Your Title Here <data:blog.pageTitle/>

Monday, September 24, 2007

Google Conversion Optimizer vs Manual Optimization

Earlier today Google announced a new AdWords service - the conversion optimizer.

Here's a snippet from the AdWords blog explaining what the conversion optimizer will do:


"Here's how Conversion Optimizer works: simply specify a maximum CPA bid and the Conversion Optimizer does the rest. It uses historical information about your campaign and automatically generates optimal CPC bids for each auction. You still pay per click, but you no longer have to manually adjust your bids to reach your CPA goals. Since the Conversion Optimizer can choose a new bid for each auction, you're provided with the additional benefit of spending money only on the search queries and sites where your ads are likely to convert. You can read more about how the Conversion Optimizer can manage your costs here."

I'm going to put the optimizer to the test. I pulled stats from one of my campaigns for the last 30 days (8/24 - 9/24):

Conversion rate: 38.60%
Cost per conversion: $0.03

This campaign has had in excess of 40,000 conversions during the previous 30 days so there should be plenty of data for the conversion optimizer to work with.

I just flipped on the conversion optimizer for this campaign and will post periodic updates as well as a complete summary when the test concludes. My hunch is good old fashioned manual brain powered optimization will prove more effective although part of me hopes I'm completely wrong and the conversion optimizer blows that conversion rate away.

If you would like to give the AdWords Conversion Optimizer a spin here are the instructions for getting started. Keep in mind for a campaign to be eligible for the conversion optimizer you need to have 300 conversions over the last 30 days in that specific campaign.

Labels: ,


21 Comments:

At 12:01 PM, Blogger mmackey said...

Can't wait to see how it goes, Jeremy!

Melissa

 
At 1:21 PM, Blogger Jeremy Mayes said...

Thank you for the comment Melissa.

I just checked today's stats and if today is any indicator of what the next 29 will bring I was right, good old fashioned manual brain powered optimization will take the prize. Today's stats:

Conversion rate: 32.61%
Cost Per Conversion: $0.05

Conversions are down and costs are up...

 
At 1:22 PM, Blogger searchquant said...

Surprise, surprise, surprise.

Something tells me Google didn't build this for advertisers who already know what they're doing.

 
At 2:48 PM, Blogger Jeremy Mayes said...

searchquant,

Something tells me you're right.

Perhaps the data will prove us both wrong...perhaps:-)

 
At 10:12 AM, Blogger Jeremy Mayes said...

Just a quick update - performance is actually getting worse instead of better. Here are the stats since I started:

Pre Conversion Optimizer:
Conversion rate: 38.60%
Cost per conversion: $0.03

Conversion Optimizer Day 1
Conversion rate: 32.61%
Cost Per Conversion: $0.05

Conversion Optimizer Day 2
Conversion rate: 24.07%
Cost Per Conversion: $0.06

 
At 5:34 PM, Anonymous Zachary Fox said...

Thanks for your comments, they led me over here, and I'll keep an eye on this post.

I'm doing some testing as well. With dramatically different conversion rates and cost per conversion. Maybe that will make a difference?

Only time will tell, hopefully this isn't one of those "maximum profits for us" optimizers that Google is so fond of.

 
At 9:31 AM, Blogger Jeremy Mayes said...

Hello Zachary,

Thanks for the comment. I'll keep on your test results as they become available, feel free to link to your test from here when you start posting results.

Jeremy

 
At 9:46 AM, Anonymous Jon said...

Hey Jeremy,

Thanks for doing this test. It really helps the rest of us :)

However, I wanted to point out that your previous conversion numbers were pretty high. Actually, I think that Conversion rate: 38.60% // Cost per conversion: $0.03 are the highest conversion numbers I've ever seen.

I'm not saying that this invalidates your research in any way. But what I ask myself, besides the obvious help for AdWords newbies, what could we expect for campaigns with more "mediocre" conversion rates (3%? 5%?).

I think that maybe, from where you standed, there's nowhere to go but down. And I congratulate you for that :)

Best,

Jon.-

 
At 11:08 AM, Anonymous Zachary Fox said...

Jon,

I think I'm testing with numbers that seem more applicable to most of us. At least I hope so. Just for an idea, here's my numbers for the 7 days prior to turning on the Conversion Optimizer:

Conversion rate: 1.9%
Cost per Conversion: 100.81
Conversions: 78

This is a campaign using a very broad match that hasn't been optimized other than adding negative keywords.

If you want to keep an eye out, my original post is here.

Jeremy, thanks. I'll update here when I post my follow up.

 
At 2:37 PM, Blogger Jeremy Mayes said...

@ Jon,

I knew the numbers were on the high side (conversions that is) but figured since I had such a high number of conversions over a 30 day period (over 40K) my account should have plenty of historical data Google's conversion optimizer could use to improve the campaign.

I don;t know that I'll continue the test much more than few more days...just have to wait and see I suppose. Thanks you for your comments.

@ Zachary,

Keep us posted on your test and results - looks like a nice campaign to test things out with.

 
At 3:04 PM, Blogger Jeremy Mayes said...

Quick update -

I realized when I wast setting something up I broke my conversion code...doh! The day 2 stats I posted are not accurate. The day 1 stats are accurate.

I have everything corrected and will post some updated results next week.

Jeremy

 
At 11:34 PM, Blogger searchquant said...

Jeremy_Mayes:

Over on Webmasterworld someone offered up that if Conversion Optimizer wasn't beating your pre-C.O. performance that you could try passing it a [<1] conversion value to try and steer it to the right point.

Do you plan to try that? I'm very interested to see the results if you do, as well as your overall results.

Also, you gave pre and post conversion rate & Cost Per Conversion, but what about daily adspend?

 
At 12:53 AM, Blogger Terry Whalen said...

Searchquant,

I think ad spend can be inferred - 40,000 conversions at $.03 cost/conversion would be about $1,200 spend in this campaign in a 30-day period.

I'd also be interested in seeing whether passing a (really, really low) conversion value would make a positive impact, although my bet is this wouldn't have much of one, given that Google is already saying they will absolutely try to maximize overall conversions given the daily/monthly budget. So, they're saying they're going to take a portfolio approach and get the cost/conversion down as much as they can, regardless of value of a conversion. But you never know...

 
At 8:30 AM, Anonymous Amit @ DewDropSEM said...

Your cost per conversion is pretty low....I think you have something different as a conversion than a sale...

 
At 8:55 AM, Blogger Jeremy Mayes said...

searchquant,

I don't have a conversion value set at all. I could try and set one but am not sure how that would impact things. Could you post a link to thread where that's being discussed?

Terry Whalen wasn't too far off on the spend.

Amit @ DewDropSEM,

A conversion is not always a sale. In this case/test I'm not selling anything, users are completing an action or series of actions.

 
At 9:43 AM, Blogger Jeremy Mayes said...

Ending the Test

I let my conversion optmizer test run over the weekend and at this point have seen enough. While it was able to increase my conversion rate in some cases my cpc was up across the board which in turn caused my cost per conversion to almost double.

Over the last 3 days + pre optimizer numbers:

Pre Conversion Optimizer:
Conversion rate: 38.60%
Cost per conversion: $0.03

Sep 28, 2007
Conversion rate: 38.61%
Cost Per Conversion: $0.05

Sep 29, 2007
Conversion rate: 38.84%
Cost Per Conversion: $0.05

Sep 30, 2007
Conversion rate: 36.40
Cost Per Conversion: $0.05

Increasing my conversion rate is great - almost doubling my cost per conversion - not so great.

If you know of any other conversion optimizer tests feel free to link to them from the comments. Would love to see how it's working for others.

Related:
http://www.blogation.net/2007/09/google-conversion-optimizer-why-being.html"

 
At 4:51 PM, Anonymous Zachary Fox said...

So here's my results. Pretty good. Certainly better than the ones that Jeremy got. Perhaps it's the difference in numbers, or how we count a conversion. Since Google won't really tell us how it works, we'll never know.

Conversion Optimizer Results

 
At 11:18 AM, Blogger Jeremy Mayes said...

Thanks for the update Zach. Are you planning to leave the conversion optimizer on for your test campaign?

 
At 4:36 AM, Anonymous Anonymous said...

The optimizer needs at least a week to settle down so any results reported before that time are meaningless.

I aim for 0.50 a conversion and yet I bid as much as 2.00. It's coming in just over budget so far after more than a month using the tool.

Manual bidding is probably more effective but if like me you have a large account with only one pair of hands, it's very worth while indeed.

 
At 12:42 AM, Anonymous PPC-Management said...

To Conversion Optimize or not to... Hmm I say no for now. Going back and forth on this for a few days now....Thanks for the input, you all helped me decided to stick with the manual bidding.

 
At 12:55 PM, Anonymous Shoffy said...

I just started dabbling in this as well. So far, day 2, i'm not impressed. Total click have dropped like a rock.

However, like any software, it needs to learn. So.... ::crossing fingers:: I'm going to gut it out a little longer.

 

Post a Comment

PPC Discussions Home

Subscribe to this site About PPC Discussions

PPC Discussions is a blog about paid search authored by Jeremy Mayes, a search marketer from Illinois.

Subscribe to PPC Discussions via e-mail

Advertising

PPC Search Engines

The Big 3 in PPC

Subscribe to this site Previous Posts

PPC Discussion Archives

Subscribe to this site Miscellaneous

Add to Technorati Favorites

Powered by Blogger