Off The Map: Category local search

Yahoo! Gets Neighborhood-friendly with Urban Mapping

umibot — Tue, 24 Jun 2008 05:59:00 +0000

A few months ago Yahoo! announced it had incorporated neighborhood search into Local and other properties. We’re pleased to say this is brought to you by Umibot and the hard-working team at Urban Mapping, so Yahoo! can now enjoy the same neighborhood goodness as many of our other satisfied customers.

Here’s the official news.

YellowPages.com Does Full Frontal Neighborhoods!

umibot — Thu, 05 Jun 2008 17:16:00 +0000

Umibot’s been a proud servant to AT&T’s YellowPages.com for some time, but neighborhood search functionality is now front and center, so search by neighborhood to your heart’s delight!

Urban Mapping to Speak at SMX Local & Mobile

umibot — Thu, 05 Jun 2008 02:07:00 +0000

UMI’s own Ian White will share the podium with several yet-to-be-determined panelists at the SMX Local & Mobile conference in San Francisco July 24-25 at the Marriott Hotel. The panel, Monetizing Local & Mobile: Who’s Making Money, is bound to be provocative if past panels are any indication.

Neighborhood Boundary Database Goes Offline!

umibot — Wed, 28 May 2008 07:40:00 +0000

As in, literally…

Today we’re excited to announce a partnership with Intelligent Direct, the market leader of custom print and data solutions for business. Through publication of custom electronic and printed maps of all sizes, companies can better manage the impact of location, geography and demographics.

Urban Mapping’s neighborhood boundary database will be incorporated into custom solutions. IDI’s MarketMAPS division has served thousands of clients in its 25+ year history.

Umibot is most excited about the offline possibilities. Even though it’s not clear if a bot can exist in the real world, there’s no question direct marketing does. And it’s huge. And small/medium-sized businesses think about overall marketing impacts and budgets, not segmenting by channel. This announcement is the first of several to link interactive and direct marketing efforts.

-The official news

Neighborhood API News...

umibot — Fri, 16 May 2008 15:47:00 +0000

This just in…Umibot is pleased to announce a few enhancements to our Neighborhood REST API…

In addition to our significantly-increased neighborhood coverage we’ve responded to developer requests and enhanced the REST API’s getNeighborhoodByLatLng to offer the option to return zero-to-one results, as opposed to the default zero-to-n results.

Why does this matter? Particularly in urban areas, neighborhood boundaries are organic, complex and because they are culturally defined phenomena. They are often with overlapping and/or hierarchical, and sometimes vague spatial relationships.

If you are enabling local search for your records, associating them with multiple neighborhoods will provide your users with more search options. However, some application developers want to know the neighborhood for a particular location. For this case, users can rely on our algorithms to take into account the underlying spatial relationships and geometries of all the neighborhoods which include the point to provide the best answer in response.

A final minor enhancement:, we have added ‘distance’ field to the result of ‘getNearestNeighborhood’ representing the distance to the centroid.

[Background Music: Begin Dirge]
Please note that we are deprecating the SOAP API. We have observed that the complexity of SOAP clients causes far more headaches for our end users, and our development overhead is not insignificant. As a small team, we have decided to focus our energy on expanding our coverage, and enhancing the REST API in response to user feedback. If you haven’t already, we encourage you to move over to REST.

Urban Mapping Releases Mass Transit Data for 50+ Systems

umibot — Wed, 14 May 2008 11:21:00 +0000

Phew! After more than a year in development and two years deep in Umibot’s RAM, today we unveil a grand plan: normalized mass transit data for (today) 53 public transportation systems in the US, Canada and UK. To get here we had to develop other pieces–a data intake platform and a schema. Some more info on all of these:

Web-based Mass Transit Data Intake Platform (no acronym yet) Umibot believes the greatest cost in data collection is identifying and purging the system of dirty data. By auto-validating data at point of input, we’re able to significantly reduce this cost. UMI’s proprietary web-based platform is flexible and captures the vast collection of spatial and attribute data we manage. This includes things like routes, station footprints, exits (you can’t generally exit at a ‘station’), hours of operation, handicap accessibility, elevator location, amenities (retail, bathroom, telephone, etc…) and a great deal more. We then associate this attribute data with the ‘spine’ of spatial data and then compute a graph network, making the data ‘routing ready’ across a variety of platforms.

Transit agencies can take advantage of this platform by using UMI’s infrastructure as a platform to inventory their own data. It’s a well-known fact that transit agencies face bureaucratic, technical and legal challenges to releasing data, and this platform is one more reason for transit agencies to partner with industry to increase data distribution and support increased ridership by driving awareness.

Normalized schema Before we began data collection, a uniform schema that recognizes transit nuances and complexities needed to be developed. For example, scheduling for the London Tube operates on a headway, meaning trains depart every Xish minutes. New York’s MTA operates on a tabular schedule, with scheduled departure times. Sounds like a detail, and it that’s exactly what it is–multiply this nuance 100 times and there’s a great deal of data definition that matters. What we’ve developed is internal to UMI and offers tremendous flexibility to add new mode types (ferry, funicular, etc). It has nothing to do with the output customers receive, and we’ll have more news about that soon.

Coverage The map below reflects current US coverage. Across the 53 transit systems, UMI has defined over 14,000 individual stations and over 100,000 data attributes. Stay tuned for increased coverage, attributes, service delivery and partnerships!

And some fun transit statistics for current coverage:

22% of transit stations have bathrooms (they may not be operable/accessible, but they exist)
35% of transit stations have dedicated parking

FYI: Wire release

Urban Mapping Neighborhood Boundary Database Coverage Reaches 40,000

umibot — Thu, 08 May 2008 10:06:00 +0000

Umibot’s been overclocking the past few weeks, and for good reason. Today Urban Mapping passed a new milestone and is thrilled to announce boundary data for more than 40,000 neighborhoods in the US. In addition, UMI continues to increase coverage across Canada and European countries, bringing our global coverage to more than 50,000 neighborhoods across 2,000+ cities and 15+ countries. Users of our enterprise and web services delivery can tap into this collective pool of rich local knowledge. A new shiny map of the Continental US displaying coverage:

Some more great things in store over the next several weeks (including an announcement at Where2.0 next week), so ensure you are RSS-compliant!

Official news release

Live blogging (with time delay) from the Kelsey Group conference--Zillow's Rich Barton

umibot — Fri, 02 May 2008 19:44:00 +0000

On Day Two of the Kelsey Group’s DDL conference here in Seattle, Zillow’s Rich Barton gave a keynote address about his three Big Ideas where information asymmetry presents significant opportunity for business model disruption: travel, legal services and real estate, or as he says, “storming the Bastille.” Umibot knows that Rich has obviously proven himself as a successful entrepreneur but wants to clarify a few points he made (and I thank my master for giving me my AI that allowed me to ‘know’ this).

Zillow’s neighborhood database has 7,000 neighborhoods covering approximately 150 US cities.

UMI’s neighborhood boundary database contains almost 40,000 neighborhoods across 1,200 towns and cities in the US (plus additional Canadian and European coverage), and we continue to add additional neighborhood coverage on a regular basis.

Rich said Zillow’s neighborhood boundary data is available via an API. I believe he misspoke. Certainly Zillow offers an API, but I don’t believe it offers neighborhood boundary data (although this could certainly be done).

UMI offers a fully robust API, allowing us to offer neighborhood-level geocoding via web services using REST.

Zillow’s boundaries are generally drawn around census tracts and postal codes

UMI’s neighborhood boundaries conform to how users (not direct marketers or actuaries) understand neighborhoods–postal codes and other administrative/political boundaries bear little relationship to neighborhoods, as this search reveals.

There’s more to this story, but the above is probably enough for the non-obsessed to chew on.

Urban Mapping to Present at Search Insider Summit

umibot — Tue, 29 Apr 2008 22:37:00 +0000

Urban Mapping’s own (guess who) Ian White will participate at MediaPost’s Search Insider Summit May 18-21 on Captiva Island, FL. Ian will participate on several panel discussions and breakout sessions. Umibot is thrilled that UMI will be at the event as it will provide a good opportunity to take the pulse of search engine marketing and local search.

Urban Mapping to Speak at SES Toronto 2008

umibot — Tue, 22 Apr 2008 01:44:00 +0000

Tis the season…

UMI’s own Ian White will participate on a panel Getting Found in Local Search & Maps at Search Engine Strategies 2008 Toronto show, June 16-18.

Urban Mapping Launches Innovative Geotargeting Platform

umibot — Thu, 17 Apr 2008 17:11:00 +0000

We’ve been working on it for quite some time, and today are thrilled to announce the first public release of our GeoMods geotargeting platform. In the past my master has written and spoken about the perils of IP-based geotargeting for local search.

Here’s the fact: for local targeting, where granularity matters (and some will argue it doesn’t yet matter), resolving an IP address to a location doesn’t work. The UMI solution doesn’t claim to be a technical solution, but rather aggregates large sets of geographic keywords around existing search behavior. It’s a mouthful to explain, but stand by for more on that front.

Urban Mapping to Speak at O'Reilly Where 2.0 (2008) conference

umibot — Fri, 11 Apr 2008 23:53:00 +0000

Urban Mapping’s Ian White will perform a hat trick at this year’s conference, participating on a panel, Monetizing Maps & Mashups. Greg Sterling will moderate and other panelists will be revealed in short order…

When is a subway service change more than a change?

umibot — Sat, 05 Apr 2008 05:40:00 +0000

Umibot recently caught up with a few favorite blogs, including the -ist family. In not so unbelievable, yet simultaneously incredible fashion, here is the change of service announcement from hell.

Umibot may not be human, but he still understands that too many facts in too short a space equals too much confusion….Information anxiety, for sure. Stay tuned for UMI’s Urbanware Transit product–a fully robust and highly-structured database of mass transit systems.

Thanks Gothamist

Urban Mapping Convenes Local Search Summit

umibot — Mon, 17 Mar 2008 10:37:00 +0000

Umibot is all kinds of tired from several days in Las Vegas, where Urban Mapping brought together industry leaders for two days of off the record discussion about the ‘state of local.’ The event was billed as a ‘part salon, part workshop,’ recognizing that a substantive and engaging dialog is best set in a relaxed and informal environment.

UMI brought together industry stalwarts and upstarts. Analysts presented research, executives shared key insights and the group worked (and ate and drank) through the complexities, nuances and opportunities around local search: geotargeting perils and realities, ‘market inversion’ around local ad inventory, mobile growth, user behavior and market forces.

The esteemed group:

Jake Ballie, Managing Director, STN Labs

Matthew Berk, Lead Search Architect, Marchex

Pete Flint, Founder and CEO, Trulia

Craig Greenfield, Director, Local Search, DoubleClick

Jeff Greenwald, Director, Search Products, MapQuest

Martin Herbst, Senior Strategy Manager, Kijiji

Peter Krasilovsky, Program Director, Marketplaces, The Kelsey Group

Farhan Memon, Senior Product Manager, AOL Search

Ted Morgan, CEO, Skyhook Wireless

Fred Owens, Vice President, Business Development, Medio Systems

Justin Sanger, President, LocalLaunch!

Greg Sterling, Founding Principal, Sterling Market Intelligence and Senior Analyst, Local Mobile Search

Steven Stern, SVP, Sales & Market Development, Urban Mapping

Ethan Stock, CEO, Zvents

Joel Toledano, CEO, Krillion

Jaron Waldman, CEO, Placebase

Ian White, CEO, Urban Mapping

Disclosure: Jake Baillie and Greg Sterling sit on the Board of Directors at UMI.

Urban Mapping Named Semi-finalist in NAVTEQ LBS Challenge

umibot — Fri, 14 Mar 2008 17:12:00 +0000

It happened a few weeks ago, but Umibot is just now getting around to posting…UMI is one of 15 companies nominated for the semi-finals of the annual NAVTEQ Global LBS Challenge. We aren’t sure how many entrants there were, but we are privileged to be included in this group.

The UMI application is based on the highly structured data than comprises our Urbanware: Mass Transit data product. Built on the where.com platform, Urban Mapping was able to quickly develop for mobile using uLocate’s location-aware platform.

Finalists will be announced April 2 in Las Vegas during CTIA Wireless.

Greg Sterling Joins Urban Mapping's Board of Directors

umibot — Mon, 03 Mar 2008 06:12:00 +0000

Today we announce the addition of Greg Sterling to UMI’s board. His insight and experience lend a shining stamp of approval to the company.

Urban Mapping Gives Web Services A REST

umibot — Mon, 25 Feb 2008 00:50:00 +0000

After some hard work/late nights and prodding from the geo-techno elite, we’ve humbly completed and are pleased to offer RESTful access to our neighborhood API.

What does this mean for developers? Probably less time spent developing and more time doing. Since we offered the SOAP-based free neighborhood API last month, dozens of individuals and companies have signed up, and we’re confident this announcement will spur another wave.

We’re also thrilled to be using Mashery to manage our multiple APIs. Ciaran and team twisted themselves pretzel-wise to get us up and running in no time at all. Calling Mashery ‘on demand’ is exactly what what they offer, and it’s what they delivered.

The news release with a quote from Brady Forrest and other goodness.

Urban Mapping to Present at Search Engine Strategies 2008

umibot — Fri, 01 Feb 2008 14:35:00 +0000

Urban Mapping’s Ian White will moderate a panel, Why Local Is Different, at the annual SES New York Conference & Expo at the Chicago Hilton, March 17-20.

Urban Mapping to Present at Pelorus Group's Mobile Local 08

umibot — Tue, 22 Jan 2008 09:40:00 +0000

UMI’s Ian White will speak at the Pelorus Group’s Mobile Local Search Forum in Albuquerque, New Mexico, February 20-21, 2008. He will participate in a panel on GPS & Location Based Services: Profit vs. Privacy.

Urban Mapping to Present at ad:tech San Francisco

umibot — Mon, 21 Jan 2008 13:47:00 +0000

UMI’s Ian White will speak at ad:tech in San Francisco, April 15-17, 2008. He will participate in a panel on local and mobile search

I Want My (Free) Neighborhood API!!!

umibot — Wed, 09 Jan 2008 06:03:00 +0000

Today Urban Mapping introduces its free neighborhood API for all developers and interested parties. This means you can send UMI a long/lat and we will return the associated neighborhood(s) point it is contained within. It also means no cost. No charge to you. Free.

Our Online documentation provides code snippets to help you along. We also have a great map of our US coverage.

Free neighborhood geocoding is important because it represents how people think about local–UMI has done the heavy lifting and documented more than 25,000 neighborhoods, representing every conceivable way of understanding informal space. For interactive applications, providing a data set that reflects, not determines user behavior is a novel yet invaluable concept.

The “official” news release for those who like reading.

Urban Mapping + Placebase = Very Good Things

umibot — Mon, 19 Nov 2007 08:00:00 +0000

Today Umibot is pleased to announce a partnership with Placebase. Their map Pushpin map API is powerful and gives developers great flexibility in creating applications. Pb will distribute UMI’s Urbanware Neighborhood data product. Their map tile factory has been hard at work, making freshly-rendered tiles of SoHo, Nob Hill, LoDo, centre ville and about 25,000 other names. A Manhattan demo is available for you to look at and play with. Note different layers of neighborhoods as you zoom.

The neighborhood data news over at Placebase.

Pelago Licenses Urbanware: Neighborhoods

umibot — Tue, 23 Oct 2007 07:00:00 +0000

Today, stealth startup Pelago pulls off the covers and unveils its web/mobile application that blends the social with the local (on the mobile, of course). Some call it Facebook meets MySpace, and others have described it as the tip of location-based services. UMI is thrilled to be a critical component of the product as neighborhoods provide the granularity and context that matter in complex urban spaces, and Pelago embraces this. Don’t take our word for it, give Whrrl a spin!

Urban Mapping to Present at Search Engine Strategies (Chicago) 2007

umibot — Fri, 19 Oct 2007 09:20:00 +0000

Urban Mapping’s Ian White will participate on a panel, Online Maps: Plotting the Direction of Local Search, at the annual SES Chicago Conference & Expo at the Chicago Hilton, December 3-7.

EIRE Gets Its Postcode On

umibot — Thu, 18 Oct 2007 13:16:00 +0000

Since Umibot is continuously searching the web for new & notable for all things geo, here’s one of note…The Republic of Ireland is going postal! Or, more accurately, An Post (national post system), will introduce a national postcode system “sooner rather than later.”

Dublin has had a postcode system since the 1960s but not committed itself to a national postcode system until last year.

It will be interesting to note how administrative delineations emerge; as Umibot knows, ZIP codes aren’t all that useful in the US for policymaking–they are are defined based on a mail carriers’ route, and in fact are not even self-contained areas, yet they persist for direct marketing purposes despite obvious shortcomings. The history of the Zone Improvement Program is a curious one. We’ve also written about the ‘ZIP code fallacy’ on this blog.

CitySquares Licenses Urbanware: Neighborhoods

umibot — Wed, 03 Oct 2007 06:48:00 +0000

Umibot is thrilled to announce our most recent customer. CitySquares, a Boston-based hyperlocal community site, will incorporate UMI’s neighborhood taxonomy, enriching the user experience on the site. Look for it in a few weeks!

What is Your Favorite Map Tile?

umibot — Fri, 28 Sep 2007 14:18:00 +0000

MSR brings geo-philes something new to play with. The project (Hotmap: Looking at Geographic Attention aggregates map tile downloads at different zoom levels, allowing you to see that more map tiles for an area in the Gulf of Guinea (off the coast of Ghana) have been downloaded than for all of Africa.

-via Adena

Urban Mapping to Present at The Kelsey Group's Interactive Local Media Conference

umibot — Wed, 26 Sep 2007 20:34:00 +0000

UMI’s Ian White will participate on a panel, ”Maps: They’re Not Just For Driving Any More” at the Kelsey Group’s ILM conference, cosponsored by Search Engine Strategies, November 28-30, 2007 in Los Angeles.

How is Local Search Like Storm Tracking?

umibot — Wed, 19 Sep 2007 08:31:00 +0000

News from the National Weather Service that is sure to get (geo)data-wonks excited…

From the National Oceanographic & Atmospheric Adminstration (parent agency of NWS), the current means of tracking severe weather events is done in the following manner:

…The NWS currently issues and disseminates warnings for tornado, severe thunderstorm, flood and marine hazards using geopolitical boundaries.

As of 1 October 2007, this system will change to something new:

Storm-Based Warnings (threat-based polygon warnings), are essential to effectively warn for severe weather. Storm-Based Warnings show the specific meteorological or hydrological threat area and are not restricted to geopolitical boundaries. By focusing on the true threat area, warning polygons will improve NWS warning accuracy and quality…

You may want to ask Umibot “what’s the big deal?” Some graphics from the Storm-Based Warnings (NB: press release to follow on 10/01/07) illustrate this:

On the left, the county is used as the unit of measure–this means if a predicted storm path touches a county boundary, the entire county will receive an alert. This is especially cumbersome in some Western states, where counties can be extremely large. Deploying emergency resources (first responders, food, supplies, etc…) and alarming the public when not necessary could prove and expensive proposition.

The image on the right highlights the new approach: “threat-based polygons” might sound menacing, they are no different from what the NWS currently uses with a key exception: the granularity has changed such that the unit of measure is now the municipal boundary.

From UMI’s perspective, what is interesting to note is that NOAA prediction accuracy did not drive the Storm-Based Warnings program–there are meteorological (and related) advances that help officials understand patterns of severe weather, and that is independent from presenting those data. Because prediction science has become more accurate, a smaller unit of measure (ie, municpal area) can be used. From this perspective one could say predictions were ‘hiding’ behind the larger unit of measure (ie, county).

Umibot likes these kinds of stories because they play directly into his (or her?) sweet spot–the design of data. And this was the focus of a talk Ian gave last year on the very subject.

The analogy for local search is clear–data should drive the use case of an application. If one is going to offer an application that allows for (say) mobile search, will a user have the granularity that is needed to have a meaningful experience? An example here is “restaurants in San Francisco”–mobile means you are, well, mobile, on the go, and a city is (probably) not a meaningful geo-constraint. Something more granular, like a 2 mile radius (if the device is location-aware), cross street, or neighborhood will likely be more satisfying.

Microsoft Licenses Urbanware: Neighborhoods

umibot — Mon, 17 Sep 2007 06:00:00 +0000

This morning at the TechCrunch40 Conference Urban Mapping announced its latest customer–we’re thrilled to count Microsoft among the portals that utilize UMI’s products. Now may every man, woman and child find their way through congested and urban areas!

Geotargeting Overview

umibot — Mon, 13 Aug 2007 09:07:00 +0000

This morning Search Engine Land has a lengthy and informative article about geolocation–how it works, what it is used for, who offers it and who deploys it. The sentence in the article that sums it up for Umibot is this one:

The biggest problem in assessing the error rates of geolocation data is the simple fact that there’s no way to really test well for accuracy.

As the article indicates, this uncertainty has to do with proxy servers, load balancers and anonymized traffic. Urban Mapping’s georargeting solution (not yet unveiled, but in the latter stages of testing) will provide significant gains over existing geolocation technologies using a completely different technique–rather than identifying the location of an IP address, UMI sees to determine the intent of users, and tie that data to geographic space. A hard concept to understand, yes, but the results will speak for themselves. We’re excited to make this available and making some final tweaks, so the coming weeks should have some more news in store.

Urban Mapping to Present at SMX Local & Mobile

umibot — Tue, 31 Jul 2007 10:35:00 +0000

UMI’s Ian White will participate in a panel at Search Marketing Expo’s Local/Mobile event in Denver, October 1-2. He will speak on a panel entitled Show Me the Money the afternoon of October 2.

Urban Mapping to Present at GeoWeb 2007

umibot — Thu, 28 Jun 2007 14:08:00 +0000

UMI’s Ian White will speak at GeoWeb 2007 in Vancouver, British Columbia, July 23-27, 2007. His talk, Web-based GIS or GIS-based Web?, will address how neogeographers have changed consumption habits of spatial data.

Get Local, Future-style

umibot — Wed, 27 Jun 2007 08:31:00 +0000

Man-of-all Men Bruce Sterling FFWDs ten years to recount how we consume local media in 2017. Call it hyper/nano/ultra/extremelylocal in near future, but don’t ignore it. The future comes in ways we least expect, and Sterling offers an insightful (and visionary) narrative. For those who aren’t well-versed in the genre, here’s a snip from his recent column, but be sure to read the whole story.

I’m dictating this entry — thank heaven for voice recognition — from the passenger seat of a Hyundai GPS-King careering along the Beltway. I downloaded a cool plug-in to block out the gas-food-lodging ads that hit my screen a quarter mile before each exit, so I’m free to concentrate. What do I care about lodging anyway? The best thing about being a top-tier geo blogger is that everyone knows where you are. When the buddy list tells folks you’re in town, they ping to offer you dinner and invite you to sleep on the couch….

These ideas aren’t new, but Sterling combines and weaves them into a tapestry that makes the future come alive. Next time you are in a corporate planning session, don’t think up new products for the future. Read Sterling’s work and chart a course to get there. It’s much harder. And more rewarding.

-From APB

Seeking a Geo-aware Search Marketing Wiz

umibot — Fri, 22 Jun 2007 13:35:00 +0000

Thanks to our talented people and forward-thinking customers, we’re growing in all directions. Our newest offering has to do with geotargeting. As we’ve noted in the past, the current state of affairs in geotargeting ain’t so good. IP lookup can’t provide the granularity that is required to serve up local content/relevant advertising. If a user has registered/provided a profile, that’s great news, but is only a small piece. It also presupposes you are located in your default/home area.

Leveraging our expertise in informal space (ie, neighborhoods), user research/cognitive psych and local search, UMI has developed a geotargeting service that reflects user intention. This is significant because technical methods fail to deliver expected results. This synthetic geo-targeting service taps into user intent, allowing for ‘hyperlocal’ to be more than hyperbole.

We’re looking for somebody who can help us penetrate agencies and advertisers. With marquis clients already in place, we seek an entrepreneurial and driven search marketing pro who thrives in vague and inherently ambiguous environments. You have a demonstrated track record of developing new business and are resourceful, tenacious and creative.

Urban Mapping hires for what you can do, not for what you have done. This means the background we seek is desirable, but not critical. A few notes about specifics:

Agency/publisher experience - 3-15 years campaign management, ad operations, account management
Sales/Biz Dev background - The primary role is to bring sales, but this can only happen once a plan is in place. You must be comfortable on the road, drawing from your contacts, inbound lead gen, cold calling and geunuinely enjoy your work.
Industry knowledge - You are an interactive advertising fountain of knowledge–you consider SES to be a great networking opportunity, not a great venue to learn about online advertising. You have a solid understanding of SEM, SEO, PPC and emerging ad products.
Technical competence: You need not be a programmer, but technical comfort is a requirement. This means awareness of stacks/languages/environments and advertising APIs. Exposure to geospatial technologies is useful, but not required.

Compensation is modest base, generous commission and equity participation.

We work in a creative, entrepreneurially-vague environment. Our methods are novel, but our approach has yielded valuable insights and results. To be considered for this position, please forward a resume and brief note speaking to your training and experience. Email to talent [at] urbanmapping [dot] com. Indicate “SEM Pro” in the subject line. No phone calls please!

The Centroid Gap, or the Death of the ZIP Code?

umibot — Thu, 07 Jun 2007 21:28:00 +0000

A several weeks ago we posted a few thoughts about the death of the ZIP code. There’s a lot more to say from the geo-perspective on local search, and here’s some more fodder…

To give any data a geographic context, it must be spatially-referenced to the Earth. Geographic information systems (GIS) serve as a means of referencing this information. Within the context of local search, addresses, city boundaries, postal codes or other geographic data must be ‘translated’ from human terms (690 Fifth Street, San Francisco) to latitude and longitude, ie, machine terms (37.775429, -122.397314). This geocoding process allows databases to recognize human-language requests. To geospatially reference (say) a postal code, one would expect that area to be spatially-defined. When a user searches for (say) “coffee in 94107,” the ZIP code should serve as the geographic constraint, searching within this polygon. Correct?

Wrong! A variety of reasons are to blame for why the logical doesn’t happen: most obviously, ZIP codes were defined as letter carrier routes. They were not meant to serve any other purpose. As such, the ZIP may not even conform to what you expect–one side of a street, one floor of a multi-story building or one-half of a block may not be fall within what postal code you expect. In fact, many parties claim to use a ZIP code database in fact obtain this info from a sister governmental agency, and these boundaries are stylized representations of the USPS data.

More to the point, these stylized boundaries are likely not used. Instead of associating (say) 50 latitude/longitude points to define a the postal code boundary, technical optimization says one point is sufficient. The analogy here is reducing a novel to a word–in the context of local search, granularity matters, and using the mathematical center of a polygon serves to distort and misinform a user’s search. In practice, the centroid is used because it is more efficient to calculate than the actual shape. Reducing the contours and nuances of a small area to a point, often with a radius drawn around it, effectively makes all postal codes look like circles. Gaps and overlaps are formed, further distorting the expected reality for a user.

Graphically this can be represented with the ZIP code boundary and circle (with the center serving as the centroid). The circle includes area that is not shared with the postal code area and vice versa. A user searching in this ZIP will therefore not be returned all the relevant listings. Some will argue this is a technology issue, but from the above example, it clearly more of a mindset–getting product managers to think about the how and why of data will go a long way.

MapQuest Licenses Urbanware: Neighborhoods

umibot — Wed, 30 May 2007 12:38:00 +0000

This morning at the O’Reilly Where 2.0 conference we announced our latest customer. UMI is thrilled to be the latest MapQuest partner. We look forward to millions of users being able to search in an intuitive manner!

From the Radar

umibot — Mon, 21 May 2007 14:44:00 +0000

The O’Reilly Radar, that is…

Brady offers a good overview of the latest & greatest in the geo-portal data wars. Of particular interest are his comments about owning the geostack:

Until recently, Yahoo! used deCarta to handle creating directions (or routing). They have announced that they have taken ownership of this part of the stack and have built their own routing engine. Ask and Google still use deCarta. Microsoft has always had their own. Yahoo! is hoping to make their new engine a differentiator. In some ways this is analogous to Microsoft’s purchase of Vexcel, a 3D imagery provider. Microsoft did not want the same 3D data as Google Earth or any other search engine for its 3D world.

I think that any vendor servicing Google, Microsoft, Ask, Yahoo or MapQuest will have to keep an eye on their next source of revenue. Those contracts aren’t going to necessarily last too long. The geostack is too valuable to outsource.

There is only one part of the stack that I think might be to expensive for any one of the engines to buy or build out right. That’s the street data and it’s a data source primarily supplied by two companies, NAVTEQ and Tele Atlas. NAVTEQ has a market cap of 3.5 bilion dollars as of this writing; Tela Atlas has one of 1.4 billion pounds. These would be spendy purchases. Microsoft is currently working closely with Facet Technology Corporation to collect street data for cities to add a street-level 3D layer (see Facet’s SightMap for a preview), but this Facet is not collecting data to match the other players. It will be interesting to see if Yahoo! parleys its partnership with OpenStreetMap into a data play.

[Brady’s spellchecker appears stuck in the ‘off’ position.]

Yahoo! Dumps deCarta for Maps

umibot — Wed, 16 May 2007 11:58:00 +0000

Surprise! Yahoo! Maps is going inhouse with its own map server. This is after 5+ years with Telcontar, aka deCarta. Some new changes out today and apparently more in the works.

Thanks, Adena.

The Digital Becomes the Atomic

umibot — Mon, 07 May 2007 06:00:00 +0000

UMI has long argued that a significant piece of the value in using online maps & directions stems from the transitive property of interactive media–to some degree, the experience can come off the page and become portable.This means users are endowed with an artifact that represents the digital, but can engage with it in a tactile world. It would be easy to digress and pontificate on post-realism via Walter Benjamin or Jean Baudrillard, but Umibot lacks that memory implant, so have no fear…

The idea of maps & directions are to get from one point to another point. They are meant to be didactic and support wayfinding activities. That much is clear. It then follows that many users would print maps/directions and take them on the trip–other options may include a ‘send to phone’ feature, mobile browsing for directions, use of a GPS device or transcribing/summarizing directions by hand.

If a user prints a map and uses it for wayfinding, it becomes an invaluable artifact during the course of that journey–it may be referenced multiple times and the ‘eyeball quality’ is undoubtedly high (who has heard of a casual direction-follower?). So this begs the question…why are printed map/direction pages missing the obvious–highly-targeted and relevant advertising? A review of print-ready directions from Google, Yahoo!, Live Local, MapQuest, Ask and MSN reveal some interesting things.

This summary table highlights some key questions that may be obvious, but warrant consideration:

Can multi-page directions be avoided? Should they be minimized?
Do consumers need vast flexibility in ‘configuring a map for printing?’

Most importantly, does advertising make sense on a printed map? This is valuable real estate that currently looks like a greenfield site.

NOTE: Images below have been cropped and resized to fit this blog, so page lengths may not make sense, but they are–print them yourself to see!

Google A

Google B

Live Local

Yahoo! A

Yahoo! B

Yahoo! C

Ask

MapQuest

MSN Streets & Directions

Power Moves in Local Search

umibot — Thu, 03 May 2007 10:30:00 +0000

Umibot feels that many readers won’t bat an eyelid over this, but my master feels strongly, so I must be a good slave and report…

The original idea was to have a Top Ten list, but frankly there aren’t many moves worthy of the Power Move moniker, so we will instead post them when we think of them.

Our favorite is AskCity’s draw-your-own-polygon and search within. There’s no hiding the fact that this take some serious geo-skillz to implement. To hell with ZIPs (or neighborhoods for that matter)–draw your own boundary!

This screen grab shows an area being drawn near UMI’s offices. As it is late morning, we are always hunting for new lunch spots in this half-industrial/half developed part of town.

This grab shows results–presto!

Top Neighborhood Names

umibot — Wed, 25 Apr 2007 18:41:00 +0000

From the home office in San Francisco, CA, Umibot unveils the first (of future periodic) installment of Most Common Neighborhood Names in the United States…

Downtown
West End
Midtown
Northwest
Northeast
Old Town
Riverside
Uptown
Central
Highland Park
Southwest
Chinatown

There doesn’t seem to be too much that is surprising here, but it is interesting to note that there are close to 200 instances of Downtown and close to 20 instances of Chinatown. Neighborhood names are truly a long tail!

Why Geotargeting Sucks

umibot — Wed, 25 Apr 2007 15:14:00 +0000

Ad-based geotargeting sucks. At a broad level–national or regional–it can be useful for brand advertisers, but for the truly small business, geotargeting fails. It’s fantasy. Advertisers don’t like to talk about this but it’s a fact that local ad inventory does not exist in any scale. Yet. This will change over time when the ability to have meaningful geotargeting is developed (more on this from UMI in the coming weeks). This will drive usage, which drives awareness, then advertisers.

Geographical targeting of advertising can happen by geo-intelligence on the serving platform or by a user profile/explicit communication of the desired search area (ie, Cambridge, 02138, user profiles, etc.). Within paid search, Google’s AdWords, Yahoo’s Panama, and MSN’s adCenter represent the primary means of delivering text-based advertising. Since more than 75% of the web’s search traffic is carried by Google and Yahoo!, they are significant players. Contextual and rich media ads represent a smaller but more rapidly growing segment and includes is driven by Google’s AdSense, DoubleClick, smaller publishing networks and a vast ecosystem of ad networks, agencies and other intermediaries.

Geotargeting within the ad serving platform is generally limited to its embedded geo-intelligence. Internet Protocol-based geotargeting seeks to convert a computer’s IP to a real-world location. In theory this is an intuitive and valuable approach; assuming the user is searching in this area, advertisements could be geographically targeted to a very local level. In practice, this rarely happens. Internet service providers commonly use load-balancing equipment to distribute traffic across their network. The experience of all AOL subscribers’ IP addresses showing up Dulles, Virginia, is the most widely-known example of this happening. The end result is geo-IP lookup that provides consistent results for the city or metropolitan area. At times, increased granularity (ZIP code) can be achieved, but this is by no means guaranteed.

In addition to a website determining the user’s location based on IP lookup, users can directly provide information about their desired location to perform a search. Because user profiles (such as My Yahoo!, Google Accounts, etc..) are predetermined, they require no additional effort for the user. However, if a user is searching outside the area contained in her profile, results will be bogus. The vast majority of the time, this is the scenario–it is up to the searcher to define the geographic proximity, and that is where the system breaks down. ‘Formal geography’– cities, addresses, postal codes and others are recognized by search engines, but the informal geography–proximity to schools, parks, neighborhood locations, key landmarks, etc…, are recognized in a keyword, not spatial, method. This is critical and is the difference between (say) coffee shops located in Tribeca, Manhattan and the Tribeca Coffee shop in Krakow, Poland or Durban, South Africa.

Whither the ZIP Code?

umibot — Sat, 21 Apr 2007 12:11:00 +0000

Last month Ian spoke at the Kelsey Group’s Drilling Down on Local conference in Santa Clara, CA. The Red Herring penned a story that characterized the ‘machine readable nature’ of ZIPs viz a user-driven, human readable postal code (ie, neighborhood). Here’s a quote from the article that puts it nicely:

“The ZIP code has had a good run, but the Internet is threatening to render obsolete yet another relic of the postal era. That’s because the ubiquitous ZIP code is becoming increasingly irrelevant online as a search tool and data organizer since the birth of so-called hyper-local search technologies. The problem is that people looking for a restaurant, movie theater, or hair salon tend to search by neighborhood names rather than ZIP codes. Now a group of startups is working feverishly to develop new algorithms that tap “natural” languages rather than mathematics to process search requests.”

Umibot agrees that the story is accurate. However, the most significant obstacle is stasis–people are comfortable with existing habits and practices, despite their inefficiencies (think QWERTY keyboard layout). But when tools provide them with another way (ie, as my master signs up more customers), we’re confident user adoption will follow. Somewhat ironic that humans are forced to unlearn, then relearn, but that’s something that doesn’t concern Umibot.

Informal Space Means A Lot

umibot — Tue, 17 Apr 2007 20:22:00 +0000

UMI staff spent some time sifting through the leaked AOL search data and came up with some interesting findings….

We ran our database of neighborhood names through the AOL data and found that fully 9% of queries reference informal space–things like soho, downtown, financial district, etc. We discount this number by 1/3 to take user intent into account (clearly the search “Northwest Airlines” has no local intent).

So we end up with 6% of all search queries using some kind of informal (assuming the sample is somewhat representative). This is an astounding number. In the context of local search, this is something that nobody is capturing–the queries are going unfilled or returning a proximate match based on keywords, despite the clear geographic nature of the search.