Can Carriers/Mobile Network Operators execute Long Tail / Web 2.0 applications?

Can Carriers/Mobile Network Operators execute Long Tail / Web 2.0 applications?

A good question .. raised at forumoxford. Here is my analysis .. I believe that most of what I am describing below can be done by a product such as Xtract today

What is the Long Tail?

The Long Tail is the opposite of the Pareto distribution(aka 80/20 rule) and is the preferred business model behind many of the recent Web successes – specifically Amazon.com.

Supply side motivations

On the supply side, we need low inventory and distribution costs. This means, unpopular products can be stocked. Netflix and Amazon are examples of this model.

Demand side motivations

Demand side motivations include search engines, recommendation engines, sampling tools etc allowing customers to search a wide range of products OR for products that they normally cannot find physically in a store due to lack of time. Again, Amazon.com and many others excel here.

Much has been said about why telcos cannot execute Long Tail applications. So, I am addressing the more challenging aspect of – How Telcos CAN execute Long tail applications.

Firstly, demand side is a software, recommender, and influencer problem. This can be addressed.

The supply side problem is more complex. The Telco needs to maintain a Long Tail inventory of some entity. The obvious answer is a Long Tail of content. A more interesting answer (for a Telco) is a Long tail of people.

A Long Tail of people is an interesting proposition for a Telco since it is accessible through voice and SMS. Remember that a social network can be built from transactions i.e. the underlying data. There are various research papers about how this can be done(for instance creating a social network from email data or from IM data). Therefore, implicit and explicit recommendations can be captured from voice data or SMS data and that can then be used to identify the long tail in an aggregated manner

Here is an example:

If many people call a specific Pizza place, then that Pizza place gets an implicit recommendation as ‘Good’. Note that individual records are not revealed ; rather aggregate records are used to create the rating. The initial rating can be displayed to the users and users can contribute. For instance – if users see this Pizza shop rated ‘Good’ and it turns out that they have been calling to complain – then users can update that information (this works same as Amazon ratings). But the difference is – this information is seeded by the Telco from voice records. Once it is seeded by the Telco, it can be enriched by the users. This is classic Web 2.0

So,

a) Supply side: The ‘Long tail content’ could be lots of shops and stores(by postcode)

b) Demand side: The social search and recommendation tools can be created

What is missing

a) Profiles!!! At the moment, very basic user information is known to the Telco

b) External feeds(for instance feeds of restaurants etc)

c) A social network built from transactions i.e. the underlying data

Like I said, all this is currently possible through products like Xtract. I think Telcos need to address this space by engaging more with the end users and building the profiles(that’s the missing link!).

Comments welcome

Comments

  1. Andreas West says:

    Hi Ajit,
    interesting idea, because i think nobody has analysed calls from consumers to business numbers so far, aggregating them and using this information for business purposes.
    But I have some questionmarks on idea.
    1) Many calls to a pizza service
    Can mean that I’m a happy customer, but also can mean that I’m talking often to somebody working there for private reasons.
    2) Many calls to a business
    Can also mean that my problem is not solved, especially with lots of calls in a short period of time. So it’s a rather negative thing, where your recommendation can come into place and destroy a business (aka the rating of Ebay)
    3) Is there enough to recommend?
    I’m not sure if there’s enough restaurants etc. where consumers call in. It probably would require a deep analysis of consumer to business calls first of all so see where the majority of those calls goes and if there is indeed a possibility to make some money with it.
    Generally I see it as a perfect example of Web 2.0 execution, but there are some hurdles in it (see above). The above mentioned aggregated information could be used for directory assistance calls, but are people who use DA service willing to pay for it?
    Still, a very interesting new aspect of what to do with Telco call detail records.
    Andreas

  2. Ajit Jaokar says:

    Thanks Andreas. Here is where the feedback loop comes into play. i.e. I totally accept that there may be some instances where someone calls becanse they have friends working there .. but realitically – how many pizza businesses do we know that have their staff on the phone chatting to non customers? :) But more to the point, the feedback loop will come into play. By that I mean that the Telco will only SEED the data – after which people may rate/comment based on their experiences like Amazon.
    At the end of the day, Calling, SMS, browsing etc are all attention streams (see my previous post about APML). A telco, if it is smart enough, could be in the business of capturing attention streams and hence be able to convert those AGGREGATED attention streams into ratings. Voice is one element of an attention stream which I used to illustruate the concept.
    Hope that helps. Thanks for the comment. kind rgds Ajit

  3. Christian says:

    What is salient really is that profiles are the key. Telcos have been really reluctant (I don’t see it changing soon) to initiate any sort of profile because of privacy concerns. Some thing that they seem so sensitive to. Ironic how willing they were to give out data to the US gov. with little prompting.

  4. Roy Sattethwaite says:

    Ajit,
    While your example on profile and ratings is valid for the long-tail model of web 2.0 (especially rich web content), the long tail of web 2.0 is largely still inhibited by supply side complexity and demand side distribution problems.
    You state above “on the supply side, we need low inventory and distribution costs”. I agree. For example the Amazon long tail model works because authors can publish (content) and users can find it, then easily consume it. This is how a niche book that only sells 1,000 copies can thrive in a long tail model (it doesn’t take too much money to create, inventory then distribute the book vs. the traditional brick & mortar book value chain model).
    This is not so with the mobile Web. Authors have high publishing cost hurdles to create mobile web content because of device / platform fragmentation. In addition there is no uniform distribution (merchandising) model across the variety of wireless carriers. Finally, there is no simple way for consumer discovery & repetitive use.
    Won’t these systemic problems needs resolution before we see rapid advancement of the long-tail model for the mobile web?
    Roy

  5. Ajit,
    While your example on profile and ratings is valid for the long-tail model of web 2.0 (especially rich web content), the long tail of web 2.0 is largely still inhibited by supply side complexity and demand side distribution problems.
    You state above “on the supply side, we need low inventory and distribution costs”. I agree. For example the Amazon long tail model works because authors can publish (content) and users can find it, then easily consume it. This is how a niche book that only sells 1,000 copies can thrive in a long tail model (it doesn’t take too much money to create, inventory then distribute the book vs. the traditional brick & mortar book value chain model).
    This is not so with the mobile Web. Authors have high publishing cost hurdles to create mobile web content because of device / platform fragmentation. In addition there is no uniform distribution (merchandising) model across the variety of wireless carriers. Finally, there is no simple way for consumer discovery & repetitive use.
    Won’t these systemic problems needs resolution before we see rapid advancement of the long-tail model for the mobile web?
    Roy