RealTimeBiddingは、「リアルタイムターゲティング」ではありません。
“Data Driven Thinking” is written by members of the media community and contains fresh ideas on the digital revolution in media.
Today’s column is written by Dave Zinman, CEO of Infolinks.
When you buy display ads in an RTB environment, you’re targeting audience based on intent. Your assumption is that you’re reaching an audience while that intent is still fresh – that the message you’re so carefully placing along their path is relevant to them the moment they see it. Truth is, it’s probably not.
Nearly all impressions purchased through RTB are targeted based on past actions using cookies that are too old to reach shoppers in their “moment of truth,” and could even be days or weeks old. Only a very small percentage of the more than 10 billion impressions per day available via RTB are based on real-time interests and intent. As an example, shoppers may read an article about home decorating trends, visit the Pottery Barn site, and then spend some time on HGTV.com. We can assume that these people are about to either decorate or renovate a home.
In an even worse scenario, they may have been searching for a particular rug or chandelier during that session – which may been have a week ago – and have since purchased one or the other and moved on. So the ads they’re seeing for rugs and chandeliers are completely moot right now. (I can honestly tell you that I’ve had a pair of running shoes follow me around the web for weeks. Long after I’d purchased a pair offline, a pair I’d viewed from a certain online retailer appeared on nearly every site I visited.)
Although RTB leverages sophisticated technology, this method of targeting is seriously limited and puts display at a disadvantage when compared to search, which truly does target real time intent. If I type “lightweight running shoes” into Google or Bing, milliseconds later, I will have my choice of thousands of pages featuring the latest and greatest in athletic footwear, where to buy them, how much they cost, and articles recommending the best running paths, socks and water bottles.
Why can’t display behave the same way? The context of many sites provides enough real-time data to deliver real-time targeting. I recently read an article that said that online advertising was the only industry in the world to be satisfied with a .1% success rate. Is this the best we’ve got, really? We can improve engagement rates for display by facilitating real time targeting at a granular level.
Think about how search works. The user types in query, and they’re served with the content they’ve requested, along with ads that are highly relevant to that query. When people consume content online, the context often provides nearly as much information about their intent.
Search also makes real time decisions about whether to even serve ads or not based on whether there are relevant advertisements to show. Nearly half of all search results pages do not contain ads because the queries contain no commercial intent. Why do we serve up display ads on every page, regardless of whether the consumer has any commercial intent at that moment? Has it occurred to anyone that the relentless delivery of irrelevance enabled by our scalable ad targeting and delivery systems contributes to anemic response rates?
Until we can get relevant and granular with display targeting, I’m not sure we’re ever going to see a reversal in click-through and engagement rate declines. Speaking from my own perspective, if display ads consistently delivered what I wanted right now – like, in this moment, a whitepaper from an ad solutions provider that can target real-time context – I’d be clicking! I’m not suggesting that ad servers need to have psychic capabilities, but there’s enough data available to serve ads based upon the current desires and interests of consumers. Let’s use it to target ads not only accurately – but with immediacy.
Follow Dave Zinman (@davebehappy) and AdExchanger (@adexchanger) on Twitter.
from:
http://www.adexchanger.com/data-driven-thinking/real-time-bidding-aint-real-time-targeting/
