So to those not in that industry here is the best I can do in summarizing what’s going on here.

The amount of display ad inventory available online is absolutely massive… far more than the supply of advertising dollars chasing it… so the price someone is willing to pay to serve any old advertisement to a random Internet user is pretty negligible. Meanwhile, the internet advertising industry long ago went down the path of selling itself as a data intensive, highly measurable and result oriented medium… and for better or worse is generally stuck with that description.

So… the name of the game nowadays is to not just serve anyone on the Internet any old ad and call it a day, but to serve a very specific group of people, sometimes a very specific ad, and measure what happened afterwards to see if it ‘worked’ in terms of driving clicks or purchases… rinsing and repeating until one gets the desired result or gives up and tries for a new result instead.  The more highly correlated a given piece of information is with some desired activity like a click or purchase, the more valuable it is.

So some folks are making tens of millions of dollars doing very little more than going to an open advertising exchange and buying low cost impressions generated by people they know, and then adding in the information they have on those folks in order to more effectively target ads in a game of information arbitrage.

The vast majority of folks are focusing on the part of the equation around WHO will be shown what ad… which can be things like people who shopped online for flat screen tvs in the past 30 days, or people who checked an online stock portfolio in the past 24 hours, or someone who just read reviews of new cars on an auto oriented site.

A great recent example of this is a company called Magnetic (http://www.magnetic.is) which just raised $5 million in funding some top VCs, and a company that PaidContent thinks could be part of the next big wave in online advertising.  What Magnetic  does is provides advertising re targeting data off of searches conducted at one’s site. So not only can site owners continue to run a Google AdSense for search program directly on their own site, but let’s say you’re running a car blog and someone searches for “Ford Mustang” on your site and later heads over to  somewhere like cnn.com to read about the disaster in the Gulf.  Ordinarily there is no way to know that a reader on cnn.com reading about the disaster in the Gulf may be interested in a Ford Mustang, but using a system like Magnetic allows CNN to directly get this information and try to use it to charge more for their ads, or alternatively ad buyers for someone like Ford may not even concern themselves with specific sites and instead simply buy people who have searched for their brand or products wherever they may go across the web, through purchases of ‘individual cookies’ via blind advertising inventory exchanges. 

Another similar example is aCerno which was recently acquired by Akamai for $95 million. aCerno uses consumer shopping data gathered from a co-operative of approximately 550 major e-commerce sites, to re target advertisements across the web based on their online shopping behavior. 

The key words to keep in mind about where the industry stands today is terms like “shopped online”, “checked an online portfolio”, “read an auto site”… notice one thing in common here… all these behaviors are taking place in front of a computer screen. But what about the vast majority (95%) of the times when all those folks walked into a Best Buy store, Fidelity retail brokerage or stepped foot onto a Ford auto lot to do their commerce the old fashioned way offline?

There is no reason why this game of information arbitrage needs to be limited to purely online behaviors, or to the traditional browser of the PC based Internet.

Is a guy who spent three and a half hours sitting in Yankee Stadium four separate times last month probably a better prospect to buy Yankee hats, mugs, and jerseys gear than the general public? You betcha. 

Is a user who spent 45 minutes at a local Ford dealer lot last Saturday, potentially someone in the market for a car with higher than average intent to purchase a Ford vehicle? Probably.

So you have to think that it won’t be long before all of that algorithmic, arbitraging media trading that we’re seeing online these days begins to bleed over into the world of offline meets online, using location data at the center, in fact it’s nearly here.

Now this could very easily turn into another rah-rah post about why mobile social applications like Foursquare, Gowalla, Loopt and MyTown are going to take over the world… they get you to fork over information about your whereabouts and that information can be digital adverting gold.

But I am not sure I am ready to concede that this is something for mobile social networks to own…  do you really need a user to push a button to tell you where they are in order to get that location information? Per a recent L.A. Times article, the latest Apple iPhone terms and conditions changed to include a section related to LBS where they declare that “Apple and our partners and licensees may collect, use, and share precise location data, including the real-time geographic location of your Apple computer or device” and the article also makes note of Google’s similar geo data collection policies for Android Phones. 

For at least a few years now companies like Sense Networks and Placecast  have been working with large volumes of aggregated location data, collected from a variety of places, in an attempt to unlock the value contained within a long history of geospatial locates.  So there are obviously other ways to get at this raw data and make it valuable beyond the self reported (and self serving?) check in, which after all is just a small snapshot of activity of a few million users at best. But how valuable is a string of user locates as stand alone data?

In the current online world, tracking a search query or information from a web page to turn it around for re targeting purposes is relatively straightforward since everything already exists digitally. But when someone goes SOMEPLACE in the real world now, the digital documentation about that place is currently pretty weak…  so going to a position in space at some point needs to get digitally mapped back to the vast reservoir of digitized knowledge that we have about that space.

Folks like Localeze have started us down this path by making business listings more rich versus the dry name, address and phone numbers of the days of the yellowpages, but they’re coming at it from a perspective of web and local search.

Meanwhile an interesting new company called PlaceIQ is coming at it from the perspective of painting a better contextual picture of the places people visit. In the same way that ContextWeb tries to understand the context of the content on a webpage to serve a better ad, PlaceIQ is looking to better understand the context of a place to serve a more relevant mobile ad to folks at that location, not based on the content within a mobile site or app, but on the geographic space surrounding the customer at that time.  Taking it a step further PlaceIQ, similar to companies like Magnetic and aCerno,  will look to extend that knowledge of place to using information about historical presence at places to better target advertising via re targeting… like a mobile ad for a Derek Jeter jersey targeted to someone who attended a game in Yankee stadium a few days earlier.

Just knowing that a person is at a given latitude and longitude alone may turn out to be about as useful as knowing someone is on the web… and from an advertisers point of view, pretty low value. But if that latitude and longitude can be resolved to a place, and a ton of other information assigned to that place, then a new rich dataset for targeting and re targeting across the mobile and geoweb will evolve with location and presence at its center.

Despite claims of hundreds of thousands if not millions of users across these apps, the vast majority of my friends aren’t using it so it can get a little boring.

Recently I’ve found myself playing more and more with MyTown, last night as I waited for my wife outside Mermaid Inn, a neighborhood restaurant, I had time to check in on all three, but ended up spending 15 minutes collecting rent and upgrading a bunch of business I own on MyTown… frankly I am not sure how much longer I’ll be doing that either but it’s not a bad way to pass 15-20 minutes. And I did take advantage of the special offer that was running on Foursqaare for a free side dish at Mermaid Inn, the restaurant we were headed to anyway… a little surprise free side order of onion rings popping up in your day is never a bad thing. So maybe I’ll stick with it all a bit longer.

On a random technical side note, I did notice that the MyTown geo location feature is noticable more off than folks like Gowalla and Foursquare. Standing on Amsterdam avenue in front of three restaurants Mermaid Inn, B Cafe and Rancho Cafe I noticed that B Cafe was MIA while the other two were right there near the top of the list on MyTown with their little green indicator lights on. After a search I found the restaurant B Cafe, and they had the correct street address, yet still MyTown thought I was nowhere near the place… and it’s not the first time I’ve had to use search to find nearby places on MyTown… definitely some kinks to get worked out there.

I’ve found myself recently thinking alot about all these folks from places like Gowalla, Foursquare, myTown, Loopt.. not to mention the bigger guys like Patch, Google, and Yelp trying to sell in digital advertising to local retailers and some of the challenges they must face. 

On Greg Sterling’s blog Screenwerk he posted a video from a company called WebVisible that speaks directly to this challenge, where many local retailers talk about the various ways they think customer find their businesses (walk-ins and yellow pages types stuff) and then consumers talk about how they really find businesses in this modern age (“the Google” and the Internet). Check out the video below… the bit near the end about one flower shop setting up a page on HappyFace (er Facebook?) is kinda telling. Good luck to the FourSquare sales guys calling on that lady!

It seems that the folk that have found this most useful so far are governments and energy companies.  According to the Metacarta site, there are millions of government documents of which over 70% contain significant geographic references.

I can see how this would be quite valuable for the folks in homeland security who could now potentially  connect the dots where a CIA agent writes up a report about a suspicious character on Mott street and another agency transcribes a phone conversation between bad guys referencing a business at a nearby location adjacent to Mott street.  It’s information that may not have been connected in the past if it weren’t for the previously unseen geographic/proximity link.  

The other category where Metacarta operates is in digital publishing, particularly around news content and this is the area I was interested in poking at a bit more. AOL’s Patch product has been getting a ton of press, Yahoo just plunked down $100 million for Associated Content, and other start ups like Outside.in are all operating in some shape or form in the area of hyper local content creation or aggregation.

Creating or aggregating all of that content for a hyper local audience of what may be just a few thousand people seems like an awful lot of work. So let’s look at some numbers…

Having worked with Associated Content before I suspect that its reasonable to get a 500 word article produced for $30, and let’s say that the article is about something happening in the town of Mamaronek, New York  a small bedroom community of NYC  with a population of 18,752 (at least as of the last census). So using some oversimplified math… in order to break even on the expense of creating the article assuming un limited advertiser demand for Mamaronek eyeballs at $15 CPM, the article would need to attract 2,000 views, assuming 1 ad per page, or 1,000 views assuming 2 ads per page… or approximately 10.6 or 5.3% of the town population respectively.

Now seeing that Quantcast estimates the domestic use of Google.com at around 157 million monthly unique out of a total population of 307 million people or about 51%, it seems a tad aggressive to assume that one publisher could get 10% of a local population engaged in their content… although just for the heck of it I looked up the circulation of my hometown newspaper the Gainesville Sun which has a circulation of around 45k against a population of around 125k or over 1/3 of the population, so maybe its not so crazy!

You can kinda see where the Excel  commandos may be getting excited, saying something like “if we can establish Patch as the next generation local newspaper and get 1/3 of the population reading the site and get articles produced at $30 a pop, we’ll be rich!”

But here are a few potential glitches to that plan. For one, as the barrier to creating local content is lowered, unlike in the traditional newspaper world, the number of competitive neighborhood sites will  be more significant, so getting 1/3 of the population to read your website like the Gainesville Sun has accomplished in Gainesville, will be much more challenging… outside.in will be right there challenging Patch  for the local eyeballs with a network of other local bloggers. 

Second and probably more importantly, is the advertising demand and pricing question.  Ad networks are awash in inventory, and geo targeting is just a simple check box away, so the eyeballs are already available at CPMs in the low single digits.  While small local businesses will be sold directly and the vast majority have never heard of advertising.com, I am not sure you can expect them to pay large “local content adjacency” premiums over the long run.

A quick perusal through the Patch job listings shows a lot of emphasis on the business directories business, so it seems that AOL may be trying to hedge their bets even further against the premium display CPM ad business  and trying to cast a wide net capturing ad dollars previously headed for local premium display, classified, and yellow pages.  Then there is the whole behavioral/re targeting opportunities when folks from the local sites show up later elsewhere on an aol property and upgrade the value of that inventory as well.

You can start to see where this local content begins to look like a valuable opportunity once you can get the production costs low enough and local ad sales folks in place… it all makes sense on a spreadsheet at least.

But what resonated with me most, was chuckling at the business listings that were purported to be surrounding me as I drove down fifth avenue on to Central Park South… home to some of the toniest hotels and shops like The Plaza and Pierre Hotels. According to my Nuvi right next door to those hotels was supposed to be a place called AAA Als Towing and then a few blocks later an auto repair place… yeah right, how many auto repair and tow places do you know paying more than a few grand per square foot for such prime real estate…  it was apparently the early days of POI spam, or at least a really bad dataset.

You’ve probably seen articles about the locksmith map spam problem on places like Google Maps, but its much more than that… the industry as a whole suffers from just really poor information related to documenting places… for too many years the map and navigation device makers were seemingly so focused on getting folks from point a to point b and documenting the roadways they’d use to get there, that they forgot about all the places in between point a and point b where you may want to stop and gas up or grab a bite to eat. 

But nowadays things have changed and folks are giving away the capability to navigate from point a to point b and looking for ways to make money from getting you interested in making stops along the way.

As I mentioned a few months back a company called Placecast is trying to help bring together the disjointed state of POI data, to help move the industry forward, by introducing a product called the MatchAPI.  What the MatchAPI does is allow developers to send in a reference to a geographic location through the API and receive in return any other references in the system that are a close match. The reason something like this is necessary is because there seems to be so many different proprietary datasets around without any centralized and unified source of reference to help connect them. Let’s start with a simple two dataset example where a company has a list of movie theatres from one source like Navteq (where they get all their POI data) and then wants to go to a different vendor like Fandango to get movie times and reviews for the movies happening in those theatres.  So if there are a few thousand movie theatres in the U.S. you can see how the task of matching up the Navteq list of theatres with the Fandango list of theatres could be pretty labor intensive… that’s one area where the Match API can come into ease some of this pain.  Now if you wanted to do something similar with say the 1 million or so restaurants from that same Navteq POI set in the U.S. and linking in all the Yelp reviews tied to those restaurants… well you can only imagine the amount of pain that the MatchAPI is alleviating.

But where things can really start to take off is when various systems that rely on “place” can more simply and seamlessly integrate and “speak” with one another on the fly, creating a much larger opportunity for all the players involved when the network effect begins to kick in. 

Right now there are hundreds if not thousands of location based services all largely operating independently with their own systems, definition of places and customers.   When the ability to share information between these services becomes more seamless, not only will the value to users increase dramatically, but folks like advertisers can begin to view this industry as a cohesive ecosystem, and one that has some meaningful  scale and reach which will expedite this becoming a viable new medium for them.

Right now even the most successful Location Based Services in the U.S. just have a few million active monthly users, which won’t put them on the radar screens of many big national advertisers.  In a way it’s all similar to the way that DoubleClick first helped cobble together a bunch of small websites into an online ad network and create a highly simplified way for advertisers to buy a large volume of eyeballs over 15 years ago. Right now if McDonalds wanted to throw in a two week promotion of their Shrek Glasses in the business listings for their 31,000 stores across all the map platforms, navigation devices, and LBS iPhone apps… well they probably couldn’t do it without a small army of buyers, designers and integrators.  This is the pain that the MatchAPI could eventually help go away.

Just today Placecast came out with news that after 60 days of having launched the MatchAPI platform, they’re finding that error rates in the data they’re seeing runs anywhere from 8% to 40% depending on whether it ‘professional’ data or ‘user generated’… finally a somewhat quantitative representation of all those towing and auto repair shops I noticed along Central Park South.   

For the nascent LBS industry there is no way that this becomes a big and viable opportunity for marketers if 8% to 40% of the time you either send someone to the wrong place, send them to a place that doesn’t exist or is closed, or give them the wrong phone number,  etc. … so fortunately there are folks hacking away at trying to help solve some of these problems, so we can move along to some of the bigger and more interesting innovations that are possible.

There are two basic underlying practices as part of the guidelines:

1. Users must receive notice about how location information will be used, protected and shared… although the form of notice is not dictated

2. LBS providers must show that users gave consent to divulge location before initiating the location based service and users must have the right to revoke consent at anytime… although the way in which consent is recorded or retracted is not dictated

Some other details of interest, and what one may potentially read between the lines: • CTIA encourages the industry to develop “new technology to empower users to exercise control”… in other words lets not just bury some words in the terms & conditions somewhere to cover our butts, someone please come up with a cool system to give consumer controls over their location data that they’ll actually use… maybe a fireeagle-ish thing?

• A wireless carrier is a LBS provider when it directly provides users with a service, not when it provides location information to an application developer who then turns around and offers a location based service. In other words the guy directly offering the service is the one bearing the LBS provider responsibilities, not the originator of the location data, so lets provide protection (and remove some risk) to the carrier providing location data, thus encouraging it.

• When location information is not linked to a specific device or person, but only used in the form of aggregated or anonymous data, notice must still be given, but the consent requirement seems to go away… ie you still need to tell people what you’re up to, but since it’s not about any individual, getting individuals consent would be over-kill.

• LBS providers must inform users of how long location data will be retained and should only retain location data as long as business needs require and should afterwards be destroyed or converted to aggregate or anonymous data. This seems to be a tricky one, in many cases it would seem to be in the business best interest to retain as much information as it can for as long as it can… so I don’t see many folks being anxious to destroy this information if there may be a valuable business purpose down the road that they haven’t come up with yet.

 • “Consent may be implicit such as when users request a service that obviously relies on the location of their device”… ie all that stuff about consent doesn’t really apply if you’re running an app called “Whats nearby me now?” where it’s obvious that it needs to know your location to perform.

The whole things reminds me quite a bit of what the web community has done with regards to browser cookies and PII online, so there seems to be a lot of precedence here. As you might expect there is nothing revolutionary in the guidelines and it is more or less common sense and doesn’t disrupt much of the way the location based services I’ve seen operate already.

It may have been my imagination, but while reading it I felt as though the CTIA was really hoping that someone would develop an innovation that would allow users to actively manage their location sharing while realizing that in reality it’s likely to go the route of check boxes on multi page terms and conditions documents that no one reads.

So anyone that’s been around the online advertising world will be familiar with the famous Google Page Rank algorithm. While maybe no one other than Larry and Sergey truly knows how it works, there are literally small armies of SEM and SEO experts that wake up in the middle of the night in a sweat wondering if they left out an important keyword, or whether they need to pay for links to get a boost for their clients.  It’s a fascinating micro economy that has developed almost exclusively around servicing customers and their interaction with Internet search providers, particularly Google and its $20B in annual revenue.

I haven’t bought search in well over a decade, before Google existed, and am by no means an expert in search, let alone local search, but if you’re looking for more information I’d suggest starting out by reading SEOmoz or Greg Sterlings Skreenwerk blog or reaching out to a local search SEO specialist like David Mihm or Mike Blumenthal who are frequent speakers on those circuits and regularly share some invaluable experiences on their blogs at Mihmorandum and Blumentahals.

But to greatly over simplify, fundamentally there are two main components in play for Google on the web, and how well they translate into a true mobile location aware search is fuzzy at best.  So for the current Google web search here are two key factors being looked at:

Continue Reading on LBX Journal

Well after giving it a bit more thought, I’ve come to the conclusion that the industry needs a more broad “presence at place of sale” (PAPOS?) metric that could capture all the events where a person actually walks into a brick and mortar retailer and that action is recorded by any available means.

The PAPOS could then be looped back into the marketing ecosystem serving as the click or conversion rate for all advertising, both on and offline, targeted at driving brick and mortar foot traffic. When consumers actively check in somewhere that is fantastic, but I think it’s a bit unrealistic to expect this to become a ubiquitous consumer behavior, and it’s certainly not the only way to know when someone is in a retail store. Besides the 100% active way of checking in, there is of course the 100% passive way where your phone location is just recorded in carrier location logs, and many, many things in between like credit card and loyalty card swipes at retail and of course the wi fi and GPS positioning directly in the phone which enables all those great mobile discovery and navigation apps.

I suspect that the active and explicit way of registering PAPOS through check ins will be pretty limited and much will be inferred or recorded through other methods, you’d figure some company will just find a way to throw up 14 million geo fences around all of the retailers of America and fire back geo pixel trackers each time a mobile device enters a place of interest!

A more systematic way of recording PAPOS would go a long way in opening up the world of mobile and local digital advertising… which currently suffers from a serious disconnect when measuring ad effectiveness when they’re looking to drive offline foot traffic… tracking click to call is ok, but how often do you find yourself calling your local McDonalds? “Uh Excuse me, do you have chicken nuggets in stock today, I just want to be sure before I drove on over?”  HA!

Per John Hanke, VP of Google Earth, Maps, and Local from a recent TechCrunch article : ”PlaceRank is like PageRank for places, it tries to figure out how prominent a place is based on factors such as references on the Web, reviews, photos, how many people know about it, how long its been around.”

By the way I think it’s notable that the thing being “figured out” here is “prominence”.

Now I understand that you’ve got to start somewhere, but I would argue that the tactics used in web search engines don’t really apply to spatial search much and it should be treated as a completely separate animal.  Web pages are about text and the authors of that text linking to (and as a result voting on) other web pages, in order to determine a pages’ ‘prominence’. 

I think there are a few key differences when looking at location and spatial oriented search:

• The “prominence” of a search result is relative to things like distance and the convenience of alternatives in local/spatial search, versus something more absolute in web search where you’re simply clicking on a link to ‘get there’
• Determining “prominence” is very important when parsing through 1 trillion pages of “always available” information, but in the more dynamic yet much more limited options of local search something as simple as solving for “highest prominence” may not be the right answer
• The true “linking” happening to a physical place is not happening on a website, but through foot traffic and phone calls… and the traffic links between places is not captured on a webpage at all, but on a handset or a carriers’ back end logs
•  The stuff being searched for could and should exist in a variety of mediums, not just html on webservers… find a person from their mobile device, find an item from an inventory system, find a bus from a location sensor.

I suspect  that there will be some incremental improvements over time with matching online information to offline stuff, but I also suspect that we’d be better off by blowing up the existing search model and starting over from the ground up with a model designed purely around location specific spatial search, that merely taps into the vast reservoir of online content only when necessary… rather than serving as the foundation.

Stat: There is a huge discrepancy between time spent on local content and advertising spent for local online advertising. If the ratio comes in line with other mediums, it’s a $30 billion market
So what? It’s potentially a $30 billion market… maybe something to pay attention to.

Stat: Last year 1.4 million companies bought search advertising versus only 44,000 that purchased online display.
So what? Well the opportunity seems to be finding a way to make it easier for those 1.4 million local companies to buy display online.
 
Stat: 80% of consumer dollars are spent within a 50-mile radius near their home, and 96% of consumer purchases  happen in retail store environments; e-commerce is still quite small in the grand scheme of things.
So what? The opportunity to use new technologoy to help connect consumers with stores and items in the stores around them is significant.

Stat: Quick service restaurants spend $3.04B per year and only 2% goes to Internet, and according to MerchantCircle research, 60% of SMBs never heard of Facebook…..
So what? There is still a long way to go with many retail oriented segments, they’re still not online, let alone on mobile, social networking or location aware advertising.

Stat: On Yelp, 3 million of 16 million local businesses have reviews.
So What? Well if you’re one running one of those 3 million businesses, you should keep an eye on what is being said. Online reputations management systems are popping up to help make this a more quick and painless process.

Stat: currently local digital advertising is getting placed via:
Self service: 49% (presumably Google AdWords primarily)
Yellow page sales team: 24%
TV and Radio sales: 9% (SMBs preferred method)
Newpaper sales team: 8%
Ad Agencies: 8%

So what? Well as you may have expected, folks that have always sold local advertising to the local SMB market, are now also selling someone else’s digital advertising to those same customers.  But many local business are buying digtital media themselves, likely out of necessity, even though they’d prefer using their TV and radio sales contacts. Looks like there may be an opportunity to provide high quality sales service selling  digital advertising to the local SMB maket.

Am I suprised that the goverment actively looks up peoples location data, no.  Am I surprised that they did it 8 million times, well yes! Maybe they should lend the login to the Sprint field sales team so they can track down and give a personal face to face pitch to the ~1 million customers that Sprint is losing each quarter!