Guy Posts

At LeWeb 2008, Marissa Mayer made the claim that an individual will conduct 70% of their economic transactions within 5 miles of their home. I agree, the market for local search — answering the question of where I should go to [eat, shop, drink, play] is the next frontier in organized search.

Theoretically, local search should be little more than attaching a few geographical filters to a broader query (“pizza” becomes “pizza in san francisco”). Practically, “local search” must take into consideration not just location, but time, weather, friends, traffic, economic incentives, and past behavior. Local search is about going beyond the critical review, the 5-star scale, and addressing this one question:

Where I should be going now?

Twitter and Foursquare will be waging war over this question in the months ahead. I’ve been watching this one very closely. I use both applications. For a long time I’ve been impressed with what twitter will be able to do with local search. Two years ago, I created a search application that displays tweets on a map based on location. Last year, I felt that twitter was going to crush local search, but now I’m not as sure.

Having used foursquare, I can see the tremendous upside in their product. People are literally telling this company where they go to do things. The commercial potential of this data makes me shiver. Furthermore, my dream job would be to use this information to develop a time + place recommendation engine that answers the following question: where should I go right now?

Twitter has a grasp on what, Foursquare has a strangle on where. While some say that “what” is more intriguing than “where”, to me the killer data is “where”. Just as google used a hyperlink as a “vote” for another website, a foursquare check-in is a “vote” for a certain place at a certain time. People are utility optimizers, where they are at any time represents their optimal allocation of place utility. On aggregate, where more people are, the better the place. The need for in depth, lengthy reviews becomes secondary. Additionally, by allowing people to find other people with similar “place-tastes” allows one to discover new places serendipitously. To me, the location data is more powerful than what people say about it. Like link tracking, it eliminates from the equation confabulation and strikes to the root of the question: where did you spend your time?

Foursquare is taking the key steps needed to build the foundation that will provide this answer. They will have it in the next 18 months.

1) Encourage People to Share Their Location. The check-in is a concise way to convey two bits of information: location and time. Additionally, by incentivizing users to check-in and earn points to edge out their friends and win goofy badges, Foursquare has taken steps to decrease the pain of having to constantly “check-in” at each location. I suspect they are going to focus their resources on making the check-in process as streamlined and painless as possible. In 2008, I wrote that twitter, by limiting us to 140 characters, was forming the foundation for a global repository of ideas that taken together will have the ability to break news and events faster than traditional outlets. Today, I contend that Foursquare will have the ability to tell us where to go far better than a coupon in a newspaper, an advertisement on tv, or a series of reviews yelp.

2) Build a Recommendation Engine. While each check-in is a unique, independent event, taken together, your check-in history represents your implicit vote for how you’ve decided to spend your time. If we’re a utility maximizing species, where we are at a given time — in most cases –represents our optimal consumption bundle of space, time, and money. Google uses hyperlinks on a page to represent a “vote” for the linked page. A check-in is a “vote” for a place. A series of check-ins is a personal voting history– what types of bars do you prefer on a Saturday night? Are you a brunch person on Sunday? Where should I go this weekend? The potential to leverage the “check-ins” of each user, as well as variables not explicitly provided by the user like weather, event, and even financial data clears a space to build a massive, yet incredibly local recommendation engine. A recommendation engine driven off of more than simply explicit categorical statements of “Thai Food in San Francisco,” and a smattering of polarized user reviews.

3) Involve Businesses. By incenting the supply side of the economic equation to use the product to lure demand, more users will end up joining to gain access to the deals offered– a mutually beneficial event whereby Foursquare can collect more location data, and companies receive a better ROI from their advertising dollars. If Foursquare also allows these businesses access to the wealth of collective intelligence contained in the place/time decisions, as well as the platform to deliver ads to users that meet a specific behavioral profile, you’ll witness nothing short of a revolution in product and service marketing.

4)Third-Party Application Development. http://www.CheckoutCheckins.com/ demonstrates both the primitive state and the potential that location data has. You can map where you’ve been. This app doesn’t even include the time data, but it provides a visual glimpse of the potential benefit of being able to place people’s behavior on a map? Where do males under 25 with an income over 40,000 spend their time?

Jesse Schell is right, games are here to stay, and as we gain the ability to measure more of what we do within our day, the ability for machines to harness collective intelligence will allow us access to better information by which we can answer the question of what we should be doing. Now.

Currently, Foursquare’s limiting factor is the act of checking in. To do so, a user must open the application, find the location they’re in from a list of potential locations, click the screen a couple more times until they’ve finally checked-in. This needs to become a zero click process. Additionally, the ability to pin-point a user’s location is still weak-sauce. GPS does not perform well in urban environments (where foursquare is most useful) and cell tower triangulation is fine for approximating, but it’ll struggle to identify what establishment someone has entered. But, the technology to detect when someone or thing has entered a store is not new. A bell on a door, or a ding when someone crosses through the entrance has been used in the analog world for years. Conceiving of a similar device that detects devices running foursquare in the background should serve a starting point for deeper thought on how to make the check-in a seamless process. There might even be a solution in mesh technology. If a business owner or store clerk broadcasts a unique code from their device running Foursquare in the background, any devices in a small radius would receive proximity notifications and be prompted to check-in at that location. In the same vein, the ability to “Bump” a check-in at a point of sale, much like a swipeless credit card could also yield faster check-ins.

If Marissa Mayer is right, and that 70% of an individual’s economic transactions take place within a 5 mile radius of our home, the ability to reach each individual with actionable information becomes a golden economic opportunity. Right now, we’re telling foursquare where we’re at. Soon, they’ll be informing us where we should be.