Too often it is still easier to search the entire Web than a single site. In-site search leaves much to be desired in its ability to provide users with accurate search results, let alone assistance in performing more accurate searches. Israeli PredictAd is attacking this challenge head-on by offering auto-complete functionality for search boxes with an added twist: monetizable real-estate.
Already deployed across 6,000 websites and generating millions of monthly impressions, PredictAd is announcing the launch of PredictAdPlus+. Targeted at media networks and high-traffic websites/blogs, the offering provides greater control of the look and feel, customization, dedicated databases, and advanced display options. Most importantly, it allows a number of sister sites to be integrated into a single search box. As an example, the TechCrunch search box at the bottom of this post incorporates CrunchGear, MobileCrunch and CrunchBoard. The auto-complete needs time to learn, but you’ll get the general drift.
PredictAd is making 250 invites available to TechCrunch readers. Sign-up here.
PredictAd’s twist on search monetization begins by helping users refine their in-site search queries by way of real-time search suggestions—think Google Toolbar’s auto-suggest feature. PredictAd continuously optimizes its suggestions based on the community search patterns for any given site. Beyond collaborative filtering PredictAd also takes into account historical searches, geo-location, time of day, common misspellings, connections between words, and more.
Not only should users benefit from more accurate searches, site-owners should benefit from increased page views as a result of user exposure to additional site content. Site-owners can also benefit from SEO insights drawn from the service’s analytics backend.
This is where ad monetization comes into play. PredictAd injects contextual advertising into the search box. The ads—text or image—change dynamically as users type out their search queries. In practice, PredictAd creates its own ad real-estate and then monetizes it. An important element in PredictAd’s offering is the fact that the ads are displayed directly in the user’s focal zone, unlike traditional ads which are susceptible to “banner blindness”. Rev-share is 50%/50% on run of network ads, and 80%/20% (for the publisher) when the advertiser selects specific publisher sites. PredictAd is claiming click-through rates in range of 0.5%-4%.