How Long is Your City's Tail?

Here’s something that everyone doing business on the web knows today, thanks to Chris Anderson: it’s all about the long tail.

When the cost of each individual transaction falls to nearly zero, marginal and low-performing items, grouped together, can account for a lot more of the overall value of a company than the top-performing ones. makes more money from the aggregate of all of the books that sell one or two copies a month than from sales of best sellers. And Amazon is a much stronger, healthier, and richer company because of the extremely long tail of books it sells.

Everybody gets that.

What almost nobody realizes yet is that the same is true for cities – or can be.

Most cities right now are models of closed, rigid systems, systems that rely on a few, top-performing agents to get civic tasks done and keep quality of life high for residents. Most of these agents are departments of the city itself, though some are outsourced. Either way, cities rely on one agent per issue, no more. To use as an analogy, cities today are like an Amazon that only allows the #1 best-selling book from each category into its system.

A good number of cites are beginning to do deals with mega companies like Google and IBM, giving them access to city data so that they can build excellent tools for residents to use. This is a great thing. I love looking at Google Maps and seeing that bulging red line right next to my house indicating that traffic there is at a standstill so I should consider biking instead. Makes my life better.

Still, to use the Amazon analogy again, now these cities have allowed not just the #1 best-selling book from each category into the system, but best-seller #2, #3 and #4. It’s still a closed system of cherry-picked agents with privileged access.

And that’s about where a lot of people would choose to end the opening of cities’ data – giving unrestricted access to the Googles, Microsofts, and IBMs of the world. The rest get limited access, or access contingent upon satisfying some governmental board or other. (New York’s Mayor Bloomberg, who launched his Big Apps contest yesterday, is seemingly among this group.)

If we do that, of course, we’re missing out on what is potentially the biggest piece of the pie – the tail. That’s where a huge chunk of the value comes from.

So, imagine instead a city that has totally open, unrestricted access to data (say, San Francisco or DC in 2011). What does it look like? It has all of the familiar city-run departments providing all of the services and assistance they’ve always provided – that’s not going away. Then it also has public services offered by the mega companies, the Google Traffic, IBM’s Smarter Cities, and so forth. Those are huge added value to these open cities – they’re used by a large percentage of residents and make life in those cities better. But THEN, it also has an insane long tail of services set up and run by anyone with an interest in doing so, just by hooking into city data, distributing it in a new way, improving on it, mashing it up, giving it back to the city, etc. These services each individually get used by a small minority of people, but collectively they get used by more than any other single source in the city.

That’s the healthy, long tail city of the future in action: head, “meaty middle” and tail, all working together, all reinforcing each other, all driving each other forward.

And that’s the future of cities.

So it might be time to ask yourself: how long is your city’s tail shaping up to be? The answer may determine, to a large degree, how much your city is a thriving place to live in decades to come.

tags: ,
  • Thanks John, that’s a great analogy.

    I’ve been thinking around Tim’s notion of ‘do it ourselves’ public services and it’s becoming easier to understand where digital design patterns can be applied to civics.

    One notion I’ve been toying with is is that as a city administration’s monopoly of public services diminishes, do we unintentionally fragment communities?

    Is a valid post-Gov2.0 role for local government to regulate the emerging ecosphere of DIO services?

  • bob

    Oh come on, the long tail was a nice theory but has been thoroughly disproved by actual events.

    Even after reading the article twice I still don’t understand how you’re trying to hammer this analogy on to opening up data sets.

  • Even with open access to public data, I think the effort it takes to utilize the data would prevent a long tail from developing. Most of the population can’t write software, design data visualizations, or translate questions about policy into questions about data. Applications at the far long end of the tail (those specific to a small locality, issue, or number of users) aren’t cost effective.

    My hope is that smarter tools will eventually empower the average person to utilize public data. Such tools would need to understand natural language, combine the necessary data sources, and provide clear output and visualizations. Ideally, someday we’ll be able to ask a tool like Wolfram Alpha for “map of water quality vs poverty in Chicago”.

  • You’re describing cities full of “situated software” to borrow Clay Shirky’s term from 2004. It’s absolutely the right way to go, and how Jane Jacobs would want us to build city apps if she were alive.

  • The number of people who do the most work in any organization including a city is very, very small. So it’s not true that there is value in the entire tail. We are are all benefiting from the law of large numbers to locate the 0.01% of the tail that does something useful.

  • John:

    Thought this was very compelling and wanted to see if you would repost or add it to our new collaborative intelligence project, the SmarterCities Scan:

  • I too thought research into “the long tail” theory proved that the top-sellers clearly outperformed the “long-tail”. I believe HBR did a nice job covering this:

    That said, I do believe opening city data to the public could yield useful apps and services. I find it sad that geo-related data remains the killer data-source for building apps. Surely cities are sitting on information that can be “consumed” into ideas and applications that can transform civic life.

    The big question: How do we convince them to open their treasure troves?

  • @Mike:

    “I think the effort it takes to utilize the data would prevent a long tail from developing”

    A long tail is already beginning to develop, in the form of apps, mashups, various websites.

    Also, one can’t assume that the effort it takes to utilize data is going to remain constant. In 2001 you could have said “the effort it takes to write fresh content on the web will prevent a long tail of web content from ever developing”. And you would have been right, until Blogger, SixApart et al came along.

    In other words, this will get easier to do. Give it time.


    “The number of people who do the most work in any organization including a city is very, very small.”

    Right. That’s exactly what the power law says. That’s true in any scale-free network. It’s also true that there will be a long tail alongside those power contributors. And there’s value in that tail.


    “I too thought research into “the long tail” theory proved that the top-sellers clearly outperformed the “long-tail”.”

    Whether the tail outperforms the head or vice versa depends on the makeup of each individual power curve. But that is of little importance in my mind. Maybe the tail outperforms the head, maybe the head outperforms the tail.

    What’s important is to realize that there IS a tail, or at least a latent tail, to city services.

  • So what do you say in response to Lessig’s arguments against transparency? (

    What data sets are least problematic to open up from a security standpoint? Which are likely to be least resisted? What’s the simplest starting point, and what are the typical internal obstacles? Fear of making a lot of municipal jobs obsolete, or taking too much IT control out of government hands? Fear of what the private sector will do? Is there a role here for the press or some new mediating institution between the public, commercial interests, and government?

  • Hi John, great post.

    I’ve been writing about the same idea – calling it the Long Tail of Public Policy.

    I also made it a big piece of my chapter for the upcoming O’Reilly Book on Gov 2.0. What you describe is precisely what I’ve been telling people here at the City of Vancouver as we have created our data portal and begun to see the beginnings of this long tail effect with applications like this one.

    Your counter to Jason is also bang on. It may be that the tail is much smaller than the head – but does it matter? The cost of sharing this data is not significant and so even small benefits become value generating.

    More importantly, the real benefactors of shared data are the governments themselves. Presently, most governments cannot easily access the data they themselves generate since it is often locked in silos and not easily distributed. Creating systems that will foster a long tail among a city’s public also helps the head of the tail – the city’s administration – by finally giving it access to data it can use to analyze problems and create new services of its own.