Facebook, Waze and the Falling Cost of Mapping Technology | Street Fight

Facebook, Waze and the Falling Cost of Mapping Technology

Facebook, Waze and the Falling Cost of Mapping Technology

mobile-phone-map-searchFacebook is reportedly in serious talks to buy social mapping app Waze for a whopping $1 billion in cash and stock, an acquisition that would underscore the growing importance of mapping and navigation functionality to the Web’s major players.

A lot of digital ink has been spilled about why Facebook would want to snap up Waze — namely, the company’s 40 million mobile users. But the discussions also verify the once-unthinkable reality that a startup could scale a navigation and mapping service into a billion-dollar business in five years. A number of technological trends have set the stage for Waze and other mapping and navigation startups to compete. Here’s a look at a few of them:

A Historical Accident
Suppose I’m a Pittsburgh-based salesperson, and I need to stop in Toledo, Cleveland, Akron and Cincinnati to see clients. What’s the shortest possible route that visits each city without backtracking and returns to Pittsburgh? That’s called the traveling salesman problem, and it’s one of the quintessential questions of algorithmic theory. The problem is often solved using graph theory, a way of storing, traversing, and processing complex relationships, which has traditionally served as the underlying technology for navigation services.

Long a backwater of computer science, the technique has seen a renewed emphasis over the past few years thanks to Facebook and its open graph. Facebook uses graphs, which are a collection of vertices and edges or links between them, to help reveal the sea of connections between people in the same way that maps help us navigate between places.

“Graph computations were cute, and everyone knew they would be important some day, but there was no serious driver for investment before Facebook showed up,” says Scott Rafer, founder and chief executive of Lumatic, which makes a pedestrian-focused navigation app called Citymaps. “So you had a million PHDs that suddenly cared to make it work.”

What happened was a “historical accident” of sorts, says Rafer.  The investment, and subsequent leap forward, in graph computation, which Facebook sparked, indirectly led to a dramatic decrease in the cost of running the complex graph database algorithms used to get people from point A to point B. That led to a what Rafer views as a classic innovator’s dilemma, in which a new set of algorithms created a more efficient way to make the same product at a fraction of cost. These new algorithms also allow smaller companies like Waze or Lumatic to create better mapping services, accounting for additional information like user-generated traffic data to optimize a user’s route in real-time.

Cheapening the ‘Stack’
What’s happening with navigation is one example of a cheapening of tools that’s occurring across the mapping “stack” (the term used to describe the combination of technologies that goes into making maps work). OpenStreetMap (OSM), for instance, has created an open source geospatial dataset that companies can use to cheaply build the underlying map. Firms like Foursquare have used the company’s data, which provides the basic geospatial information indicating a river or city boundary, to build the customized maps that underlay the function as the foundation for their services. A number of other open-source technologies have emerged to make rendering this information and showing it on a website easy as well.

Waze doesn’t use OpenStreetMap, but the company has thrived by finding a cheap alternative to traffic data. By crowdsourcing traffic information from users’ mobile devices, the company was able to forgo the massive costs associated with hiring and managing a fleet of cars to survey the roadways.

Unbundling the ‘Stack’
Today, Google sells its mapping services as a bundle in the same way that a cable company might package ESPN, AMC and the Oprah Winfrey Network into one flat month. Given its strong position in the marketplace, the Web giant has an incentive to lump all of its features into one product, whether you need it or not.

But Rafer believes that the cheapening of tools and the increasing need for mapping and navigation services on mobile will trigger an unbundling of the stack of mapping services, resulting in a number of layer-specific services cropping up. Companies like Lumatic will be able to focus on a specific layer like pedestrian routing (or driving, in the case of Waze), and bring that technology to a host of different consumer and marketing use cases on a software-as-a-service model.

Rafer’s argument is, of course, self-serving, but he’s peddling a familiar story for information-related business on the web. The web fragments industries that justify scale based on either the collecting (mapping) or distributing (media) of information. And, as Andy Weissman points out, unbundling is a necessary corollary to that process: “What if the power of connected networks such as the Internet is that they unbundle all that came before them? They disintermediate incumbent industries but also do the same to any new attempts at re-aggregation?

Steven Jacobs is Street Fight’s deputy editor.

Waze’s business development chief Andy Ellwood will be joining us on June 4th in San Francisco at Street Fight Summit West. Register now to learn from and network with some of the brightest minds in hyperlocal. Click here for more info.

2 thoughts on “Facebook, Waze and the Falling Cost of Mapping Technology

  1. what is currently the best service to solve the actual traveling salesperson dilemma? I have tried many of them, just looking for the very most efficient for optimizing 100-150 stops at a time.

Leave a Reply

Your email address will not be published. Required fields are marked *

Name *

2 thoughts on “Facebook, Waze and the Falling Cost of Mapping Technology

  1. what is currently the best service to solve the actual traveling salesperson dilemma? I have tried many of them, just looking for the very most efficient for optimizing 100-150 stops at a time.

Leave a Reply

Your email address will not be published. Required fields are marked *

Name *