The Future of Location-Based Advertising Looks Like Twitter

I promised a look at the challenges we faced in the evolving local advertising space since 2005. So here it is, I hope this helps avoid similar pitfalls and help accelerate faster evolution from the advertising platforms toward a lower friction, more mature ecosystem of direct exchange.

Challenges and Complexities: An Immature Ecosystem

The traditional network model doesn't have enough local efficiency built into the system to scale geographically. Meaning that as networks sell to advertisers and attempt to match them to a publisher network, very few of those ads carry any location at all. Those that offer "local" targeting have meant city and in some cases zipcode which is fine for tests but doesn't warrant a premium for location.

Current reality is different from current perception that; 1. advertising is currently location-targeted, (even location-aware),  2. there is advertiser demand for lat/long level targeting and 3. there are flexible mechanisms in place to sprinkle these ads around the web or to mobile devices.

When Platial launched, the location-based advertising ecosystem was immature. We had the technical capacity to target by lat/long AND context but there were few ways for advertisers to buy these ad units. Additionally, it was difficult to define the value, so what will be a premium in 2010/2011 was experimental (read-free or cheap) previously.  How much can you charge for location-targeting when there are no case studies for driving traffic to point of sale? One test of location+context was a pay per call campaign using several verticals like legal and real estate. We could target legal ads to maps, places and specific sites within our own network related to "Lawyers" in "Pittsburg". Very smooth case study with eCPMs of around $7. However, this number could not be scaled because the amount of inventory needed to match relevance against our entire network was huge. In fact, its feasible that even the largest aggregators today, would not have inventory to match all of our content. Even at our peak w/impression levels which were astonishingly high, the slicing and dicing of context and location could not work.

Some of our specific challenges:

1. Ad networks who have a large inventory of location-based advertising

It is still difficult to find. Most mobile aggregators and networks are just now launching or will launch their location based advertising initiatives. It is easy to count all ad inventory but drilling down to location and any associated metadata is very difficult, particularly within free form text of user-generated content.

Still a problem

2. Map languages and ad formats

Because our users were everywhere on the web, our maps were in Flash and javascript and in customizable sizes which is maximum value for the site owner/map owner but many of those formats are not monetizable. Most ad networks still do not have a Flash API.

Still a problem

3. Distributed model- widgets

Early local advertising networks could match an advertiser to a pre-screened/pre-approved site, however, we were distributed across the web on about 500,000 sites per month and millions of blogs, profiles and websites. There were no APIs for controlling a subnetwork within their network.

Better

Mobile advertising and the emergence of powerful platforms like Apple and Facebook has made this partially irrelevant.  Lots of apps are able to participate in systems that can and have scaled. With the power and brand of these platforms, it has created high advertiser demand, and a need for inventory on mobile devices and within social networks which require distributed models.

4. 3rd party vs. creation of your own sales force/system

We decided early on that we would not build up an ad sales force. We created a self serve ad platform but decided that the development and sale of this product would compete too heavily for resources creating our core map offering. The right answer is different for every company, but I still see lots of people throwing an ad network into their product mix before their product or platform is even solid or successful. It is not an easy task to build. Technically it is simple but there is a lot of legwork in the creation of a smooth running ecosystem at a scale that makes money for its participants.

Better

Message to publishers and apps- if you're looking for short term revenues, work with an established network or deep-pocketed, cynical upstart. Decide if you want to be the one generating the case studies or the money because the gap between the two can take years. We fell into a common start-up trap- we like inventing the future more than analyzing the past. Both are absolutely critical. We were also overly optimistic about our ability to make the future come faster :)

5. How to value of location-aware advertising

Between a maturing ecosystem, advertiser demand, mobile platforms and location-obsessiveness we have some good case studies for driving value and pricing. This is a question which is constantly changing and many of the networks are optimized around this today.

Better

6. Platform support

Its amazing the range of perspectives in this world. Mypace disallowed us to link out from any widget, meaning as a developer, we could not monetize. Facebook on the other hand treats developers like a platform more than an ad network and believe its in their best interest to help developers monetize.

Google has built its business on developer support and shared revenue, although, honestly the pay out on location and maps its still immature.

Better

So, in the old days, publishers were "inventory" for advertisers, I would avoid this mentality today where platforms are supporting shared revenue. Apple has the best history of helping its developers make money. I believe they will be able to extend this to advertising.

Platial Levers and Map eCPMs

For Platial, the levers are the same in an immature and a mature ecosystem:

1. # Maps served
a. Revenue-generating: Large format, javascript
b. Not yet revenue generating: Sidebar js (too small), Flash

2. Avg eCPM (for comparison sake we backed CPC, cost per call etc into eCPM)
a. Text and display ads on a map/map page
b. Sponsored pins
c. Text ads on places/place pages

I've averaged eCPMs over a period of time across various networks to give an overview and purposefully not drill down on a single network, ad type etc. Its a simple model for how many maps served and corresponding revenue that would yield at our avg eCPMs.

Its easy to see how to improve advertising revenue. I assure you, each has its own challenges and you are seeing only a glimpse in this overview:

1. Serve more maps
2. Increase engagement
3. Increase place pages
4. Drive more traffic from widget to destination
5. Sell sponsorships; innovate on our own sales process
   
6. Mobile emphasis
7. Contexts that would be most relevant, real estate- kind of place/kind of map. Running maps, for example
   
8. Work only with third parties who get "platform" vs. "inventory"
   

A Mature Ecosystem is Coming: Single Community Location Aware Exchange

Networks sell to "local" advertisers and create networks of publishers. At scale there is enough money moving through the system for publishers and networks to make money. We experimented a lot but there was not enough value in the fractions of context targeting and location targeting.

Maps and distributed bits can be better monetized for better efficiency when we can have self serve platforms which are as much tools as networks.  This enabled aggregators and map monetization potential.



Maps, apps and distributed bits can be better monetized for better efficiency with self serve platforms which are as much tools as networks.  But there is a missing piece. This is the piece Platial was so excited about from the beginning when we called our Platform "Support Local Business". That piece is A SINGLE UNIFIED COMMUNITY- not advertisers, publishers, developers, platforms. A single community where our roles change with our needs; the seller, the buyer, the marketer, the consumer, the business, the customer but always, a member of the community. Local advertising, allows us to cater to this and makes it a unique opportunity. To facilitate this you need a powerful exchange NOT a powerful ad network.

Platial decided to cut out much of the inefficiency by creating "local self-serve networks". The opportunity is even greater with location-aware units as opposed to  local targeting. We can track our exchange, our transaction, our value in better ways. An exchange can facilitate transactions or messages between members of a community. For example, I have bamboo and you can buy it--more like classified models, small businesses, independents. For this reason, I think Twitter may hold the key to the future of location-aware advertising. I have a special of the day and you are hungry, I have your friends here and you want to see them. Additionally, the economics work- it is cheap for me to buy a posting for $5.00 but expensive at a CPM if you rely on advertisers sending bulk messages. Twitter already facilitate supply and demand in a new way seamlessly between online and offline.  They have created a dynamic potential marketplace, with real time need fulfillment-- Long tail of supply and demand as ad network. The big chains can only sustain so much of a network. So, here's to reducing friction, local economies, dynamic need fulfillment and greater transparency between buys and sellers. Good luck to everyone forging ahead on this! Good times and happy mapping.