Over the past month, I've been noticing "waves" of high and low conversion that follow no cross pattern (time of day, day of week, etc.), other than that a low period precedes a high period and vise versa.
Conversion rate always jumps around...that's just the nature of the beast. But it's becoming almost predictable; if I see a period of way-above-normal conversion, it will almost certainly be followed by a period of way-below-normal conversion.
Certainly, G has enough data at this point (thanks to the generosity of conversion tracking users) to know conversion by kw, position, time of day, etc. And certainly, they have an economic reason to keep us all close to breakeven rather than having some of us win big and others of us lose big. Could it be that they are leveling the playing field by manipulating how/where/when kws are fed in a way that achieves a target conversion rate.
Anyone else seeing the big waves?
Yes. I'm looking mainly at US Search traffic.
Do you separate your conversion measure between Google Searches, Search Network and Content Network, and then by country?
For the first time ever, my Yahoo campaigns are more predictable, and are out performing my Google campaigns. I'm definitely going to start focusing more time on my Yahoo campaigns. It's becoming quite obvious that the new Google algorithms are designed to squeeze every last dime out of everyone - And have absolutely nothing to do with quality.
Yes. I'm looking mainly at US Search traffic.
I am asking because we see huge fluctuations on the Content Network, but not on Google searches... Google searches are quite stable for us.
I am asking because we see huge fluctuations on the Content Network, but not on Google searches.
I'm seeing the big swings on Search, though, which would be a more difficult beast to manipulate I think.
After the Quality Score -- seeing how it acts to increase expense and normalize roi across a variety of accounts, and seeing how much trash it lets pass through its supposed "user experience" filter -- I'm just assuming there's a team of freshly minted MBAs working angles to optimize revenue. I'm assuming they know enough about math to work some pretty complex schemes, but not enough about running a business to understand the kind of long-term fallout that all this manipulation can have. So, any theory is worth considering at this point.
I may need to join the tinfoil hat club on this one. That makes perfect sense for G to maximize their revenue by keeping everyone afloat. Our fluctuations are week to week mostly. Last week - WAY above normal conversions (2 times), this week, WAY below (50% of norm). This pattern goes back about 5 weeks and is becoming very predictable.
Same exact keywords, same ad times defined, same competitors ads (i check them every day), same products, same pricing, same number of visitors (more or less). Everything that i can control or observe is identical...but conversions move wildly.
But I'd like to think that somebody at that company is bright enough to know what the consequences would be if it ever came out that Google were manipulating ad placement beyond what it disclosed to advertisers.
The word disastrous comes to mind.
AWA
we don't use your conversion tracking data in this way.
AWA, thanks for clearing that up. You never know when the stockholders will apply pressure for that to change, so I just keep my critical stats internal (and I advise everyone to do the same). But it's good to know the data isn't currently used.
Does the same rule apply to using the Smart Pricing algo and bid behavior to intuit relative conversion value of different Content partners and different Search positions?
For example, by watching how people bid towards different search positions on diffferent keywords, I could probably extrapolate a breakeven CPC and a relative conversion value for each keyword/position.
And by using whatever drives the Smart Pricing algo to reach its decisions, I could probably intuit which Content partners convert well for a keyword and which convert poorly.
And I could use this information to manipluate Search position and Content distribution to adjust conversion towards our breakeven CPAs.
It sounds wacky, but it would be an easy way to increase revenue, and it might explain why conversion has become cyclical.
Again, thanks for the clarification. Unfortunately, the wild speculation comes with the territory in using a "model of opacity".
The latter part of this statement is interesting - In this way? In what way do they use it?
