Technology Raises the Targeting Bar"Dramatic technological changes in the online media marketplace will soon change the way all media is bought, sold and utilized" is a phrase we hear over and over. Online advertising became the only electronic medium to exceed $4 billion in revenue in its first five years. And Jupiter Communications, New York, predicts online advertising revenues will grow to $28 billion worldwide by 2005. But how will this happen?
The promise of cost-effective direct marketing to individual consumers is the driving force, and it's possible today. Advertisers that use optimization technology already reach only those customers most interested in their product or service and build one-to-one relationships with them.
So what is optimization technology? In the beginning (about three years ago), advertisers rushed to place campaigns across the Web with no clear goal or return on investment metric, other than possibly blanket brand building. Cost-per-thousand impressions pricing was used because it was the model for other broadcast media pricing. Early Web campaigns were simply faster and cheaper versions of offline mass marketing.
One of the first steps toward giving advertisers greater accountability was cost-per-click pricing using banner click-through rates as a quantifiable metric. Click-through rates record a consumer's initial interaction with an ad, but they are insufficient to supply advertisers with measurable ROI.
What is needed is old-fashioned math and the latest computing systems to analyze the huge amounts of data that the Web collects. This combination is referred to as optimization technology.
There are two parts to the equation: deciding where the ad goes and how it gets there.
With real-time optimization, you can maximize for any function the advertiser wants: geography, time spent on the page, registration, sale, time of day, etc. The system learns who should see an advertiser's message and adjusts the campaign. The learning progresses faster if different delivery channels -- Web, e-mail and desktop -- contribute data to a central decision-making location.
Once optimization technology is in place, you can pinpoint and maximize advertiser ROI. Optimization can occur at the post-action stage, which we call conversion tracking and optimization.
Here's an example of a test program for a leading financial client:
• Criteria are set as clicks and resulting qualified leads. Qualified leads are those leaving name, address and contact information for follow-up. The client has to commit to tracking fulfillment through traditional channels, such as branch office and phone.
• Collect click-stream data off the pages of an advertiser's Web site.
• Crunch numbers and decide if the campaign is ready for full rollout. Key issues are conversion and close rates on qualified leads, and closing ratio for all channels -- phone follow-ups, branch visits and online inquiries.
• Cost per qualified lead can be pinpointed and a decision made to proceed or continue testing. Aside from an initial start-up fee, the program cost is based on performance.
This dynamic will revolutionize offline media buying when the broadband future arrives. Offline media will have a new meaning. The pricing model of broadcast today is predicated on finite inventory (24 hours a day). Waste is unavoidable when purchasing television and radio, and there is no interactive component. Digitized transmission allows for the almost infinite expansion of inventory.
The entire concept of what we call a network soon becomes obsolete. "Networks" become aggregated users who match the criteria of the product. They are dynamically created for each individual campaign and are refined as more data are collected. Optimization technology makes this possible and brings the pricing metric full circle. Only effective optimization technology provides precise enough targeting to make cost-per-conversion pricing viable.
Reach-based pricing will never disappear, and some sort of hybrid pricing will probably evolve, as is now happening on the Web. But pricing will be based predominantly on the richness of the selects desired and on proven, quantifiable results.