I had an interesting experience last month that has culminated in some pretty spectacular results in my AdWords campaigns using Conversion Optimizer (CPA) bidding.
This experience was with my Never Cold Call Again site. My business model on that site is as follows: Users reach the landing page, and the offer is to download a free 37-page PDF preview of the full course. They then receive a follow-up email series enticing them to purchase the full course, which is a hardcopy workbook and 2-CD set that ships to them along with some additional downloads.
I’ve always tracked conversions via signups in AdWords Conversion Tracking, striving for about $3.00 per signup CPA. However, last month I decided to change my strategy, based on the reasoning that just because someone signs up for a free download doesn’t mean they’ll necessarily buy the full product, and in fact a huge percentage of signups are probably just freebie-seekers who have no intention of buying down the road.
So with that in mind, I removed the conversion tracking snippet from the signup thank-you page and let AdWords track only sales conversions.
But, upon reviewing a large amount of tracking data for the year, I quickly reversed course, and here’s why:
1. A huge percentage of customers purchase more than 30 days after the signup date, and the AdWords conversion tracking cookie only lives for 30 days.
2. An even larger percentage of customers’ sales are not tracked at all, because they either sign up and purchase from two different computers, or because they’re blocking third-party cookies in their browsers.
So, about 36 hours after removing the tracking snippet from the signup thank-you page, I added it back in.
And something interesting happened:
First, traffic plummeted to a very low volume. Apparently AdWords saw that signup conversions dropped from several hundred a day down to zero, and stopped sending traffic as a result. I emailed one of my reps at Google who reassured me that things would come back up in a few days, which they did.
And here’s the second, and more interesting, effect that this had: Traffic not only came back, but conversion rate nearly doubled! And on top of that, over the past week, traffic volume itself has nearly doubled as well, meaning an astronomical increase in conversions!
There’s no way to know, specifically, why this happened, but my best guess is that removing the code for a short period of time caused Conversion Optimizer to “reset” itself and begin tracking conversions all over again from scratch. This may have had the effect of purging “stale” sites from the historical data, that may have converted well in the past, but since stopped performing.
In the future, if I see Conversion Optimizer results starting to sag, I may try this again and keep my fingers crossed. I’ll certainly report the results here on my AdWords blog.