Grokking a New Approach to Search Advertising Experience: Sea Dragon and More

Gary Flake (I posted on my meeting with him earlier this week here) has sent me files from his demo at SAS. I missed that demo, but you can get the picture, literally, by checking out these images. Here is an image of LiveSearch with the SeaDragon/SivlerLight mashup overlay….

Gary Flake (I posted on my meeting with him earlier this week here) has sent me files from his demo at SAS. I missed that demo, but you can get the picture, literally, by checking out these images.

Here is an image of LiveSearch with the SeaDragon/SivlerLight mashup overlay.


When you mouse over or click on a favicon-sized object at the beginning of the paid search ad on the right (in this image, you can just see one of them at the bottom right), a window opens up with tons more information, turning a “text ad” into an interactive display ad, voila! Yes, this is like Ajax, Flash, etc. But the feel is far more fluid, at least, it was in the demo…

Now, imagine this technology for display ads or online representations of traditional publications. Here’s an example of a digital newspaper:

Flake Ex 2

Imagine clicking or mousing over that BMW ad and getting tons more info instantly. All of sudden, real estate gets far more fungible. It could change the business model a bit.

PhotoSynth is still in early dev stages, but the possibilities are very interesting. Using the technology you can start to build virtual stores that integrate the way we are wired – visually – to the way the web is wired – datily. Er, data-ly. You get the picture. Well, no, here’s the picture:


In this example you see a very realistic rendering (it should be, it’s from a photo) of a high-end plumbing fixture store. Highlighted via the PhotoSynth interface is a faucet, and on the left is a ton of specs, data, etc. on that item. Imagine shopping in this interface online. It’s compelling….the glimmerings of a tangible link between the physical and the data worlds.

It’s all the matrix anyway. We just process it one way. Connecting the way we process data to the way machines process it is…well, fun to think about.

I bet if we asked nicely, Gary’d be happy to take your questions and input in the comments section….

10 thoughts on “Grokking a New Approach to Search Advertising Experience: Sea Dragon and More”

  1. I predict CTR for advertisers will drop with Flash ads on SERPs so unless they charge a premium for this type of ad it will reducing search ad revenue for MSFT and volume from advertisers…two already large issues with Live.

    Let’s think about what we know:

    Users click text ads in SERPS – Users ignore display ads
    Text ads have high conversion (and text links have higher CTR just about everywhere) – Display ads and images, Flash have low conversion & CTR

    The goal of the SERP should be to get people off of it as quickly as possible. This approach keeps users on the page.

    As far as the PhotoSynth example it loks like a bad category page to me.

    MSFT is doing a lot of really great things. The adCenter tools, Virtual Earth, Collections, to name a few… but I’m not as impressed as you are John with these ideas.

  2. Regarding the newspaper example:

    Why are we sticking to the physical paradigm. Their are digital benefits that formatting the same way as on paper make silly.

  3. Jonathan:

    Some of what you say is clearly true, but I don’t think that I agree with some of the assumptions and implications in your comments. First the assumption: you are looking at advertising and marketing as a “one size fits all” sort of thing. It is not. Different products and services have entirely different properties in how to best connect with a potential customer. Some product lines seek repeat business, which suggests that brand awareness is important. Other products follow a research, compare, convert pattern – which suggests both brand, targeting and direct marketing are all important. Other products are but one-time actions, which means that something more direct like paid search is best.

    The point: flash and other display technologies give better ROI to those who need to build brand awareness and to support a longer purchase cycle. Text ads often help at the tail end, but they do not fulfill the entire process. Finally, many of these advertising products are sold on an impression basis, so the CTR argument is really not even appropriate.

    As for Seadragon and advertising, of course this would be something entirely new: not an impression basis, nor a click basis, but something else. Also, the Photosynth page will not make much sense in static form. Take a look at to see what it is really about.

    — GWF

  4. Kevin:

    I hear you. We’re all about new visualization models. The point of that portion of the demo was to show that you can embed an unbounded amount of information in a finite amount of space. What you are not seeing in the static image is that we can zoom into the car ad at the bottom right, viewing details at many different scales. I’ve found that if you show a new application in a familiar context, it’s often easier to grok what’s different than if you show many new things at once.

    Look for new content visualization that breaks physical constraints at a later date.

    — GWF

  5. is this what web innovation has come to ? new ways to push landrovers. if this is the revolution, gak, it’s dreadful.

    microsoft’s search interface is such a blatant ripoff of google that they ought to be sued. they won one look&feel suit back in the day, and they think they can copypaste everyone else’s ideas into their own borg. i am absolutely underwhelmed.

    the whole POINT of cpc as google implemented by google was not really bothering the user ! it’s the user experience, dummy. this technology is just another damn thing for me to avoid with my mouse. screw that !

  6. I hope I’m allowed to meander a little here…

    I have so little expertise WRT graphic design and/or visualization issues, I guess I could basically say that “I know nussink”. ;D So that’s NOT what I want to talk/ask about here.

    My POV is that computers are good for computation — and in order to compute something, they need it in “machine readable form” (yes, Capt. James T. Kirk has a totally neat-o and courteous computer — I’ve never seen a “blue screen” on the Enterprise computer [but then again: that’s scifi, right? ;]).

    Well, at any rate, I see more and more how “search” is becoming more and more “bifurcated” into two segments:

    1. usability (human interface) issues

    2. info-retrieval (“classical” search/algorithm) issues

    For example: traditional search engines do not normally do “on the fly” and/or “dynamic” searching — i.e., they do not “perform” searches on other sites (the exception being the long-not-heard-of “meta-search” engines — of which there were actually a couple quite good ones back in the day). And besides: as more and more of the data is becoming presented in some visual form (e.g. a picture of a water faucet, or perhaps even a “video” showing how the faucet functions), there is less and less data that is “grokkable” by some kind of algorithm.

    I suspect that quite a few highly ranked websites have almost no “machine readable” content (i.e. content that a search engine could “grok”). If this site becomes very “popular” (has a bunch of links “pointing in”), then the relative *DEARTH* of machine readable content will allow very precise *targeting* of the small amount of information to be “configured”. At the *extreme* end of this scale would be a website where the only machine readable content is the URL of the website/webpage itself.

    And that’s exactly what I’m banking on! ;D (BTW: domain names are also *especially reliable* — since each domain name is listed in the registry’s database, along with details about the domain registrant [e.g. contact details])

    I guess if I had a question it would be: does anyone actually believe (IMHO “the myth”) that visual search engines are “just around the corner”? And what are the “best practices” for website development (and especially WRT so-called “SEO” issues) in a world with little to no “machine-readable” content?

    Would anyone wager a less-than-book-length answer to such questions?

  7. I think Sea Dragon is a somewhat interesting innovation in big ticket rich media advertisement. Its going to be used on major content providers, and it’ll provide a nice alternative to the current paradigm of views and clicks. Now you can get – and simultaneously measure – the full interactive website experience without the risk of losing the article you came to read in the first place.
    However, so I don’t come off as a total microsoft shill, I think there is a lot to be said for open formats and interoperability. What is to be done when your shiny new ads don’t work everywhere because Microsoft changes the pricing model for Silverlight? Not very likely, but the essential problem of infrastructure in proprietary hands remains.
    All in all, its some flashy tech, and it will be interesting to see if its used more effectively than the js and popups we despise so much.

  8. Mendez is probably right on the money, as far as charging premiums for these types of ads. Besides, there are lots of “innovative ways” for displaying ads… last time I was chatting with someone about broadcast ads, there was talk about having ad placements in shows that would be capable of being clickable (on pause on a DVR… think the wii remote acting as a mouse to select an ad item — that blunt coke bottle, for example, on the American Idol panel’s table — during playback. point/click the remote to coke bottle = visit You could construct the most SEO/SEM/Flexed-out ad ever, but really, at the end of the day, consumers will visit a destination site because of its original content, not its adverts. People won’t read The Guardian because it features rich media ads. Likewise, a fancy ad won’t entice consumers unless the ad messaging and actual product is compelling… it’s going back to the basics. Try spending 700K marketing dollars on a rabbit hole ad for Lincoln’s MKZ and see what the ROI is on that. Exactly.

  9. Photosynth looks similar to real estate virtual tour software. Nothing really new except it’s being repurposed to retail. I do think it will be accepted by users and advertisers.

    Seadragon/SilverLight mashup/overlay reminds me of Idealab’s (Snap’s) link previews. Honestly I find that widget/tool annoying at times. Isn’t it just another form of pop-ups?
    This too will fly with advertisers but I am not sure users will be well served.

Leave a Reply

Your email address will not be published. Required fields are marked *