Our vision for map search has always been about discovery. During the development of NavDog we always had the saying “Tell me something I don’t know” posted on the wall in order to remind us of why people search for things in the first place.  Virtual Earth and Google maps give you this rich and dynamic backdrop of the world that you can control in real time but for some reason the big map search engines chose to utilize it in such a basic way.  In the current map search format the visitor types in a search, gets the item they are searching for and that’s about it. The rich experience stops at the map and does not translate to the content in a meaningful way and you get what you typed in to the search box and nothing more.  By contrast we embraced the nature of these mapping platforms and built our system to work in concert with them.

We chose to make the map work more like a map and less like a search engine, we included a map legend and personalized map items by giving them unique icons, then we decided to show you more than just one item and even more than one type of item at once. We wanted to bring in what you were looking for, plus related suggestions and then include additional items that related by a broader interest, the “things you didn’t know”.  In order to pull this off we needed to build a processor that could sift through the body of available items and bring back those that mattered most to the visitor at that moment.  We use a little contextual search with keywords but we rely more on associative search based on our engine.  The simplistic explanation: we track everything a visitor does on the site and assign a weight to each action and weigh the actions against the content they choose to interact with.

Google sample screen shot

Based on our logic certain types of content are promoted up to the visible layer or pushed down and out of the visitors view.  This creates a content hierarchy where items that are closer to your interest are available and items that are further from your interests fall down to deeper zoom levels.  This being said if you were to search for “Pizza” in the Mohave dessert you would most likely get our closest contextual match for that area, like “Pizza Rock” and the map may be littered with back fill content which does not match your interest in “Pizza” but can help you to stay engaged in your session.  There are many aspects of the engine that we have not turned on yet but we will be unveiling more features and systems over the next few months.

NavDog sample screenshot

When it comes to user interface we abandoned the Virtual Earth API in favor of a custom built interface where we could control every aspect of the experience. This design also allows us to use Google maps with a small amount of effort. From the placement of map icons to the popups to the control panels it’s all designed to work with our methodology.  Popups can contain custom HTML, Flash, or Media and they are basically blank slates that can be formatted any way we want them to be.  They can be detached from their icons and saved to the interface, and in the case of the live radio feeds it works great for listening to radio from one state while you explore another. .  We expose your search history in every detail in order to let your replay everything you have done on the website and we have created a way to exploit this feature to make tours of locations with the exact map data that you saw during your session.

What’s up next for NavDog?

We will be enhancing the interface and working on better browser compatibility for IE 6 as well as Safari.  We will be adding more partner data in the form of venues and ticketing, movie theaters, and more retail stores as well as much more content around the world.  We have a membership system that allows members to create mashups and tours and add their own public or private map items to the map but we have not turned it on yet. We are estimating that it will roll out in February-March 2009.  In the meantime we will be doing some promotion and working on the Business development front engaging possible partners.