Long Tail Worries May 7, 2007
Posted by Simeon Simeonov in Digital Media, Long Tail, The Long Tail, Web 2.0.4 comments
Found an interesting study by Lex Miron from CIBC World Markets that tries to test for the existence of a long tail in interactive media properties using Nielsen Net Ratings (NNR) data. Lex measures user engagement as total number of minutes spent on a site + then looks at the distribution.
These findings alone do not support the existence of a long tail in interactive media. In fact, they support just the opposite: Interactive media is indeed the “land of the large.”
No surprise. I’m not sure anyone has claimed that there is (or should be) a long tail in brands, which is what Lex is measuring.
Long tails are enabled by the diversity in human interests. Most of us will like at least one of the songs on the current top 20 hits list. Thus begins the head of the distribution. From there on, our likes will distribute over a wider and wider universe of music. Hence, the long tail in music consumption. In short, long tail distributions are likely to appear when consumers are faced with a concrete set of product/content choices.
Most large online interactive media brands are diversified. They aggregate large amounts of content to become more attractive destinations for consumers and keep them engaged longer. Choosing between brands operating across content categories is not the same as choosing content within the same category.
Just consider the following. Brands can merge, combining their standing in the usage distribution and altering its shape without altering the underlying set of consumer content choices. It hurts to contemplate songs/artists merging…
Making Money From the Paradox of Choice February 19, 2007
Posted by Simeon Simeonov in Advertising, Digital Media, Long Tail, Mobile, The Long Tail, VC, Venture Capital, Web 2.0, e-commerce, startups.2 comments
In parallel to 3GSM, Ogilvy runs a telco conference in the beautiful Dolce Sitges Hotel, about 35km from Barcelona where the main even is. Sitges is a vacation town for the Barcelonians, a quiet place fit for a more exclusive event of 150 as opposed to the madhouse of 50,000 that 3GSM is. Thanks to my friend Ajit Jaokar who was speaking at the conference, I went along and ended up having dinner with Rory Sutherland, the owner of one of the longest job titles–Executive Creative Director and Vice-Chairman, OgilvyOne London and Vice-Chairman, Ogilvy Group UK. The humor of corporate hierarchy aside, Rory is a super-sharp Renaissance man who’s able to switch subjects at dinner faster than a croupier moves chips. Having had the chance to think through some of the topics in more detail, I wanted to share some thoughts that are applicable to a world where content supply and demand are spiraling out of control.
The Paradox of Choice re-introduced an old idea that more variety is not necessarily better. Sorting through choices takes effort. Further, since a “perfect” choice may not exist, even the knowledge of alternatives which are better in some respect can make us less happy with our ultimate choice. (There are lots of examples of this in published work. Rory’s blog links to a fascinating piece of research which suggests that one of the key reasons why people like to shop the organic sections of supermarkets is that they are presented with less choice.) This type of reasoning flies in the face of traditional neoclassical economics, which considers humans to be rational utility-maximizing creatures. More choice is always Pareto optimal, i.e., no worse than less choice. Psychology experiments suggest that’s clearly not the case. (One of the reasons why I abandoned economics was that I didn’t see many Homo economicus walking the world. It’s more interesting to think about how people actually behave than about how they should behave.)
Well, if we are not running triple stochastic integrals in our heads to optimize decisions based on all available information, how do we make choices? Without getting into details, the basic idea is that we use heuristics. That finding is supported by both theory and experiments from Herbert Simon’s work on satisficing (a combo of satisfying and optimizing) which spanned cognitive science and AI to evolutionary psychology to behavioral economics. The basic theme of heuristics is that they are quick and dirty (computationally efficient from a wetware standpoint), use local information as opposed to all available information and are much influenced by emotions.
Side note: for anyone interested in these ideas, especially in the context of how humans fail to accurately estimate probability and risk, I highly recommend Fooled by Randomness. I had missed that book despite my interest in the area and have to thank Gourdon Gould of ThisNext for bringing it to my attention. It’s a must read for anyone spending time in the financial markets, for economists and for entrepreneurs.
We live in a world where content choices are increasing at an increasing rate, from more cable/satellite channels to user-generated content to everyone dumping their content vaults on the Net. What hasn’t changed is the 24hr day (though I hear Fox execs are offering a lot of money for ideas on how to eliminate that constraint. 
). Are we better off? Sure, but… There are at least three reasons why more does not always equal better for individuals:
- The ratio of signal to noise has gone up. I have no good data to support this other than the logical argument that signal (meaningful data) changes relatively slowly as it is tied to things of importance and lasting value. I just can’t imagine there being enough new “signal” to justify, say, the rapid rise in the blogosphere content. When someone who’s not an expert on a subject (for example, me on most of the matter covered in this post) writes, the resulting text tells more about the writer than about the subject. Now, noise to some is signal to others, which is absolutely true, but that just brings the point that…
- Search and discovery costs have gone up. Again, I don’t have any good data on this, but my personal experience is that I spend more time searching for the right thing because I assume it exists and I’m less likely to accept the top choice my search engine gives me. Past data I’ve seen (I can’t find good links to it to share here) suggested that the average person puts 2-3 words in a search query + clicks the top 3-5 links in order until she satisfices her search request. More recently, I’ve heard from search engine startups and the likes of Google and Yahoo that average search term length is going up. That’s used to suggest that search engine users are becoming more productive. In addition, they may be getting more frustrated with the poor quality of the search results from simple queries and have to work harder to find what they are looking for.
- Information asymmetry has increased. In the past, there was not only less information available but, also, much of the same information was available to most people. (Think of the days of radio.) People from businessmen to your next door neighbor could make stronger assumptions about what other people knew about. Nowadays, I’m constantly caught having discussions with businesspeople and friends from the standpoint of significant information asymmetry. It takes time and effort to get closer to parity so that joint decisions can be made, e.g., which tech conferences should Polaris sponsor this year.
So, what’s the solution? There are four axes to consider:
- Go someplace where this is not a problem. What I’m describing is primarily a developed economy problem. People in Cuba aren’t troubled by too much choice. (I grew up in Communism so I know what I’m talking about.)
- Increase the available time to consume content. The supply of disposable time to consume content is bounded by the 24hr day but has been growing steadily over the past decade. TiVo and DVRs in general have allowed people to watch TV at odd hours. Despite the best efforts of some larger corporations, the Net has brought entertainment to the workforce. Mobile phones increasingly fill spare minutes with entertainment. And, yes, we have product placements in TV shows and virtual worlds as well as ads in elevators (a true startup innovation) and bathroom stalls. Unfortunately, we are getting to the point of strongly diminishing returns. It is very difficult to come up with another disposable hour of time in our busy lives. Therefore, much of content consumption will be replacement-based, which leads to…
- Make content more attractive to the audience. The Net’s “infinite number of channels” and low bandwidth costs (broadband penetration is what enabled MySpace and YouTube) have presented distribution options for both niche and user-generated content. Technology is also becoming much better at putting content in context. (An interesting point that came up at the Ogilvy dinner is that everyone is chasing mobile entertainment while they should be chasing content in context. Rory cited some research on the subject indicating overwhelming user preference for the latter.) Still, the more fragmented the content ecosystem, we have to…
- Make discovery significantly easier. This is probably the most exciting area and one where I’m focusing some of my time as an investor. Because content consumption often begins with discovery, the impact of better discovery tools is significant. Google’s market cap is built not on the fact that they sell advertising but on the fact that for many people discovery begins with their search engine. In short, to make money from the paradox of choice you have to make the large choice set seem small and relevant and that’s what discovery is all about.
So far, I’ve focused primarily on content but the same arguments apply to products. Especially in developed countries, it’s becoming harder and harder to find reasons why a new product should be purchased (the full kitchen/wardrobe/house problem is the real-world equivalent to the 24hr constraint). Manufacturers are developing systems for mass customization of products. For example, at 3GSM I saw a startup, which specialized in producing phones for niche audiences. They had a roadmap of dozens of designs, all on a common platform. And product discovery is perhaps an even more relevant problem than content discovery since products are often much more complex to evaluate.
More on the opportunities in discovery in my next post.
Long Tail Aggregators January 31, 2007
Posted by Simeon Simeonov in Advertising, Digital Media, Long Tail, The Long Tail, Web 2.0.add a comment
I’m at the AlwaysOn Media Conference in NYC today. At breakfast, Esther Dyson (who just recently left CNet), Chris Dobbrow (now at GoingOn, the platform powering AlwaysOn), Melinda Gipson (GateHouse Media) and I had an interesting discussion about how to build long tail aggregator businesses. This type of business does not have access to the head and torso of the distribution–it only works with the tail.
The basic premise is to aggregate the value under the tail. Revenue per unit (of whatever you’re measuring on the X axis) will be small on the tail portion but there are many units, hence the opportunity to build a big business, provided you can keep costs low enough to make great margins/unit.
It was impossible to build these types of businesses in the real world–costs are simply too high. You need the draw of the hits to bring an audience in and support fixed costs such as rent. On the Net, the early wisdom was that you cannot launch without the hits–what would Amazon and Netflix be if they had tried to launch without the bestsellers and blockbusters? Things have changed in a decade. Search and social discovery mechanisms have made niche content much easier to discover, bringing customer acquisition costs down enough to make a meaningful difference.
At the highest level, to become successful as a long tail aggregator, you have to be very easy to do business with at every level: how you’re discovered, how customers and partners engage with you initially and over time. If you’re not, there will be increased sales, marketing, account management and support costs, which directly affect margins. Part of being easy to do business with has to do with you providing a whole solution (in a Crossing the Chasm sense) to your customers and partners.
The obvious, Business 1.0 approach is vertical integration. That’s the Apple way. If you can execute it with near perfection over time, it’s a fantastic way to go. It’s also very expensive and very risky. iPod/iTunes couldn’t have launched outside the shadow of a large company such as Apple.
The Business 2.0 way is orchestration and clean APIs. You need to orchestrate the solution because you must control your customers’ experience. You partner with best-of-breed providers to fill out the areas where you need help. By itself, that’s no different than what Geoffrey Moore preaches. In a Web 2.0 world, the difference comes through clean APIs, which allow you to maintain efficient, fluid and broad partnerships while keeping costs low.
Bob Metcalfe on Internet Video January 8, 2007
Posted by Simeon Simeonov in Digital Media, Long Tail, Polaris Venture Partners, The Long Tail.1 comment so far
Beet.TV caught up with my partner Bob Metcalfe and did a nice interview on the evolution of the Internet to the video Internet. Bob makes three key points:
- The Internet has been getting richer since the days of uppercase ASCII
- As communication becomes richer it replaces transportation
- As the pipes and infrastructure improve, the Long Tail demand can be satisfied.
Link to Ethernet Inventor Says Transformation to Internet Video is Here – Google Video
Mobile Advertising: Not If, But When And How November 8, 2006
Posted by Simeon Simeonov in Industry News, Long Tail, Mobile, The Long Tail, VC, Venture Capital, startups.4 comments
People seem to fall in two camps with respect to mobile advertising: those who see it as an imminent opportunity of significant size and those who say it’s a few years off. No one is willing to say that mobile advertising won’t work.
Certainly, there is a lot of experimentation in the space. MS toyed with Third Screen Media. Yahoo is experimenting with mobile advertising in search and yesterday announced graphical ads for e-mail and news. Google has tested mobile ads in Japan. AOL allows sponsors on AOL mobile search.
The publishers (both carriers and brands) are jumping in the mix, attracted by CPMs in the $25-50 range (anecdotal data I’ve heard from people in the industry). Since, especially in the US, they control the users, the carriers will definitely make a lot of money from mobile advertising. Not just through SMS but also WAP (data I’m getting from carriers suggests 40+% of new plans sold include HTTP access) and branded downloadable applications.
The user experience will be figured out–there is too much money at stake. The good news is that mobile as a platform is even more measurable than the Net. Therefore, many mistakes can be made quickly and relatively cheaply. (As an aside for entrepreneurs, follow Esther Dyson’s “always make new mistakes”.)
So it’s no surprise that there are the slew of mobile ad startups entering the space, teased by predictions of $1-10B (yeah, that’s quite a spread) market size for mobile advertising by 2010 and month-to-month growth rates in excess of 50% (according to AdMob). There are too many companies to list here. The ones I’m tracking have teams with deep experience in both online advertising and the carrier world, which is a rare combination. The typical pitch is that locking up relationships & inventory early offers lasting advantages. I’m not so sure about this. The Net has taught publishers a lot about optimizing ad revenues by playing vendors against one another and making real-time decisions about which ad to show when and to whom. That creates an opportunity for technology companies to offer the equivalent of DART for Publishers on mobile.
At the current pace of company formation, within two years, I’d expect the mobile advertising platform and technology space to be quite crowded and experiencing some margin pressure. As on the Net, the best place to be would be as a large publisher–it’s the only guaranteed way to make money.
Who will be the large publishers in mobile? Certainly the carriers and certainly the big brands. What about everyone else? On the Net there is a Long Tail of content providers. You don’t need to be a Top 5 brand in a category to do very well. There is no such thing in mobile. The hits make money. There is no Long Tail. In fact, there isn’t even a short tail. Something worth thinking more about…
Metcalfe’s Law: more misunderstood than wrong? July 26, 2006
Posted by Simeon Simeonov in Long Tail, Metcalfe's Law, Polaris Venture Partners, The Long Tail, Web 2.0, network effect, social networking.20 comments
The industry is at it again–trying to figure out what to make of Metcalfe’s Law. This time it’s IEEE Spectrum with a controversially titled “Metcalfe’s Law is Wrong”. The main thrust of the argument is that the value of a network grows O(nlogn) as opposed to O(n2). Unfortunately, the authors’ O(nlogn) suggeston is no more accurate or insightful than the original proposal.
There are three issues to consider:
- The difference between what Bob Metcalfe claimed and what ended up becoming Metcalfe’s Law
- The units of measurement
- What happens with large networks
The typical statement of the law is “the value of a network increases proportionately with the square of the number of its users.” That’s what you’ll find at the Wikipedia link above. It happens to not be what Bob Metcalfe claimed in the first place. These days I work with Bob at Polaris Venture Partners. I have seen a copy of the original (circa 1980) transparency that Bob created to communicate his idea. IEEE Spectrum has a good reproduction, shown here.
The unit of measurement along the X-axis is “compatibly communicating devices”, not users. The credit for the “users” formulation goes to George Gilder who wrote about Metcalfe’s Law in Forbes ASAP on September 13, 1993. However, Gilder’s article talks about machines and not users. Anyway, both the “users” and “machines” formulations miss the subtlety imposed by the “compatibly communicating” qualifier, which is the key to understanding the concept.
Bob, who invented Ethernet, was addressing small LANs where machines are visible to one another and share services such as discovery, email, etc. He recalls that his goal was to have companies install networks with at least three nodes. Now, that’s a far cry from the Internet, which is huge, where most machines cannot see one another and/or have nothing to communicate about… So, if you’re talking about a smallish network where indeed nodes are “compatibly communicating”, I’d argue that the original suggestion holds pretty well.
The authors of the IEEE article take the “users” formulation and suggest that the value of a network should grow on the order of O(nlogn) as opposed to O(n2). Are they correct? It depends. Is their proposal a meaningful improvement on the original idea? No.
To justify the logn factor, the authors apply Zipf’s Law to large networks. Again, the issue I have is with the unit of measurement. Zipf’s Law applies to homogeneous populations (the original research was on natural language). You can apply it to books, movies and songs. It’s meaningless to apply it to the population of books, movies and songs put together or, for that matter, to the Internet, which is perhaps the most heterogeneous collection of nodes, people, communities, interests, etc. one can point to. For the same reason, you cannot apply it to MySpace, which is a group of sub-communities hosted on the same online community infrastructure (OCI), or to the Cingular / AT&T Wireless merger.
The main point of Metcalfe’s Law is that the value of networks exhibits super-linear growth. If you measure the size of networks in users, the value definitely does not grow O(n2) but I’m not sure O(nlogn) is a significantly better approximation, especially for large networks. A better approximation of value would be something along the lines of O(SumC(O(mclogmc))), where C is the set of homogeneous sub-networks/communities and mc is the size of the particular sub-community/network. Since the same user can be a member of multiple social networks, and since |C| is a function of N (there are more communities in larger networks), it’s not clear what the total value will end up being. That’s a Long Tail argument if you want one…
Very large networks pose a further problem. Size introduces friction and complicates connectivity, discovery, identity management, trust provisioning, etc. Does this mean that at some point the value of a network starts going down (as another good illustration from the IEEE article shows)? It depends on infrastructure. Clients and servers play different roles in networks. (For more on this in the context of Metcalfe’s Law, see Integration is the Killer App, an article I wrote for XML Journal in 2003, having spent less time thinking about the problem ). P2P sharing, search engines and portals, anti-spam tools and federated identity management schemes are just but a few examples of the myriad of technologies that have all come about to address scaling problems on the Internet. MySpace and LinkedIn have very different rules of engagement and policing schemes. These communities will grow and increase in value very differently. That’s another argument for the value of a network aggregating across a myriad of sub-networks.
Bottom line, the article attacks Metcalfe’s Law but fails to propose a meaningful alternative.
