Michael Fives

  1. Thu 19 January 2012


    Originally published on Streetfight as Sounds and Places: Experiencing Bluebrain’s ‘Central Park’

    Bluebrain is not a startup. There’s no technical co-founder or venture capital dollars. In fact, they’re not even a company. Bluebrain is a Washington DC-based band, and they’ve created one of the most compelling location-based apps of the past twelve months. The app is a location-based album called Central Park (Listen to the Light), and it accomplishes what very few others have to date been able to: namely, to build an app that actually augments reality vis a vis location. More than that, though, I’d like to suggest that CP(LTTL) is a model for the possibilities of what the next generation of location-based apps should look like.

    First, a couple of words about the project. Bluebrain is a two person ensemble from brothers Ryan and Hays Holladay. CP(LTTL) is their second location-based album, following the release of National Mall in May 2011. The concept behind the album is simple: it’s an app that users can download to their iPhone, and that is designed to be listened to within the context of a specific location - in this case, Central Park, NYC. Then, as the user is moving through Central Park, the app “tracks their location via the iPhone’s built-in GPS capabilities; the melody and rhythm of the music varies in accordance with the user’s path.” CP(LTTL), in other words, is a soundtrack for your experience in Central Park. And it is, uniquely, your soundtrack - because although the music is contextually tied to your location in the park, the progression and sequencing of that music is determined by factors like route and pace that have led you to that spot.

    Central Park builds on the rich layer of metadata surrounding your location (route, speed, topography) to craft a wholly unique experience. But consider the effect: it’s a soundtrack for your exploration of the park. The app itself fades into the background of that exploration, augmenting it and making it new, but never becoming the focal point of your experience. “It’s a choose your own adventure album,” and the adventure is layered onto your reality. It might not qualify as capital-A Augmented Reality, but there’s no doubt that Central Park is a far more powerful user experience than most other Augmented Reality apps on the market. This, I want to suggest, is the real potential of “location-based apps” and “augmented reality” - the potential to layer brand new experiences on top of the rich data that exists within everything that we call ‘location’.

    Compare Central Park’s use case to the broad category of apps that we refer to as ‘location based’, i.e. restaurant review apps, social networking apps, daily deals apps, etc... By and large, these products understand location as a simple reference point (literally) of latitude and longitude. In these cases, location is something to be catalogued, and not something to be interacted with. Rather than fading away into the background, the app becomes the center point of an experience, the user dutifully picking away at tiny keys to check in, update a status, or search. Rather than creating experience, the app is merely representing it. Central Park, on the other hand, understands that the app doesn’t need to be the center of the user’s attention in order to create a magical experience, but can instead use a variety of data about your location to embed itself into the very experience itself.

    Ultimately, I think, what we’re talking about are apps that implicitly use all of the rich metadata of location (things like route, environment, topography, proximity to other people, places, and things, etc...) to create magical, serendipitous experiences in day-to-day life. We’re talking about things like autonomous search. Serendipity engines. That kind of thing. Now, this isn’t to suggest that there aren’t beautiful technologies, well designed apps, and very profitable business models that are being built on top of the location-as-reference-point model. But personally? I would love to see what’s possible when we consider all of the ways that location creates experience. Or when we think in terms of creating experience through location, and not merely representing it. A next generation of location based apps that incorporates all of the rich data of location, rather than just a spot on a map, to create magical new experiences and to tell new stories.

  2. Thu 22 December 2011


    Originally published on Streetfight as Doubling down on SinglePlatform

    So here’s the thing: it turns out that selling online services to small businesses is really hard. OK, so that’s not exactly a revelation for most people who read blogs about “the business of hyperlocal”, but banalities aside, the thing about truisms is that they are fundamentally … true. And so far this one has been no different. Ask anybody who’s walked into a small business selling hyperlocal advertising and they’ll tell you - your average small business has a limited budget and even less time to spend on their online presence, and those that do want to see short term, tangible returns on their investment. And yet ... everybody from Google to Groupon to Foursquare is fighting to “own” the relationship to the local business because of the multiples of billions of dollars in associated advertising spend. We’ve seen resources allocated at a massive scale in the online hyperlocal space, and yet there is no single company that can come close to saying that they own that small business relationship. But there is a company out there that looks like they might have a shot to pull it off, and they’re called SinglePlatform.

    NYC-based SinglePlatform was launched in January 2010 by CEO Wiley Cerilli, previously a founding partner and EVP of SeamlessWeb. Cerilli founded SinglePlatform to help restaurants to expand online by providing a one stop shop for building a digital presence (i.e. websites, Twitter, and Facebook). Almost two years later, they claim to have information from hundreds of thousands of menus in over 13,000 cities across the United States in their listings database, which would indicate that they’ve been noticeably successful in the space. But ascribing SinglePlatform’s success to a niche play in the restaurant space would seem to miss the point, especially since they’ve started to move beyond the world of restaurants and are beginning to offer services to businesses in other verticals who are selling other products. So they’re not all about restaurant menus. In fact a better way to think about SinglePlatform might be as the distribution point for a business’s digital storefront - ‘distribution’ being the key term.

    It’s important to note that there’s actually two constituencies in the SinglePlatform universe: the small businesses that provide the information, and the publishers that distribute it. SinglePlatform has built a broad network of digital publishers, from major media companies and IYPs to data providers, mobile applications, hotels, and universities, who all use SinglePlatform to distribute content about the local businesses in their network. For example, if you click on the menu tab on a restaurant’s review in The New York Times, you’ll see a menu that is powered by SinglePlatform. The value proposition to the distributor is clear: one of the first things that any user who’s researching restaurants is going to look for is the food that’s on the menu. By providing that content onsite as a menu instead of linking to a restaurant’s web page, they can improve core metrics like user retention while offering a superior user experience. SinglePlatform can also provide analytics that shows users accessing menu content in real time, so ROI is demonstrable and immediate.

    There’s immense value in this ecosystem. At some level, local businesses are beginning to realize that the effect of the Internet is to essentially comoditize the sale, and so it has become crucial to reach potential customers at the point at which they’re making a purchasing decision. Thus, wide-scale distribution and access to the publisher network is perhaps the most compelling reason for SMBs to join the network. What we’re talking about now are network effects, and network effects are the stuff that great businesses are built on. Specifically, SinglePlatform has built a network of two distinct groups - businesses and publishers - that exist symbiotically … otherwise known as a two-sided network. The value of a two sided network is that, besides being self-perpetuating, they are reliant on the platform that connects them to establish a pricing structure that can maximize the value of the network. For example, the more businesses that list their products on SinglePlatform, the more incentive there is to potential publishers to join the platform. And the more publishers that join the network, the more value there is to existing and prospective businesses. And so if SinglePlatform has figured out the right way to maximize value for each of the participants, it’s going to be incredibly difficult for a challenger to come along and replace them in that ecosystem.

    The real test for SinglePlatform, as with most companies in this space, is going to be to figure out how to continue to scale and grow. The fact remains that even for an organization that really understands how to sell to a local business (as it appears that SinglePlatform does), building a local sales force at scale is an immensely expensive proposition. This is doubly so when you expand outside of the major DMAs (as a company like Patch will attest) since you can’t exact the same efficiencies in smaller towns as are available in more densely populated areas.

    So the question for SinglePlatform is ultimately one of ROI - how to manage costs while at the same time continuing to increase the rate of return for their themselves and their customers. In this case, managing the “Investment” side of the equation really means managing the cost of sale. Presumably there are opportunities to move existing customers towards a more self serve option, so that the sales organization is mainly focused on new customer acquisition. Similarly, integration at the point of sale would enable SinglePlatform to automate some parts of the data acquisition process, while at the same time reducing the investment (in terms of time) required from businesses to keep their listings up to date. Both of these solution have their own complexities, however, and driving adoption with the SMBs here is far from a given.

    Alternatively, the “Return” side of the equation is mostly a function of distribution. How can SinglePlatform continue to grow their network in order to extract a much value as possible from each business listing? This, in my opinion, is where the really interesting product opportunities exist for SinglePlatform, because what we’re really talking about is monetizing a structured database of product inventory (what are menus, after all, but a list of a restaurant’s products?). But that’s a topic for another blog post.

    So it’s true - selling to small businesses is really hard. But here’s something else that is capital-T true: local is a massively huge market, and so are the rewards for any company that figures out how to own that relationship to the small and medium business. Can SinglePlatform do it? I wouldn’t bet against them.

  3. Wed 07 December 2011


    (Originally published on Streetfight as Hyperpublic: Structuring Place Data, Redefining Hyperlocal’s Scope)

    Hi. So are you up for a quick thought experiment? Let’s say that you’ve met some friends for drinks at Finnegan’s Pub after work, and, while you’re waiting to order a beer you check in to Finnegan’s on Foursquare. What data did you just create about yourself?

    Well, at a very basic level you’ve created an association between you and Finnegan’s at a certain time, i.e. You are at Finnegan’s. But if you traverse the graph of that association a level further, and if you start to examine the metadata surrounding that check-in, it becomes clear that that data reaches far beyond place and time. For example, there’s the fact that Finnegan’s is an establishment of type Bar. Or the fact that it’s after work, and that it’s happy hour. And then of course there are the drinks and the food that are at that bar, one of which you maybe tweeted about after your ordered. And there’s the neighborhood that the bar is a part of. And there’s all of the other people who are at the bar with you right now. And so on, and so on... until we’re no longer just talking about “You at Finnegan’s”, but instead we’re talking about ‘You After Work at a Bar called Finnegan’s that serves Guinness and Wings during Happy Hour and cetera...” .

    All of this is just a long way of saying that there’s a shit ton of mostly unstructured hyperlocal data floating around out there on the web that is immensely valuable to businesses, advertisers, developers, and pretty much anyone who’s building a business in the hyperlocal space. Or, we should say, it would be immensely valuable if only that data was structured so that one could ask, for example, for “the Twitter handle of everybody who, after work, goes to bars in the East Village that serve Guinness during Happy Hours.” Enter HyperPublic, a New York City based startup founded by Jordan Cooper and Doug Petkanics that wants to be the open platform for the web’s local data.

    HyperPublic is a platform that “collects, organizes and structures geo-local information from the open web and makes it available to the world for free.” If you imagine the web as a giant database of unstrcutured data, then HyperPublic is an index that is structuring relevant information by location (although there is presumably some hard core data gymnastics being performed in the background that makes all this possible). Put another way: HyperPublic is to a lat/long as Google is to a link (or as Facebook is to a friend). They are a platform in the sense that they do not provide a true consumer facing product (yet), but rather provide data via their APIs to developers or companies who are building hyperlocal apps and websites. For example, a company that wants to build an app that aggregates all daily deals in Brooklyn might use HyperPublic’s platform to query for Brooklyn deals rather than scraping Groupon, LivingSocial, Amazon, and the bazillions of other daily deal sites out there separately. In fact, GeoDeals and Events (where developers actually get paid to display daily deals) is currently one of HyperPublic’s three core products - the other two being Places+, which is essentially a POI database, and Data on Demand, which is kind of like a catch-all for ‘everything else’. Thus, at its essence, HyperPublic is hyperlocal data.

    So why is this exciting? Besides enabling a slew of new daily deal aggregators, what might we see spring forth from the ether when local data is properly structured? It’s easy to imagine various applications of search, i.e. “Show me all wine bars that serve croquettes.” (N.B. this seems to be more or less where the thinking is headed at HyperPublic Labs, at least for the moment). Similarly, one can imagine that this type of data would be immensely valuable to advertisers who want to target, for example, everybody who drinks beer and is in walking distance of an establishment. The range of uses for this type of data is, in

    The trick for HyperPublic, however, is going to be to figure out just how exactly to monetize all of this data at scale and beyond the crowded marketplace of Daily Deals distribution. We’ve already seen other companies, such as SimpleGeo who was recently acquired by UrbanAirship, struggle to build a robust business model by product-izing location data (although, to be fair, this was basically a partnership via acquisition and a good move for SimpleGeo ... but still the acquirer vs acquiree sides of the equation are telling). That particular wicket is made all the more sticky by companies like Foursquare who are essentially giving away venue data and customer insight to build an ecosystem around their core business. So where do the opportunities lie?

    Part of HyperPublic’s key value proposition as an aggregator is that they are data-source agnostic, and so can focus on building an infrastructure that allows them to correlate data at a massive scale. Similarly, by focusing on correlating all of the rich data available around a location, and not just on the venue, they broaden the scope of potential use cases. While a company like Foursquare might be interested in aggregating some data from other sources, they are principally interested in working within the scope of their own data. Aggregation and correlation, however, are HyperPublic’s bread and butter, and it’s super interesting to think about what’s possible when you start to look at that mass of data in the aggregate. For example, by leveraging data from across multiple platforms, HyperPublic could think about building a decision engine that would be by definition more useful than one that focused on a single source of data (and, as eBay’s recent acquisition of Hunch has demonstrated, there’s always going to be a robust market for acquisitions of data platforms that can drive decision making). Or, alternatively, forget about driving decisions about what’s happening right now, and imagine what’s possible if you can start to infer patterns of movements between people, places, and things. Conceivably we can start to envision services that would enable predictive analytics and future-optimization. For example, imagine a company that’s planning to open a new show store. Wouldn’t this company be interested in licensing data that could help them to predict traffic patterns and user tastes in various neighborhoods and that might influence their choice of locations? Or wouldn’t they be interested to be able to use this data for inventory planning, based on planned rather than perceived market demands? The possibilities are, as they say, endless. It’s a great, big, exciting problem for a company that has the talent to take it on and the patience to solve it, and it’s not much of a leap to suggest that if a company like HyperPublic can be the ones to solve it, they’ll redefine the scope of the possible in hyperlocal.