The Fwix technology and data has now been ported over to Radius, the leading sales intelligence provider on small businesses. Radius is the first SMB sales intelligence platform, and it is proving to be a great business model. We are continuing to build upon our valuable technology and data, but have changed our name and visual identity to better suit our new market.
Notice: All Fwix services will discontinue on May 31st, 2012
As our final blog post for 2011, a big thanks to each of you who read our blog.
We're very excited for what's to come in 2012 :)
The history of web organization is important to understanding how information is found on the web. As new organization methods develop, new site optimization techniques (e.g. SEO, SMO) evolve as well.
The newest site optimization method is Location Search Optimization. LSO is directly related to the growth of smart-phones and location-based searches. Websites are geotagging their content to become LSO compliant -- so people can find information based on where they are, not just what they want.
This info-graphic below highlights key points from 20 years of web organization and site optimization. We hope you enjoy this info-graphic!
With mobile phones in every pocket, the Web will be entirely mobile within the next 10 years. Just as Search Engine Optimization (SEO) served the needs of computers in the early 2000s, Location Search Optimization (LSO) serves the search needs of today's smart mobile devices. LSO is directly related to the growth of smart phones and location-based searches.
LSO is the newest site optimization method, following in the evolution of SEO and SMO (Social Media Optimization). Site optimization methods improve the visilibility of a website in search engines and finding information online. With SEO, publishers optimized around keywords, and with SMO, publishers optimized sites by adding Facebook and Twitter Share buttons. LSO is optimizing around location so your content will be more easily discovered on location-aware devices.
Why LSO Matters?
More people are accessing the web via mobile devices than computers. By 2014, mobile internet usage is expected to surpass PC internet usage. People turn to their mobile phones to learn what is happening around them and nearby. Location-based searches are now a key way people find information online, and specifically, Google reports that 40% of mobile searches are about a location.
Did you know there will be a concert at the park near your house? What are people saying about that restaurant across the street? What news articles and webpages mention that restaurant? These questions all boil down to location.
More people now are discovering information through location and mobile devices. Location is most relevant in delivering what information users need on their location-enabled devices.
How to Become LSO Compliant?
It takes less than 5 minutes to become LSO compliant. Here are the steps:
That's it.
What types of content can be optimized (e.g. LSO compliant)?
Learn More about LSO
Fwix CEO Darian Shirazi recently gave the keynote talk about LSO at the Street Fight Summit in New York City, where he provides an informative overview. The slides he used in his presentation are not apparent in the video, but we embedded the slides below as well.
View the slides from his keynote below.
Finally, just remember
The web will be entirely mobile in the next 10 years, and LSO is what SEO was in 1999. Don't get left behind!
While the promise and reality of online local advertising is bright, the success of this market is dependent on the performance of local ads. As we will see in this analysis, the vast majority of geotargeting capabilities are still only targeted to a DMA (Designated Market Area, e.g. the San Francisco Bay Area). Is that local enough? What if our geotagging technology can help improve targeted advertising to a neighborhood level?
Let’s quickly recap some of the market data and industry trends. In 2011, Borrell Associates reports the US online local advertising market to be $23 billion, and the majority of this ad spend is divided up between search and display advertising on the Web.
Industry changes are reflecting this opportunity as well. Google, Amazon, AOL, and AT&T Interactive have entered the daily deals space in addition to bulking up their general local advertising sales efforts. There has also been the meteoric rise of daily deal companies, such as Groupon and Living Social.
As companies increase their local ad efforts, it will be important that local ads become fully optimized. In order to optimize local ads, geotargeting must be effective in displaying a relevant, contextual ad. Let’s review the current geotargeting methods and respective limitations:
|
Description
|
Limitation
|
||
|
IP Based | Ads are delivered using a user’s geographic location as determined by the IP address. | There are widely-known issues with the quality and precision of this method. IP-based can at best target to a DMA. |
|
Mobile/Location-Based | Ads are delivered using the geographic coordinates via the smartphone’s GPS capabilities. | IAC reports that the mobile inventory is by far mostly mobile Web (70%) rather than application (30%). Consequently, most available mobile ad inventory has only DMA-level targeting because mobile Web publishers generally have access to only derived-location data. In-app advertising, on the other hand, has been too expensive for advertisers to participate in (e.g. iAd). |
|
User Profile | Ads are delivered based upon the explicit user registration info submitted for a particular site or service. | The scale is limited to the service which required user registration, and there are other issues such as outdated information, or lacking location context. The targeting capability is city level, but typically targeted to a DMA. |
|
Behavioral | Ads are delivered based on the digital footprint (via cookies) of the user, and they are based upon the user’s browsing actions. | As these are derived locations, the precision of the geotargeting is inconsistent, especially if the browser cookies are often cleared. Behavioral ads typically has at best DMA-level targeting. |
|
Search | Ads are delivered based on the search query entered by users on search engines. | While these are very effective ads, it relies on the user entering the specified location within the search query, and if it the location is not included in the search query, it ultimately relies on IP-based targeting. |
|
Contextual | Ads are delivered based on the location context of the webpage, typically hyperlocal content that is relevant to the user. | The inconsistency with this form of geotargeting is limited to local news sites or weather sites. These sites cannot target beyond the city level, as they are inherently organized by zip code. |
The untapped opportunity in advertising is contextual targeting.
There is a significant opportunity in location-targeted advertising and moving past DMA and city level to one step further: neighborhoods and places. Through geotagging, we can better understand the location context of a webpage and identify the street intersection, neighborhood, and city via the referenced places and locations. In turn, we can retarget local ads more effectively.
This type of contextual targeting also allows advertisers to also get more granular categorical information. If a detected place in a piece of content is a hospital or a car dealership, categorical inference yields Health or Automotive categories, respectively. This type of contextually targeted advertising allows advertisements from national advertisers to be more focused and delivered to local areas. Combining categorical and location information allows advertisers like BMW to advertise for a specific dealership in Santa Rosa on an article about Mercedes Benz’s auto show at the Infineon Raceway.
The impact of such an improvement in targeting can be astounding for media and advertising companies.
In the case of national advertisers, targeting advertisements for national companies that have local presence continues to be a multi-billion dollar opportunity. Bank of America, The Home Depot, Wal-mart, and the like are only a few examples of advertisers that need local intent and local targeting to bring online customers to stores. The ability for The Home Depot to advertise for a lawn-mower deal at the West Los Angeles store on an article about home improvement in West L.A. is impossible without contextually-targeted geotagging.
In the case of daily deals, Groupon spent $180 million in online marketing in the first quarter alone and acquired approximately 33 million new subscribers in that time. That is about $5.40 per new subscriber, and Groupon is on path to spend $720 million annually in local advertising. If Groupon and other services could understand the location context, targeted down to a neighborhood, they can buy advertisements on pages that are relevant to the deals they have inventory for, thereby increasing conversion and reducing marketing expenditures.
So, how does it work? Well, it’s pretty simple. Advertisers submit a URL to us for the page they want to advertise on and we’ll return a list of places and local categories associated with the page. We can also return latitudes and longitudes associated with a page. This meta data about the page can then be used to determine which ad to show on that particular page.
Advertising agencies want this data and advertisers want to find ways to reach real customers in local areas more effectively. We’re excited about this massive opportunity, and we’re excited that geotagging can bring the right ads to the right content and to the right people.
Continuing the series of blog posts about the benefits of geotagging your site, today we'll focus on the benefit of providing advanced mapping features.
By geotagging your content, you can provide new ways to display your content through an interactive map. Specifically, you could imagine your site's users clicking on a pin on a map and they'll be presented with a pop-up balloon containing a list of the relevant articles for that place.
In an era where devices like the iPad with beautiful, interactive touch-screens are revolutionizing how media is consumed. New apps are being built with innovative user interfaces (e.g. Flipboard, Pulse). As a result, there is a demand for new and interactive ways to discover online content. Below is one example of news content overlayed on a map.
Publishers are catching on to this trend of providing advanced mapping features. The New York Times recently just launched their Longitude product. With Longitude, users can interact through a map to find news content about places and locations they care about. New York Times Readers, for example, can find news articles about their favorite restaurant or their local neighborhood. Below is a screenshot of Longitude.
Geographical interfaces are fast becoming a part of your future news reading experience!
With our Geotagger Button annoucement last week, you likely noticed our new Fwix.com. The new site deserves a separate post. While the whole site was redesigned, there are 2 significant changes to the site:
1.The Geotagger Button is now available for everyone and free to use.
2. Brand New Place Pages for over 25,000,000 places.
You can check out the brand new place pages by simply entering your favorite place or location in the search box. In the example below, I'll type in "SoMa" -- which is a neighborhood here in San Francisco.
Once you select the location, you can then view all the content (news, events, reviews, photos, tweets, social media) we have geotagged about that place in a news feed. The news feed is awesome -- either view the 'featured' news feed or the 'most recent' feed of geotagged content. We're excited about these place pages because you can see the breadth and depth of our geotagger in action. It demonstrates the capabilities of this location-based search index where users can find relevant information about a place. You can view this referenced page here.
Check it out yourself and see what's happening at the places you care about. Better yet, tap into all this great data by using our API.
We're excited to announce the Geotagger Button is now available for everyone and free to use. In June, we launched the Geotagger Button in a public beta program with NBC and other media publishers.
Any publisher, blogger, or developer can now geotag a piece of content without writing a single line of code. Geotagger lets you deliver a new layer of location-based relevancy to your web property, for free. After a publisher pastes a snippet of Javascript code into a page, we'll automatically geotag that page and publishers can then access the geotags, e.g. the location data of the referenced places through our API. Below is a screenshot of the Geotagger Button on a sample webpage.
What are the Benefits of Geotagger?
In a series of future blog posts, I will dive into the details of each benefit listed above so stay tuned for them. Also, I will discuss our new site in more detail in a separate blog post! Today, I'll focus on the first benefit.
Reach More Readers on Mobile
The keyword here is mobile. With smart phones outselling desktop computers, we are in an era of real-time connectivity, in the palm of our hands. More people are now interacting with the Internet and search engines through mobile devices than desktop computers.
People are turning to their mobile phone to learn what is happening around them. Did you know there will be a concert at the park nearby? What are people saying about that restaurant across the street? What news articles and webpages mention that restaurant? These questions all boil down to location, and location is equally if not more important than the search query itself. People are turning to their mobile phones to answer these questions. Google reports that over 40% of mobile search queries are about a place, and Bing reports 50% of mobile search queries are about a place.
As a result, location is most relevant in delivering what information users need on their location-enabled devices. But is your content indexed by location? Is your website ready for search queries about a location?
The Geotagger Button is the solution to making your content location-relevant. When your content is geotagged, your content will automatically be added to our search index that allows people, developers, and partners to access information based on where they are. Our search index is open and free to use by developers and partners, and I'm proud to announce that we are serving over 10,000,000 API calls each day to our search index. Also incorporating the geotags (e.g. location data) on your website will boost your SEO and improve your rankings in mobile search queries about a location, which I will speak about in more detail in another post.
Fwix has geotagged over 100,000,000 million webpages so don't get left behind.
One of the most important steps in our modern, mobile, and location-based app world is building and maintaining a places database: a collection of all the physical locations, businesses, points of interest, and structures in the US and worldwide. All location-based services, such as a check-in application (e.g. Foursquare), restaurant reviews app (e.g. Yelp), local discovery app (e.g. Loopt), social networks (e.g. Facebook), etc all utilize such a database because places are core to the product functionality.
At Fwix, we use a places database, unsurprisingly, for our geotagging technology. As mentioned in a previous blog post, we partnered with Factual as our baseline places database provider. Factual continues to be a great partner company, and their places data is open and free to use through our API.
One might think, however, that because we have partnered with another company to provide our baseline places data, our work is complete in building and maintaining a places database. Is that the case? Certainly not. We have a devoted, full-time team working on maintaining a high-quality and constantly evolving places database. And this blog post will discuss the key challenges in doing so because several developers and companies have reached out to us seeking insight on this topic. It’s a lengthy explanation, but I hope it provides insight on the challenges our great engineers have been working on.
1. Data Quality
While we only distribute our Factual places data through our public API, we actually consolidate from many (10+) data sources in forming our internal, canonical places database. A canonical place simply means reconstructing an authoritative place listing from the 10+ sources. For example, each source has data for the restaurant Chez Panisse in Berkeley, CA. In forming the canonical place listing, we might grab the phone number from one source and the street address for Chez Panisse from another source. (Note: we only distribute the Factual places data through our API. We do not distribute the data we collect from the other sources.)
Consolidating from multiple sources improves our breadth, depth, and quality of data, but building a canonical place has its challenges. How do you determine data quality? How do you know which data source is better than the other? One source might consistently have an accurate phone number for a business listing but lacks a street address. One might infer that a source is poor in quality if a source is missing a big percentage of attributes (e.g. phone number, address). As a result, determining source quality is difficult, and building a canonical place becomes a dynamic value that is a mathematical function of the quality of sources.
In another instance below, two maps are shown of two different neighborhoods within San Francisco: SOMA and South Beach. While these are 2 distinct neighborhoods within San Francisco, some sources state that SOMA includes South Beach and some sources show them as 2 separate neighborhoods.
2. Data Constantly Changing
Places data change all the time, and stale data must be updated. New businesses spring up, businesses shut down, businesses change their phone number, businesses move to a new location, businesses add a storefront URL, and the list goes on. But it’s not just businesses either. Zip codes continually change, neighborhood boundaries change, and more. The data is never static and must be continually refreshed to maintain data quality. On average, we are modifying and updating 2 million existing place entries per month.
3. Scale
Processing data for millions, and millions, of entries in a reasonable time is not a simple task. We have over 30 million place entries in our database, and this number is growing. No matter how much hardware you throw at this problem, it still requires a complex infrastructure for distributed processing. We use proprietary software tools for distributed processing to optimize performance. The chart below gives a sense of this constant fluctuation:
4. Removing Duplicates
Programatically identifying duplicate place records referring to the same place is critical in maintaining a clean dataset. Does the name of this place already exist in our database? If there are multiple occurrences, are they actual duplicates, or perhaps part of a chain (e.g. Starbucks Coffee)?
For example, there are multiple “Osha Thai” restaurants here in San Francisco. One source may refer to this place as “Osha Thai Restaurant”, while another one simply refers to it as “Osha Thai”. Are they referring to the same place? Another example is the address: one source may state the street address to be “149 2nd Street” while another source may reference the street as “149 Second Street”. Obviously, there are many similar examples.
Resolving these duplicates, a.ka. “de-duping”, requires fuzzy string matching, confining geospatial searches to a geographic region, and calculating a score based on a function of parameters to measure and determine the most accurate match.
5. Lack of Data, Going Global
It’s difficult to find data for every corner within the US, let alone worldwide. Census.gov is a good source for US data, but even within the US, there are hundreds of rural towns that lack this basic information. This problem multiplies dramatically in dealing with a global places data.
6. Data Inferences
Another challenge arises from the lack of data in some geographic areas. Our technology will algorithmically make data inferences on what it thinks the value should be.
An example will help illustrate. The map below displays the neighborhoods of Daly City (outlined in blue), which is a city just south of San Francisco. Suppose there is a business located by the dot below and we had the street address but not the neighborhood information. In that case, we would have to infer the neighborhood. It is not an easy inference to do given the complex and coiled shape of Daly City. Inferring the value then turns into a series of geospatial mapping calculations relating to the centroid of neighborhoods in close proximity to the business.
Today we're releasing the Fwix Geotagger white paper to help customers learn more about our geotagging technology, which automatically identifies physical locations referred to in webpages. This paper provides a clear overview on the technology and examples of how customers are using the Geotagger Button.
Geotagging the web is challenging and not an easy problem to solve. No other company is geotagging the entire web or making a geotagging product, i.e. Geotagger Button, freely available for publishers and website owners. Our patented technology geotags, processes, and filteres hundreds of thousands of pieces of content per day, and it is a technology we've proudly built by our team.
We hope this paper helps you have a better insight into our technology and vision.
Download the free white paper (PDF).
