Real-time bidding (RTB) is quickly becoming one of the leading methods of selling and purchasing online ad inventory – it is expected to grow 72% this year, and will be a $7 billion market by 2016. A recent paper from the University College London sheds new light on the mechanics, subtle patterns, and areas of improvement in real-time bidding.
Real-time bidding was introduced in 2009 as a way to instantaneously sell a publisher’s ad inventory to a large pool of advertisers. This auction system allows individual impressions to be bid on, thus streamlining the media buying process, increasing exposure between advertisers and publishers, and increasing ad relevancy for customers.
As soon as a user visits a page, information about the user and the ad inventory is delivered from a Supply Side Platform (SSP) to an Ad Exchange, which holds an auction for each placement. Advertisers instruct a Demand Side Platform (DSP) to make smart bids on their desired audience. As in a traditional auction, the inventory is awarded to the highest bidder. Where this differs drastically, however, is that this auction takes place entirely over the span of 300 milliseconds.
Traffic trends & bidding behaviors
In this June 2013 study, the researchers picked apart patterns of impressions, clicks, and bids placed by advertisers. They found:
- Websites often had fewer visitors on the weekends
- Conversion rates (CVR) were generally higher in the daytime vs. sleeping hours, but they were highest between 6PM and 9PM.
- The number of impressions peaks between 9AM and 12PM
- In contrast, the number of bidders peaks at 9AM
Since winning bids are selected by an auction mechanism, the cost of winning bids is driven up during the morning – an unproductive trend, given that we’ve already learned that CVR peaks in the evening.
So why does this happen? Advertisers who want to avoid the low CVR crowd of 12AM to 6AM may choose to opt for no daily pacing of their advertising budget. This set-up depletes the budget as quickly as possible, thus winning lots of impressions in the morning, albeit at the price of inflated bids.
A logical solution to this conundrum involves dynamic pacing: in other words, spending more of your advertising budget during the hours in which you know there is a higher potential for clicks, conversions, etc. Unfortunately, no current DSP platform is able to analyze CVR in realtime and reactively change its budget pacing. However, advertisers could look at CVR analytics within their existing DSP and manually set up bid multipliers that vary according to the time of day.
Frequency & Recency
Two other factors that are integral to a campaign’s success are frequency and recency.
- Frequency factor (FC) refers to the number of times a creative is shown to a single user.
- Recency factor (RC) defines the time period before an ad can be shown to the same user again.
Just as ideal FC and RC times can vary drastically between campaigns, they can also differ for each individual; some people may prefer to only see an ad once, while others may require fervent repetition to drive home the message. At the moment, RTB systems can adjust FC and RC on a campaign-by-campaign basis, but lack the ability to fine-tune these variables for individuals. This is a gaping omission from a system that was designed to deliver customized content to visitors.
To get the most out of an ad, advertisers should test out various FC and RC times for their campaigns – the last thing you want to do is to waste advertising dollars on customers who are tired of your message. Some DSPs offer Frequency/Recency distribution reports, which detail the impressions and conversions logged with each FC/RC setting. Use this information to optimize your existing campaigns!
What’s next in the world of RTB?
While RTB has greatly streamlined the bidding process for display advertising, there are still many areas under research that could further enhance advertiser efficiency. For example, Turn.com is working on a system that automatically adjusts budget delivery based on conversion data, and Amazon and Groupon have published optimization strategies for bids on group-buying ads.
For now, advertisers can turn to solutions such as SiteScout, which offers a number of options to track conversions, cap ad frequency, and specify budget pacing. While technically not a true RTB system, similar elements are also available in the Google Display Network through the Adwords Conversion Tracker tool and custom ad scheduling feature.