RSS feed Get our RSS feed

News by Topic

BizReport : Advertising : March 15, 2012

How to get more from RTB

Proponents of Real Time Bidding (RTB) say it puts the right before the right consumer are the right time. Opponents say doesn't work properly. The key, of course, is somewhere in the middle and that 'you need more that a good ad' factor has left some brands wondering if RTB is worth it. Here's one expert opinion:

by Kristina Knight

thinkrealtime.pngKristina: RTB has gotten a lot of positive buzz over the past year, but the negative is now beginning to surface. What do you say to those who don't believe RTB options benefit brands?

Ezra K. Doty, President, Think RealTime: RTB is also about efficient pricing for ad real estate and paying the right price for each individual impression. What determines the right price needs to be based on both the probability that the impression will hit the mark, and the monetary value of the desired user action. For example, a user who browses multiple pages on the site is more engaged, and therefore more valuable, than a user that bounces off of the homepage after only a few seconds. This second part is often overlooked by Brand advertisers, but is something they should focus on more;- similar to the approach that some Brands have been taking when using paid search to drive brand objectives. Once the space is bought, delivering the right message is another (but separate) driver of performance.

Kristina: How should brands approach RTB as we move farther into 2012?

Ezra: Demand more. Define clear goals that play to the strength of RTB, which is data optimization against defined goals. Ask your vendor to focus on things like site interaction from users who've seen your ads - time on site - producing a certain number of page views, visiting key areas like the store locator, contact us page, etc. Put a value on these behaviors and task your RTB partner with the objective of attracting more prospects who exhibit them. The automotive industry (and others) do this with KPI (Key Performance Indicator) goals. Brand advertisers now have the ability to benefit from direct response technology and they should take advantage of it.

Kristina: What are your top 3 tips for advertisers interested in RTB options?

Ezra: 1.) Be open to new things. It is still very early, and a number of DSP's and Trading Desks that advertise RTB capabilities built their core platforms prior to RTB's existence. Marketers rushed to try these services on the heels of exciting press that focused on the promise of RTB, and at times have been underwhelmed by the results of technology that was not fully baked. The truth is that designing systems that can leverage true RTB is complex and time consuming work. Getting a seat on an ad exchange, building a simple bidder that allows you to check a box that says "we have RTB capabilities" is easy, we did it in about 2 weeks. Creating a self-learning platform that can calculate the probability of an individual impression driving an action, enter a bid based on a combination of this probability and the value of that action, and do this with any degree of accuracy at a rate of 50,000 to 100,000 times per second, is intensive in terms of hardware, software, and network infrastructure. It takes time, and the race didn't even start until October of 2009. Good results are very possible and will be occur more frequently as the industry continues to mature and advance.

2.) Set clear goals as described above and ask questions. Some vendors claim to be RTB enabled but really only use a fraction of the available data to make bid decisions. Ask your RTB vendor how many impressions per user per month they serve on average? What data are they looking at to make decisions? Do they bid a different price on each ad unit? Do these different bid decision prices vary widely based on the data, time, location, intent score of the user? Ask them to describe RTB to you so you can determine their level of sophistication?

3.) Let the magic happen. Applying insights from first generation ad network buys may hurt more than help. Setting things like narrow geographic restrictions and frequency caps is useful with manual optimization, but may work counter to the logic of a programmatic system which analyzes data at a more granular level. A well designed machine learning system is open to all possibilities, and will uncover insights about consumer behavior than could ever be found through the traditional test and control approach.

Tags: ad targeting, advertising tips, advertising trends, online advertising, real time bidding, RTB trends, Think RealTime

Subscribe to BizReport



Copyright © 1999- BizReport. All rights reserved.
Republication or redistribution of BizReport content is expressly prohibited without the prior written consent.
BizReport shall not be liable for any errors in the content, or for any actions taken in reliance thereon.