Microsoft Corp.’s online advertising researchers will spend this year teaching computers to be smart about sticking ads into video clips, and to be even smarter about targeting ads to specific web surfers.
Microsoft showed off a handful of early-stage advertising projects at its headquarters Tuesday that may or may not turn up as part of Microsoft’s Web advertising platform.
The demonstrations come just days after Microsoft’s $44.6 billion bid for Yahoo Inc., which, if successful, will boost the software maker’s Web traffic and online ad revenue.
With its 2006 acquisition of aQuantive, the software maker gained a broader network of Web sites on which to sell ads, and tools to help marketers buy them.
A few of Microsoft’s projects were aimed at helping advertisers get better at reaching their ideal customers online, particularly using search keywords.
The company showed a dashboard advertisers could use to forecast the success of certain keyword advertising campaigns and a system it says will make it easier for advertisers think about key ideas, rather than hundreds of individual keywords.
But most of the adCenter Labs prototypes had little to do with search.
“Search itself gets a lot of attention because of Google,” said Tarek Najm, a technical fellow at Microsoft. “Advertising in search, as a result, gets a lot of attention.”
Najm said spending on search keyword ads will be dwarfed by what marketers spend on other types of online advertising, such as placement based on ‘audience intelligence’ and display ads including video.
Microsoft – along with Google and other competitors – is also hard at work on new ways for companies to advertise their brands to Web surfers watching video clips.
One crunched a clip, looking for the most appropriate stretch of time and spot on the screen for an advertiser’s ‘bug’, or logo. For example, if a car company wanted to show its logo for 10 seconds in the bottom right corner of the screen, the computer program would find the 10 seconds in which the logo interferes least with the action in the video.
Another used speech recognition to make a transcript of a video, then served up ads – in the demonstration, they were text links – alongside the video. As the topics discussed on screen changed, so did the ads.
The third program scanned a video for surfaces where ads or product images could be inserted later. The demo showed how the same frames could display a Coke ad one moment and a Pepsi ad the next, without having to reshoot the video.
Other experiments included an interactive shopping kiosk that used elements of Microsoft Surface, a next-generation touch screen, to show ads and coupons, and a computer program that helped marketers avoid accidentally putting their brand on a web page with distasteful content.