was also available ‘for free’ on their own website), is that the black market is even more hit-centric.

As Eric would argue, popular music is popular wherever it is popular, in that you can’t be a hit on iTunes without being a hit on Bit Torrent …and vice versa. For further reading, I’d suggest the sociologist William McPhee’s groundbreaking theory of exposure, found in his 1963 book ‘Formal Theories of Mass Behavior’.

2.) Digital Inventory Costs
These costs are often overlooked by those claiming the long tail is a ‘panacea’ for artists. Making recorded music has many independent costs, some have gone down, others have gone up (what economists call ‘cost disease’).

Similarly, there are administrative costs to uploading tracks onto digital sales platforms and getting the money back to the creator. For example, indie ‘niche’ labels need ‘aggregators’ before they can join the main digital music platforms, which is a wholesale market, just like in any other business, digital or bricks and mortar. The same old rules of transaction costs and economies of scale apply there too.

I wanted the audience last week to consider another old rule of economics – cost-benefit analysis. Do the net benefits outweigh the costs (both upfront and commission based) of joining the long tail? It’s a simple question to ask which few bother to do and hence it’s infuriating when you read propositions like ‘all you need is 1,000 true fans’.

3.) Scarcity
There was a wonderful comment from one anonymous participant in the room who said:
“…Scarcity forces a ‘competition’-like structure to pass the cut-off point, which paradoxically creates value by increasing the effort of content suppliers to win….��?

This really sums up the point of my presentation. What I said was not particularly new, in fact its basic common economic sense, about which we sometimes need a reminder. This quote points to where I’m going to take the research next. Not the ‘head’, nor the ‘tail’ – but what happens to the ‘body’?

My hunch is that without scarcity, the body is underexploited. The quote provides context for a point I made on stage and in an article in The Register: “Is the ‘future of business’ really selling less of more? Absolutely not. If Top of the Pops still existed, it would feature the Top 14, not Top 40.��?

Telco 2.0: You’ve definitely got the debate started as there’s been significant press coverage since the publication. How have you found the reaction in the media by those who were and weren’t there?

Will: Tricky. Many journalists have agendas – some you agree with, some you don’t. Sometimes the meaning of your work gets lost in the differences within those agendas. My agenda is an academic one, like the great Scottish philosopher Hume would of wanted, one of conjectures and refutations. Let’s take a theory and put it to an impartial evidence-based test.

Firstly. it was great to see Eric Schmidt, CEO of Google, putting forward a strikingly similar argument to myself on McKinsey’s website recently.

In addition, I was pleased to see The Register picked up on the role of a collecting society, an institution that receives surprisingly little coverage in Long Tail debate, yet has pioneered the creation of long tail markets for musical copyright through patent pooling and blanket licensing for almost a century.

It was great to finally get a mention on Chris Anderson’s blog – given that I published my first set of long tail statistics (pdf) back in November 2006!

Given that the source of my data cannot be disclosed at this stage and the slide deck from the Telco 2.0 event cannot be circulated (and I’m genuinely grateful to those in the audience for understanding this point), I thought he did a pretty good job at blogging about a presentation he wasn’t present at. Hopefully this interview will help him fill the gaps.

However, I still think he’s focusing his arguments on the less-relevant volume-based data, and not looking at value in all of its definitions, Or, as an impatient CEO might say ‘show me the money!’ Volume-based discounting has, is and always will be prevalent in any market, be it online or offline. It is a simple and widely accepted fact, and once you accept that it becomes increasingly difficult to hold a conversation about why the future of businesses is selling less of more.

On the downside, one of your participants published a blog article that was so far off the mark, it made me wonder if he was actually paying attention (or more likely, understood the complexity of the music industry). For example, he describes my focus on individual sales as very ‘old economy’. Yet the erosion of the unit value of musical copyright is the biggest issue facing the membership of the MCPS PRS Alliance.

Why? Because we’re a membership-governed not-for-profit organisation that licenses, collects, processes and pays out royalties for our 50,000 individual song writers. He also goes on to say I used a data set where the concept of margin is irrelevant, which is the completely reverse of what I actually did. I presented data and then introduced the concepts of marginal costs (the real costs of managing and processing digital inventory) and marginal benefits (how much of that unit value actually gets back to the creator) from the outset.

Trying to have a balanced debate about the long tail, and avoiding knee jerk reactions and hysterical claims, is hard, very hard! Everyone immediately becomes an expert in a specific market or a statistical rule that they actually know relatively little about.

I feel for Chris on this who pioneered this concept, as he must have had to stare this problem in the face for a lot longer than me. In the music industry, which has experienced the force majeure of disruptive technologies like no other, and for over a decade now, you get a little tired of arm chair critics telling you what to do with the benefit of hindsight, and little understanding of what options were available at the time.

Nevertheless, as the legendary Peter Jenner would always say every time meetings between the content and connectivity industries collapsed into disagreement and disarray the important thing is that we keep all the parties talking, exchanging ideas, evidence and advice.

Telco 2.0: Finally, how transferable is your work? What does it mean for other areas which Telco companies are looking to get involved in, such as Television, Film and Books as well as applications?

Will: Very! Discovering a log normal distribution in one area of media provides a template for evidence-based gathering, interpretation and decision making in others.

I’d stress caution though – you need to order the questions correctly. Just as when you look at international comparisons in order to learn lessons for domestic issues, you need to ask ‘what works over there’ and ONLY then ask ‘of what works over there, and what could work here?’.

There’s been far too much decision-based evidence making to date along the lines of “the long tail must work so find me a great example of it working��?. That’s in no way the fault of Chris Anderson. He (like myself) goes to great pains to correct people’s knee jerk reactions, but it’s standard fare when a new economic theory comes along, people get a wee bit hysterical about it.

I think that the most important lesson to come out of this work is a real back-to-basics question: is scarcity a constraint or is it a discipline? I think that you can ask this question from the outset, regardless of what type (and what size) of media inventory you intend to carry across your network.

Finally, I’d like to reemphasise the importance of impartial evidence-based analysis. My work is not about trying to prove anyone wrong. I’m looking at how well their case stands up when presented with evidence. On that note, perhaps it would be apt to end with a suggestion to those proponents of the long tail theory, by drawing upon a quote from the late great John Maynard Keynes: “When the facts change, I change my mind. What do you do, sir?��?

[Ed. – After the interview we asked the Telco 2.0 analysts to comment on the transferability of Will’s analysis and the ‘so what’ for telcos:

Chris Barraclough, Consulting Director: Value comes from: catalogue breadth/depth + distribution (which includes searchability and multiple customer touchpoints, including affiliates) + evaluation. Amazon has market power because it works hard on both distribution and evaluation. You can exploit a longer tail than your competitors if you can a.) lower your cost base further than them (ie afford to carry more inventory than them) and b.) price cleverly (link price to volume so that lower volume items are priced higher).

Martin Geddes, Chief Analyst: We must appreciate the uniqueness of different content types and distribution networks. On the latter, for example, iTunes differs from last.fm due to pricing policies and content recommendation systems. In terms of content types music is weak in metadata (people don’t write much about individual songs), unlike richer media like movies (where there are lots of reviews and information on the participants), games, software apps, and TV shows. So, my advice is be careful about extrapolating lessons between different media and indeed even between sub-genres within the same medium. Sport, porn, and news video all have very different dynamics, for example.

The big ‘so what?’ for telcos is that a lot of ‘long tail’ content may have no commercial value, but may have considerable social value to users (e.g. photos of your kids). It still needs to be transported. This makes it all the more important to cater not just for ‘high QoS’ material like streaming HD movies, but also to be able to dynamically subload other content. Watch this space for interesting developments in this latter category…!

James Enck, Senior Associate Analyst: It would be useful to analyse how the value of content changes over time. Van Gogh didn’t sell much during his lifetime, Grateful Dead fans favor bootlegs rather than studio recordings, and we all know the story of the Arctic Monkeys. In other words, it’s conceivable that content moves from the tail to the head over time, and those who don’t spend time in the tail will always be surprised at what appears in the head.

The depressing truth for telcos is that replicating the head does nothing for differentiation. Moreover, if content strategies are geared to churn reduction rather than incremental revenue, then what disincentive to churn is there if all competitors have the same 4,000 film library? Long tail content can be highly appealing as a differentiator if it maintains a local flavor. www.pod3.tv is one example in the UK, which to my knowledge, no telco has sought to engage with. I’m baffled as to why Telekom Austria seems to have stopped the Buntes Fernsehen project, which to my mind was a very interesting way to differentiate on long-tail content in a way that’s highly relevant to a local customer base.

Keith McMahon, Senior Analyst, Content Distribution 2.0: The key message for me is that there is no silver bullet in merely loading content onto the net. The challenge is beyond simply distribution. The promotional aspects will be a really hard skill for telcos to replicate over a wide range of content. They are probably much better partnering, developing ‘two-sided’ enabling business models and shifting the demand risk to parties who know better.

Alan Patrick, Senior Associate Analyst, Content Distribution 2.0: I did Mechanical Engineering at University and studied inventory management theory. The thing I recall is that nearly every inventory based demand curve was Log Normal. The big issue in the online world is the lower transaction costs which supports a “positive returns��? power law dynamic, ie. the big get bigger. This drives an increased rush to the ‘Hit Head’. In other words any service which had a long tail distribution would rapidly move to a bigger hit head in any online world.]

Full Article: Telco 2.0 Business Model Map: Links & Q&As

At the February 2007 Telco 2.0 event we presented our Business Model Map. Links and Q&As are reproduced here to help clarify the map and its implications.

You can read the background to our Telco 2.0 Business Model Map in the following four-part article series: Introduction, The axes of the map, The business models, and The consequences.

In a nutshell:

  • Network operators are delivery/distribution businesses: they deliver valuable bits from A to B.
  • There are many ways of delivering those bits. For example, a video could be sent on a DVD, via IPTV, a peer-to-peer download, streamed from a content delivery cache, etc.
  • The map documents these distribution channels for valuable bits. Which ones are you as a telco going to invest in?
  • Each one is assessed on two criteria: does payment automatically flow between connectivity and the content/service (“commercial integration), and is the delivery network hard-wired to that particular media delivery, or general purpose (“technical integration).

Can you clarify the Embedded Business Model?

It’s the reverse of the handset subsidy model. You buy a device which comes with connectivity embedded – a bit like those self-adjusting clocks. As we in-fill the device market between mobile handsets and PCs, we’ll see a huge range of appliances, many with relatively short lifespans (under 2 years). Embedded a fixed period of connectivity into the device eliminates a huge amount of billing and revenue management cost.

In other words, as we have more tailored devices deployed, we’ll see many variants on how the money flows between the hardware, service and network. In this case, it’s the hardware sale that funds the rest.

Is it fair to read the matrix and bubbles in the following way: no single new business model, nor a combination, will compensate for cash flow lost on traditional business (bubbles are revenue right?)

That depends on how optimistic you are that operators can grow the “new opportunity” bubbles. In the chart their sizes all add up to the same amount in every scenario, so I’m just showing relative (not absolute) changes.

We do not understand the “protection��? zone

You try to protect the existing revenue model of vertically integrated telephony and messaging. Do this by gentle feature innovation, extending the commercial footprint (e.g. non-geographic numbers, short codes) and resisting competition.

We deliberately don’t call it “product innovation” because it’s not about seeking new online services products (best left to Internet players) but extending the lifespan and economic model of the ones you’ve got.

Q&A: Tiered access and QoS

How do we deal with the situation on Tiered service (Paris Metro model) where disputes arise – proof of QOS etc.?

The point of systems like Paris Metro Pricing is that they retain the cost structure and efficiency of best-effort delivery. If you don’t make a promise of guaranteed delivery/capacity (just better statistical odds of delivery), there’s no dispute possible or proof needed compared to the “reserved capacity” model of QoS.
You can read more at here.

Can you detail the Tiered commercial model?

Think “multiple virtual dumb pipes”. Today my computer talks to the Internet on device eth0 or wlan0. What if there were five “virtual” interfaces (“vnet0″ etc.), and I could choose any one of them, but some had higher cost/lower contention? And I could dynamically switch between them? Think of it as like TCP/IP’s back-off mechanism, but you step up and down priority levels as well as transmission rate.

A lot of the complexity is going to be hidden from the user. Say Skype get together with a bunch of carriers to offer Mobile Skype. Rather than re-write Skype to fit into IMS (no chance), they use a high-priority virtual pipe for the VoIP part, and a low priority one for the IM/file transfer.

One day, you might have hundred or even millions of “virtual Internets” (“VWANs”, to mirror Virtual LANs). A bunch of Xbox gaming devices might talk to each other on a virtual network, safe from denial-of-service attacked from the general Internet, and with the priority and latency they desire.

Does traffic prioritisation have some value as a business model at least in the short term?

Short term, yes. Enterprises, absolutely. And there will always be some friction in the system between different generations of technology that some network smarts will be needed to smooth over.

The comment was made that it was desirable to move away from deep packet inspection – what model would be better?

DPI isn’t all bad. For example, I’d retain it as a means of identifying “bad” packets like DDoS attacks, spam, fraud management. However, as a means of value-based pricing, it’s kaput. Even from a physics standpoint, it’s a non-starter: move all the bits from the optical transmission domain to the expensive and slow electronic processing domain. I’m sure Cisco would like you to do it, but it doesn’t mean it’s a good idea. You’re just in an arms race with your own customers to cloak the traffic. A better model is either (i) spend the money on more capacity and stick to all-you-can-eat broadband, (ii) go to the Paris Metro Pricing style of tiered dumb pipes plus offload the heavyweight traffic to CDNs, (iii) lock down the edge devices to eliminate the problem at source.

Was the Paris Metro Pricing a way of Bell Labs ingratiating themselves with the new French owners

Bien sur! Err, no. The research was done back in the days of Bell Labs, before croissants came on the cafeteria menu.

Q&A: Low vs high integration models

Is the future in the radical advances of end-user devices (high memory, CPU power, client-based apps)? Does this accelerate the ‘dumb pipe’ scenario for infrastructure?

Yes – ever more applications become possible to deliver over a general-purpose network with a general-purpose handheld computing device.

How do you move from a world of one size fits all broadband into a slicing and dicing world? If broadband is unlimited, who would want to bundle it into a device? Will users be willing to move off their unlimited plans to tiered plans?

As “legitimate” P2P and video download traffic moves onto content delivery networks and separately packaged offerings (that avoid peak hour, keeping capex down) the price of “unlimited” broadband will probably rise. For the average email/web/Youtube user, they won’t notice a thing. But the broadband “hogs” most certainly will. We’ve already seen the first steps in the UK with BT wholesale going from an unmetered to a metered product. (I’d expect a congestion-based pricing scheme to be next.)

Do consumers really dislike vertically integrated offers? It is the cost and service that matters. Given cheap offering with good service, one probably will buy everything from one place.

They love them! However, technical vertical integration makes for incredibly slow progress (e.g. voice telephony network) compared to “dumb pipe” applications (e.g. social networking). The trick is to preserve some of the commercial integration of service and delivery without the technical integration.

Can you really integrate commercially without any technical integration/leverage? Explain…

Sure – you’re just accounting for the packets differently based on their origin and destination. There’s no magic – as long as there’s some compensation that flows between the device, content or service and the underlying connectivity provision, then you have some form of commercial integration.

Packaging service/devices etc with connectivity is a good answer, but the worry is that the examples James gave are actually about startups that help users unbundle.

These two trends are not mutually exclusive. Artificial bundles will be torn apart, but where they create convenience to users then they’ll buy up packaged offerings.

What are the fundamental technical drivers of business model & structural change? For example, does Moore’s Law and smarter end-devices MANDATE that a dumb pipe model will evolve?

Blame Claude Shannon, the father of information theory. The information carried by a signal is independent of how it was delivered. Dumb pipes are less complex than smart ones, all other things being equal. So if you can get the application to work over the cheap pipe, why pay more for a specialised one?

Q&A: Telco culture and organisation

How bad does the crisis have to get internally/externally at the telcos before they actually embrace change? Do we need to go through a real period of agony before the willingness to change becomes real?

We’re already there. The ISP industry is already a world of hurt except for some niche players, with exits and M&A on the up. The crisis is as much for the users as for the operators, who can simply delay investment. The users miss out of the experience they require, and the economy never gets to grow new industries.

What is the future for incumbents

In the short-medium term, they’re OK as there are such large barriers to entry and a regulatory tar-pit to negotiate. In the longer term most will disappear (consolidated into a few mega-operators with the necessary scale). Many will exit the application services space or become effectively different businesses (e.g. BT’s global services, where the network is just an enabler, not the product). Probably a similar story to what’s already happened in other utility and media industries. British Gas used to retail cookers – but no longer.

What should telcos do today? What’s the first big tipping point?

Decide which core Telco 2.0 strategy you want to follow: diversification, protection, platform or pipe. Then exit the non-core stuff (but preserve the revenue streams via partnerships). “Tipping points” are a bit of a semantic illusion, it’s a continuous process of evolution.

In terms of broadband and mobile who owns the customer today?

The customer owns the customer. At best, you own some expensive glass and electronic boxes, plus some databases of variable quality. If “2.0″ is about anything, it’s the users being in control of their destiny.

What might the cultural changes be for Telcos to become more innovative?

Innovation does not equal invention. Innovation is simply applying inventive ideas to bring those things to reality. Telcos are plenty innovative enough, just not in end-user services. Most of the innovation is in deployment and operational management of networks, which you never get to see. The evidence is in how well they’ve adapted to extremely rapid technological change. I don’t think it’s possible to create an innovation or invention culture in a mature organisation which doesn’t have one.

What’s the underlying structure out of your perspective which makes change so difficult? No moralistic judgement — a structural explanation

I don’t think change is difficult. The telecoms industry has undergone incredible change in the past 25 years (mobile, broadband, optical transmission, regulatory) – as much if not more than most other industries. How much has Google really changed? You underestimate telcos.

Why are Telcos so low in R&D spend?

Shareholders have done really well out of gatekeeping other people’s innovations being distributed over their pipes. Maybe telcos are like travel agents: high turnover as you act as a handling agent for 3rd party products, but that artificially inflates your revenue and expectations for research spending. Given the efforts of the Internet companies and telco equipment vendors, overall there’s no lack of R&D in communications.

How do telco’s strike a balance between commercial innovation and innovation in technology which is often their bargaining card/source of competitive advantage?

Easy – exit the services innovation space and partner. Focus on integration and overall user experience across multiple services and payment/delivery methods.

Q&A: Peer-to-peer & video distribution

Can you expand on the telcos’ exploitation of peer to peer and content delivery networks?

“If you can’t beat them, join them.” Telco IPTV isn’t going to deliver the Lithuanian shows I want my kids to watch; or give my brother his dose of British TV out in California. There are going to be many different aggregators, editors and recommendation engines. Operators need to make it easy and advantageous for those video distribution networks to minimise backhaul costs and contention. Today there’s no incentive to make your P2P application operator network friendly.

Since P2P is so successful, can Martin and James suggest ways for Telcos to monetise it?

Just deliver the damn bits! Someone else made them valuable, so don’t expect the extract the premium for content creation. Build content delivery networks that are cheap to run and flexible to support many different aggregation and delivery models. Work to lower capex by creating incentives to avoid transfers during peak network activity.

What should the service providers do about the cost to them of “Over the Top��? content traversing their networks? Deep Packet Inspection does control this and requires intelligence in the network.

Throttling users in arbitrary ways is just going to lead to network neutrality legislation and consumer backlash. Find ways and incentives to get those P2P apps to deliver their content in cheaper and more efficient ways. Stick 100Tb of BitTorrent cache in every central office!

Operators need to block P2P traffic! The people that pay the bills for broadband won’t care! I would happily pay for fast content delivery…but you need to free up the network first.

This just isn’t feasible in the long run. Users don’t know or care what the delivery technology is, and if you give them a bad experience (or bill shock) then you lose too.

How do we reconcile the cost of carrying bits with flat-rate in a video dominated traffic model?

You can’t which is why the current Internet model will fragment and video distribution will be done via cheaper methods. Just like you send most parcels by parcel post or UPS ground service and not next-day 1st class mail.

Q&A Wholesale

Moving complexity from retail models to wholesale is far from simple. Peering agreements are complex – how can this be driven?

I think this is an unsolved problem, and the industry is going to have to work out new structures for aggregating wholesale connectivity of varying quantities and qualities and packaging it together. (Telecom probably ends up looking more like the mortgage re-sale market – build a network, and then leave the problem of filling the pipes to someone else.)

Please explain further the shift of complexity to wholesale from retail – give a practical example?

Easy – look at the Blackberry or Three’s X-series. You don’t need to sign up to a data plan, read the terms and conditions to see if you can do the things you want, and watch the megabyte meter all the time.

Is there an advantage for network operators to break themselves up (like BT)? Is there more shareholder value if operators split into retail and wholesale?

Yes…. But. In the case of BT, despite the Chinese walls, the CEO and board can effectively synchronise investment across the two halves of the business. BT has a good balance between the discipline of separation (wholesale can’t rely on one captive customer, retail can’t rely on a distribution monopoly) and the need for services and capacity to be deployed in sync. If there was no such synergy, we’d probably have seen a break-up already. We heard some evidence at the event that the City rather likes the structural separation model, as utilities have higher multiples than volatile telco services.

Regarding shifting of complexity from retail to wholesale: why should it be done, and how can it be done? Isn’t it just about shifting the risk? The risk is still there!

The WHY is because you want to manage the capex cost associated with the delivery of low-value (and often pirated!) video content. This is generally time-insensitive and needs to be shifted out of peak hour (or where economically unsustainable, eliminated via raising metered broadband prices at higher usage levels). The HOW is documented in the business model map – go read on the Telco 2.0 blog my 4 articles on it. You’re mitigating risk by taking some control over it, and aggregating risk to smooth it out. For example, with the BT Openreach structure, they make money no matter which one of Sky, Virgin Media, BT Retail, Carphone Warehouse, or anyone else turns out to be the winner.

Q&A Customer data and relationship

What about the future of data ownership and control? Isn’t this the future cash cow that perhaps telcos, if they can become platform players, can excel in?

Yes, in principle. Just as Google extracted the latent value from the collective effort of hyperlinking, telcos could feed social network applications with the squeezed juice of CDRs. Whether they can finesse all he legal, branding and cultural issues is another thing.

Google has been successful by storing, analyzing and monetizing what users are actually doing. Telcos have/own lots of data, what are the challenges in building business models around this data versus charging for access?

Privacy, regulatory, brand, encryption technology, user experience,…

It’s going to be tough.

How to provide user-centric solutions when telco model still largely views the household, not the individual, as the customer

That’s both a problem and an opportunity. Yahoo! don’t have a clue who I am at all, so the telcos should see their glass as half full here.

Q&A Partners, portals and services

When will web services stop being free? What will be the catalyst

They’re not free already: you have to buy an expensive PC and broadband connection, and then watch a bunch of ads (or give away valuable private data to enable those ads). We’re just moving to a world of many different business models linking the devices, services and connectivity.

Should telcos partner with Google, MSN, Yahoo on Consumer apps and focus on Enterprise / Business opportunities?

Depends on each telco. These aren’t mutually exclusive.

How can the telco better leverage the customer’s various contact directories?

Sync with network, offer an API, enable 3rd parties to access it, take out the barriers to usability!

How about this as a business strategy: rather than development & investment in new services, adopt a simpler strategy of “seed and buy” successful small competitors and using the telco capital for acquisitions?

Yes, telcos could become like Cisco and vacuum up the good services ideas. However, what’s the point if they are then limited to one operator’s distribution network? It’s like building supermarkets for Toyota car owners. Totally unnatural. Although maybe Cyworld and SK Telecom is the exception that proves the rule.

Partnering is obviously a critical success factor in the new world – how should this be embraced?

Big question. The first step of anyone embarking on a “platform” route is probably to separate partner management into its own exec position independent of BizDev.

Q&A Service innovation

Isn’t innovation with big Telcos more like “we are because we are expected to” Is it still a “show business” or real one? Why are there no ground breaking innovative products, if the above case is not true?

It’s been very comfortable selling vanilla voice and messaging. The risk/reward for telco managers doesn’t make it worth trying to jump that shark.

Where is the innovation coming from? Generate it from the inside or stimulate external creation? How to drive, channel and capture innovation?

All the above – depends on each operator’s situation and what opportunities present themselves.

My hypothesis: the endgame will be: there will be money in Access, in Devices, but not in the three C: Communication, Communities, or Content. These will be free, paid by ads or thrown in for free by the telcos.

I think all five of them will be fine businesses. Content will become more personalised and interactive, meaning much of the copyright crisis will naturally subside. Is World of Warcraft content?

Q&A …and the rest

What value or cost are users attributing to access the internet to make use of all these services?

They don’t care about the Internet per se. Campaigns to “save the Internet” have zero resonance with the public.

How quickly will regulatory bodies and policy makers react to this change?

About five years too late, if history is any guide.

To what extent can the overall telecoms spend purse be grown? How many other purses can telecoms raid?

Banking is ripe for disruption, and DoCoMo are probably the ones to watch.

What of the telecoms vendors? Especially the big complex integrated ones?

I think I’ll need consulting dollars to answer that one.

It’s going to be a whole lot smaller as an industry. Again, other industries have led the way, e.g. GE into “power by the hour” rather than turbine sales and blending services with hardware.

What are we going to do with all the copper?

Put Chilean copper miners out of business by flooding the scrap market…. There’s almost 70m tons of it in the legacy access networks.

What are the payment models for future fragmented services?

Equally fragmented, which will probably drive a user experience crisis in the medium term as the user switches between metered, free, ad-funded and “bundled” service connectivity.

What tech features telcos need to have on client platforms (PCs) to enable Telco 2.0 services?

The Virtual SIM that Intel described would be a great start, so we could provision applications with network access, rather than whole devices.