|
Wednesday, October 18, 2006
 This screen shot is from a test campaign I ran this week (to see full size versions of screen shots here, click on them). The drilldown on any ad group's bid or individual keyword bid allows you to project how bids will translate into ad positions and their expected click volumes in the upcoming month. This is just for a single keyword, "interior decorating."
Slide the bar over a bit, to find out what it would cost to generate double the number of clicks (686 up from 343) on this keyword. Interestingly, the average amount I'd have to *bid* to generate that number of projected clicks in a month on this keyword is well more than double the current bid - $2.70, up from $1.03. That would achieve 94% of the highest potential click volume, twice the 47% achieved with my current bid. By bidding another 35 cents higher than that, you'd eke out a little more volume, to fully 100% of the highest potential clicks, with an average ad position of near 1.0.
But get this: due to ad discounting and the behavior of other advertisers, my actual average cost on this keyword when bidding that high is only projected to rise by 21.5%, to $1.13. In fact, the tool projects an average actual cost per click of $1.13 even if you bid all the way up to the 1.0 average position, projected at $3.05. Leaving a moderately high bid "on" all the time could be at least as cost-effective as bidding very high and governing daily impressions with strict budget limits, and volume is gained without a huge ROI hit in this case (or, depending on conversion rates, ROI could turn out to be around the same, or much worse if you convert better out of lower positions). With ad discounting and complex rank algorithms, many AdWords advertisers have learned that actual average bids aren't nearly as scary as the max bids you enter into the system. It's the actual amounts that need to be considered when making projections.
 Will playing around with this tool teach advertisers about the demand curve and about optimal ad positions for their purposes? Will it help them to understand that bidding properly is a form of budgeting to be used in tandem with actual budget limits (also built into Panama for daily and monthly amounts)? You bet. What it specifically teaches them about their own account strategy is up in the air, though. That's going to vary wildly by keyword and by business. All we can really say for sure is that a lot of changes are triggered by changes in bids.
On high volume keywords with a lot of competition, these projection tools are likely to be a bit out of whack. On low volume or geotargeted keywords, they'll be a lot out of whack. But as the system gathers more data for your account as well as for advertiser behavior in general, the projections will become more certain. You have to start somewhere, and this intuitive interface is a great start.
With a visual like this, will there still be advertisers who want to bid 20 cents on core keywords, and still generate as many clicks as possible (obviously, in a competitive environment, these are mutually exclusive goals, all other things like CTR being equal)? You would think that a picture here says 1,000 words to this type of wishful thinking.
Posted by
Andrew Goodman
View Posts by Category |
|
D'oh!
|
