In an earlier column, I advised readers to monitor performance of keyword-targeted campaigns on a site-by-site basis using Placement Performance reports. Based on these reports, poorly-performing sites can be excluded using the Site Exclusion tool, and sites performing well can be "peeled off" and placed into a separate placement-targeted campaign. By doing so, you can set maximum bid prices separately per site, optimizing ROI for each one.
Several readers have written to report that they've attempted to implement this strategy, but found that their new placement-targeted campaigns failed to produce many (or any) impressions and clicks, when they set maximum bid prices equal to the original keyword-targeted campaigns.
(Remember that after peeling the top-performing sites from a keyword-targeted campaign, make sure to exclude them from the source campaign using the Site Exclusion tool. Otherwise the original ad group may "compete" with the new placement-targeted one.)
We did some research and discussed the issue with our sources at Google, and here's what we found:
Placement-Targeted Campaign Bidding Strategies
As discussed in past columns, click-through-rate (CTR) is the most important factor when Google is determining quality score for placement-targeted campaigns. According to Google, a new placement-targeted campaign starts out with a low quality score – an intentional "demerit" designed to inhibit "spammers" from deluging AdSense publisher sites with low-cost, low-quality ads.
Conventional wisdom among many content network advertisers says it's prudent to employ a "start low, go high" bidding strategy. Their reasoning is that content ad clicks are inherently "low-quality," so it's safest to start maximum bids lower than for search campaigns, and then gradually move bid prices higher as ROI data is collected. That way runaway losses – high click prices and poor ROI – can be avoided.
Our latest thinking is that the "start low, go high" strategy might doom some content campaigns – placement-targeted ones, and even keyword-targeted ones if non-text ads are used – by cursing them with low quality scores.
In fact, the strategy we're testing now is "start high, go low" – starting peeled-off placement-targeted ad groups with maximum bids higher than their keyword-targeted counterparts. The thinking is that we'll buy higher quality scores right away, and can then lower them if the CPCs cause unacceptably high costs-per-conversion.
How high should you set your initial bids? Our tests have indicated that bid increases of 50%-200% may be necessary. The goal should be to ensure your ads appear within positions 1-4.
CPC or CPM for Placement-Targeted Ads?
One reader reported that switching from CPC to CPM pricing in the new, peeled-off placement-targeted ad groups ensured that ads would begin running sooner. Other readers have asked whether CPM or CPC bidding is, in general, preferable for placement-targeted ads. Here's Google's advice for content ad bidding methods in general:
The CPC model is well-suited for conversion- or traffic-oriented campaigns, where advertisers are focused on clicks and tend to track metrics such as sales, leads, sign-ups, etc.
The CPM model is geared towards campaigns that are focused on maximizing impressions, delivering the advertiser's message to buyers early in the buying cycle by increasing visibility and brand awareness.Since different types of sites encourage different types of user behavior, we also recommend that you select your bid type for different groups of placements based on the likelihood of users clicking on your ads. For example, if you're targeting placements on sites with reviews about your product, and you're running a relevant ad, the likelihood of users clicking on your ad is high. In this case CPC bidding may be your best option. If you're targeting placements visited by your target audience that aren't directly related to your business, such as placements on news or general interest sites, your ads aren't likely to have high CTRs. In this second case CPM bidding would be a more appropriate choice.
We're currently testing CPM vs. CPC pricing for placement-targeted campaigns, and will report back on our results in a future column. Having other problems with content campaigns? Send me comments and questions via the feedback form below or in the SEW Forums Content Advertising thread.