Google Conversion Optimiser

[AmeliaDawson]

Hi Guys,

I saw that Google Conversion Optimiser (beta) is now available on my Adwords account.

Has anyone tried this?

Would be great to know any results, or opinions on this.

Cheers,

Amelia
_____________________________

Added: Link to information about Google Conversion Optimizer (http://adwords.google.com/support/bin/topic.py?topic=12045)

[Shmuel]

I'm trying it right now (turned it on yesterday). I haven't really seen any change yet. I hope to see something by the end of the week.

[AmeliaDawson]

I would be interested to hear your results. Do you set the CPA's at Account or Ad Group level?

[Shmuel]

It's set at the campaign level. The campaign has had to have 300 conversions (so yes, you'll need the code in in order to use this) in the last 30 days. Only one of my campaigns fits this bill, so that's the only one that's on.

[AdWordsRep]

I saw that Google Conversion Optimiser (beta) is now available on my Adwords account. Glad to see an active thread on this subject - thanks for starting it AmeliaDawson. ;)

I hope you all will feel free to add your feedback here, whether positive or not-so-positive. I'll make sure that the team behind Conversion Optimizer sees it.

I would be interested to hear your results. Do you set the CPA's at Account or Ad Group level? It's set at the campaign level.

I just wanted to jump in to clarify. While Conversion Optimizer is turned on at the campaign level, different bids are set at the ad group level.

AWR

[Shmuel]

I just wanted to jump in to clarify. While Conversion Optimizer is turned on at the campaign level, different bids are set at the ad group level.

Right. For the campaign I have it turned on for, we have one set CPA, based on our requirements.

[GuyFromChicago]

Here's what I've seen with the conversion optimizer so far;

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

[Shmuel]

Here's what I've seen with the conversion optimizer so far;

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

What are you selling that your CPA is 3¢!?

[GuyFromChicago]

What are you selling that your CPA is 3¢!?

It's not a sale, it's a specific action.

[Mel66]

Here's what I've seen with the conversion optimizer so far;

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
Yikes. Not exactly a vote of confidence for Conversion Optimizer! I'll be watching for more.

Melissa

[Shmuel]

Just an update of my results:
Averages for this month (excluding test)
CTR - 5%
CR - 11.4%
CPA - $10.26

Monday (turned on mid-day) saw a severe drop in CR and a rise in CPA.
CTR - 4.9%
CR - 8%
CPA - $14.63

Tuesday saw slight increases in performance.
CTR - 6.2%
CR - 9.2%
CPA - $12.99

Wednesday had continued improvement
CTR - 5.5%
CR - 10.3%
CPA - $11.61

Thursday was yet even better.
CTR - 5.8%
CR - 10.8%
CPA - $11.29

So far today shows even more improvement. I'm going to keep this on over the weekend, and I will report back on Monday.

As to the poor performance initially and the subsequent improvements, my assumption is that the system needed time to learn. I believe I will continue to see improvements over the next week or two, and am willing to wait it out.

One thing I'm not so happy about AWR, but I understand why it's being done, is the requirement of needing the campaign to have had at least 300 conversions in the last 30 days. I have many campaigns with far less than that, but would still like to test this on them (in particular one which has been a thorn in my side the last few months). I assume that as this is more and more implemented, that the number required will be lessened?

[psurplus]

I've been keeping a close eye on this feature available in conjunction with AdWords. I am glad I found this thread to hear what users are actually saying. I'm not using this feature, but what are the tagets you can choose to optimize for? Just CPA? Conversions, ROI, etc?


Pros-
1. its free when other third parties charge hundreds, if not thousands of dollars for a service like this.
2. its simple to use as with much of Google's AdWords advertising platform. Over time account performance should improve as the optimizer learns.

Cons-
1. possible limited choices of objectives, only getting revenue based data, when some other services can utilize actual order profit. I suppose you could replace revenue data in the conversion tracking code to use order profit done at the sacrifice of accurate revenue reporting.
2. the requirements for 300+ conversions in a month is steep.
3. without the functionality of advanced tools like AdWords Editor, Position preference, Budget Optimizer, Site targeting, Advanced ad scheduling, Preferred cost bidding.

[Shmuel]

I've been keeping a close eye on this feature available in conjunction with AdWords. I am glad I found this thread to hear what users are actually saying. I'm not using this feature, but what are the tagets you can choose to optimize for? Just CPA? Conversions, ROI, etc?

It's just for whatever your consider a conversion, be it sign up or purchase. In my case, it's purchase.

[Shmuel]

Monday Morning Update

So has this worked? Yes and no. I had a 7% increase in conversions with a negligible rise in CPA. Why no then? The increase wasn't that big. September was a semi-flat month for us.

I am going to keep this on for another week at least, so only time will tell.

[GuyFromChicago]

I dropped my test today. The CO seemed to do ok in terms of the conversion rate but nearly doubled my cost per conversion.

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

[plking]

The above doesn't sound too promising, but a 3 cent conversion cost might be a corner case. I'd be surprised if too many of us had costs this low.

Anyone else out there using it??

[Discovery]

300 conversion min for the campaign to use the optimizer?

Sorry, that wont work for many unless you count a visit or a free ring tone download a conversion.

Most advertisers have very targeted small campaigns where 300 meaningful sales/conversions is a high mark. And if you have hit this mark it is not a campaign you wish to optimize and risk a poor result. If you're like me you want to optimize campaigns that are not performing.

Discovery

[AdWordsRep]

300 conversion min for the campaign to use the optimizer?Point well taken Discovery.

Please note, though, that this is not an arbitrary number. It's the amount of data needed in order to accurately predict when and where conversions are likely to occur - so that it can show ads under those conditions.

That said, I'll certainly pass your feedback along.

AWR

[Discovery]

AWR - Understood, predictions are based on data - (unless you're a politician). So am I green in thinking G couldnt simply stretch out the date range from which the conversions take place?

Discovery

[AdWordsRep]

...So am I green in thinking G couldnt simply stretch out the date range from which the conversions take place? Discovery, not being a part of the team that created the Conversion Optimizer (and, beyond that, not being particularly adept at statistical science either) I'm not really able to answer that.

And to be even more straightforward, if I did know the particulars I almost certainly wouldn't be able to mention them - given that our competitors surely read this forum. ;)

One thing I have learned over time, though, is that things are rarely as simple as they seem at first glance - whether we're talking about AdWords algos, or pinpointing an electrical problem in an aging car - to name just two.

AWR

[FSPInvest]

Agreed on the conversions required being a little steep but fortunately in the business I am in it's very achievable with a little wangling (Google make a few extra quid on people wangling their campaigns I am sure).

I now have two campaigns running the conversion optimiser and I can say that the results I have seen so far have been good.

Week 1 (prior to optimiser)
78 Conversions
£3.26 CPA

Week 2
65 Conversions
£2.19 CPA

Week 3
78 Conversions
£1.92 CPA

Week 4
104 Conversions
£2.47

My targeted CPA is £2.50, so we can see in week one after the optimiser my conversions were down but was well within the CPA, week two I was back to my original conversion numbers at an even lower CPA and the last week my conversions were at an all time high at just below the targeted CPA.

So this begs the question, can I get the same level of conversions with a much lower CPA?

Google being Google will always try and spend my actual CPA to make as much money as they can from me when actually it' may not be necessary.

As the weeks go my aim is to push my targeted CPA down until the system breaks and I can't get the same numbers of conversions for the lower CPA, I'll let you know how it goes :)

[Reseo]

I've just taken over managing a large AdWords account and have set it up with Conversion Optimiser. One of the campaigns is doing about 1,000 conversions per month - Here's hoping it will help Google get up to speed pretty quickly. Will keep you posted.

...and thanks to you all for supplying your case studies, it's good to get real-world feedback rather than 'spin'.

[FSPInvest]

Set your CPA's to what you want them to be, let the system work itself out then keep pushing your CPA's down gradually, eventually you will push it to breaking point, but then you will also realise you can get your CPA's a lot lower than you thought you could ... handy little tool if you use it in the right way ... my advice is use the website optimiser to get the best converting page before you use the conversion optimiser if you haven't already!!

Good luck let us know how you got on :)

[larryg]

Hi All,

I don't post much as I am a noob with little to offer yet. However, we just had one of our campaigns become eligible for CO so I started it this weekend. It definitely eats budget but I am waiting a few days to decide if it was worth it.

The scary part is I had to do it on my best performing campaign because my other ones aren't eligible yet.

Larry

[Discovery]

I was considering using the optimizer for one campaign that has lumbered along with just over 200 conversions at our highest acceptable CPA for months.

When going through the set up Google gave me their estimate for a target CPA... Nearly 40% higher than our current CPA. In the help sections and in other Google posts they state that it's recommended that you use Googles recommended CPA bid price.

A simple question.

Initially what should I expect if I:

Use my current CPA - get more conversions at this cost?
Use a lower CPA Bid and get same # of conversions?
Use Lower CPA and raise the budget for More conversions at a lower CPA?
Use Googles CPA bid - get 40% more conversions at a totally unacceptable CPA (Ha).

Discovery

[larryg]

FWIW, here has been my experience so far. The recommended CPA was well above our average CPA for the campaign but I went with it anyway. Initially, the daily average CPA came in above the CPA bid but has steadily been coming down. I am also slowly dropping the bid to see if I can get it into a better range without loosing conversions.

Overall, the CO is generating more conversions but not every day. Our customers are a bit fickle in thier buying habits and I think that might be throwing the CO off a bit. CTR's are up about 2% and that has to help ad rank. I can see that spending really does fluctuate throughout the day compared to my campaigns that are on a straight schedule. I am going to allow it to keep running and learning and see how it does.

Larry

[Discovery]

Larry,

Thanks for your input. How long have you been testing this.

How long was the CPA higher
When did it start to decrease
When did you level out to higher conversions at the same CPA
Overall since you started are you ahead or behind in regards to overall cost per acquisition?

Thanks - Discovery

[larryg]

Well, I actually wrote down some numbers to share. For starters, my CTR isn't that much better, I had glanced at the wrong campaign...

I have only been using the CO just over a week so I have a while before I decide how good it really is. However, at this point it is already heading in the right direction and that seems like a decent "learning curve".

For comparison, here are the numbers from the last 7 days vs the 60 days before starting CO...

CTR increased from 5.21% to 5.34%
Avg CPC increased from $0.52 to $0.54
Conversion Rate increased from 1.44% to 1.67%
Cost/Conversion decreased from $36.17 to $32.58 ($41.48 was the recommended bid, currently at $39)

Larry

[JBrown]

I recently turned off CO for a campaign. It was performing ok, but I want to manage bids in a more hands-on manner for awhile.

Strangely enough, I had CO running for maybe 2 months and I now have a QS of 'Great' for nearly all keywords and a minimum bid of $.01. I don't know if this is temporary, but that's a vast improvement over where they were before CO.

Maybe CO is a way to boost QS.

[tdwhalen]

I recently shared Conversion Optimizer performance with some of the adwords agency folks. Here is what I shared with them:

Just a quick note about a small conversion optimizer test I did in case anyone is interested. Overall I view performance as poor, and this may be partly because there are not hundreds of keywords that bring this account conversions (there are really only a few dozen), and also b/c targeting is U.S. only, and I suspect some of the value of the CO is the geo-optimization.

Benchmark is 10 days preceding the period that CO was turned on. CO was turned on for about 19 days, and my CO data is taken for the 17 days in between.

Here are pre-CO data:

Date Impressions Clicks CTR Avg CPC Cost Avg Position Conversions Conversion Rate Cost / Conversion
total 203994 3680 1.80% $1.10 $4,041.66 3.2 411 11.17% $9.83




Here is data from the period when CO was turned on:

Date Impressions Clicks CTR Avg CPC Cost Avg Position Conversions Conversion Rate Cost / Conversion
total 643356 9498 1.48% $1.27 $12,098.42 3.3 826 8.70% $14.65


So, average daily cost/conversion increased by 49% and average daily conversions increased by 18%. Another way to look at it: average daily cost increased 76% and resulted in 18% improvement in average daily conversions. Doesn’t appear to be a solid tradeoff. Small amount of data, but enough to make a decision to turn it off in the case of this client.

[NewKidOnTheBlock]

Isn't this a simple case of statistical significance? If there isn't enough data the confidence interval of the prediction (the range in which the true value is) will be too big to draw any conclusions from it?

What did you mean by stretching out the data range discovery? Using more than 30 days? I think that could be a problem because of seasonality (more people surfing in winter than in summer, Christmas sales, holidays, etc. so maybe that's why they're using 30 days)?

[larryg]

Well, I actually wrote down some numbers to share. For starters, my CTR isn't that much better, I had glanced at the wrong campaign...

I have only been using the CO just over a week so I have a while before I decide how good it really is. However, at this point it is already heading in the right direction and that seems like a decent "learning curve".

For comparison, here are the numbers from the last 7 days vs the 60 days before starting CO...

CTR increased from 5.21% to 5.34%
Avg CPC increased from $0.52 to $0.54
Conversion Rate increased from 1.44% to 1.67%
Cost/Conversion decreased from $36.17 to $32.58 ($41.48 was the recommended bid, currently at $39)

Larry

Update last 7 days...

CTR = 5.33%
Avg CPC = $0.55
Conversion Rate = 1.95%
Cost/Conversion = $28.31
Current bid left alone at $39

Larry

[Discovery]

I decided that since the conversion requirement was dropped to 200 in a month and I had a campaign that was flat for months at just over 200 conversions I had a good test case.

Our Goal was to achieve a CPA of about $28 for this campaign. It had been at about $30 for the past 4 months. Optimizer recommended we set our CPA bid at 31.15, but we went for our target of $28.

Pre-test
Single day:
Clicks: 136
Impressions: 47,997
CTR 0.28%
CR: 5.88%
CPA: $31.30
Conversions: 8

First full day with optimizer
Clicks: 122
Impressions:43,495
CTR: 0.28%
CR: 0.82%
CPA: $182.44
Conversions: 1

So by the looks of the first day of change the best we can hope for is that Google needs a day to do some serious adjusting which crushes performance. Will it then start to get it right and come out of the woods?

I'll give it a few days. A few hundred dollars lost is in the noise if ultimately it achieves the goal. If optimizer pulls even with normal performance then at least I would know it saves my team a few hours a week in management time. This alone is well worth the test cost.

My cynical side says this is another tool for the newbie "set it and forget it" crowd who can't properly bid and not a bidding tool to offload some work from the SEM. It's likely that the cost of Manual tuning by an SEM will be more profitable than this bidding tool. Either that or the use of an advanced 3rd party bidding tool is more appropriate for this role.

Has anyone else taken the plunge and tried out Conversion Optimizer?

Time will tell.

Discovery

[larryg]

Update last 7 days...

CTR = 5.33%
Avg CPC = $0.55
Conversion Rate = 1.95%
Cost/Conversion = $28.31
Current bid left alone at $39

Larry

Here is another update...last 7 days

CTR = 6.11%
Avg CPC = $0.60
Conversion Rate = 2.39%
Cost/Conversion = $25.26
Current bid at $30

FWIW, it seems to be working.

Larry

[Discovery]

Okay, 3 days, I cry for mercy.

Google optimizer was a horrible flop. Bottom line is that our number of clicks stayed about the same, Avg. CPC dropped slightly, but our cost per conversion sky rocketed from $30 to $192. I can make the assumption from this performance and the reports from others that Googles bidding adjustments are probably fine, but their interpretation of where to place those bids is way off.

If a client hired me to increase conversions and keep conversion costs the same and I instead increased CPA 6 fold I would be fired. Optimizer - You're Fired.

I'm a little embarrassed that I even tried it.
I would not recommend that you make my mistake.

The attraction of saving time is a strong one. Will have to focus more on that element in my own marketing!

This brings up a couple interesting questions:

When you turn optimizer on we know it turns off all of your demographic settings. Does it also ignore your current exclude site and category settings?

Obviously something changes in your network make-up. What effects this change?

I would like to see the actual bid data behind my optimization.
What went so horribly wrong.

Why do other bid management tools work where googles does not?

Are there other ways to meet half way with the optimizer?
Review its assumptions and then adjust where the SEM deems necessary.

I noticed recently that Yahoo has the ability to find campaigns and ads with certain performance criteria. Seach by CPA, CPC, CPA, Conversion thresholds and so on and get a list that you can instantly modify - bids and so on. This kind of ability is missing from Googles tool set.

Am I right in assuming the only advanced tool Google has released for advanced SEMs is the adwords editor? Anything web based is not for us?

Discovery

[larryg]

My CPA went up a bit initially, but not 6 fold. It's interesting that it seems to adapt better to certain campaigns than others especially considering it is working off of your historical data.

Just a thought though...would you let a client fire you after just three days running a campaign or would you convince them that it takes longer for everything to settle in?

Larry

[Discovery]

Just a thought though...would you let a client fire you after just three days running a campaign or would you convince them that it takes longer for everything to settle in?

Fair question

Of course I would never let a client fire me at any time.
However, I would expect a severe knee jerk reaction if I increased their CPA 6 fold. Keeping a client happy is all about properly setting expectations, in this case I would have to explain to them upfront that their CPA could increase dramatically for the first week or two, but should then level out... But I am not willing to bet that this will be the case after two weeks, thus I canned it.

Discovery

[AdWordsRep]

Okay, 3 days, I cry for mercy. [...] Anything web based is not for us?

Discovery, I'm passing along your feedback this evening in the Advertiser Feedback Report,verbatim.

That said, I'm sure glad to see a moderate 'success story' in the thread as well. ;)

AWR

[shorebreak]

Discovery, not being a part of the team that created the Conversion Optimizer (and, beyond that, not being particularly adept at statistical science either) I'm not really able to answer that.

And to be even more straightforward, if I did know the particulars I almost certainly wouldn't be able to mention them - given that our competitors surely read this forum. ;)

One thing I have learned over time, though, is that things are rarely as simple as they seem at first glance - whether we're talking about AdWords algos, or pinpointing an electrical problem in an aging car - to name just two.

AWR

:confused:
AWR, I have to say that I have a big problem with a Google employee saying they consider other SEM tools vendors 'competitors', and I think it speaks volumes to what Google's really trying to accomplish with Conversion Optimizer. Incidentally, I heard Doug Raymond, Product Manager at Google for AdWords Editor and AdWords API say the same thing at a conference my company hosted for our 5,000 clients two weeks ago.

Google says "The Conversion Optimizer is an AdWords feature that uses your AdWords Conversion Tracking data to get you more conversions at a lower cost."

Beautiful, and given that Google does no evil, I think it's safe to assume that Google's realized that helping advertisers get more conversions at a lower cost will result in those advertisers reinvesting those saved $$$ into AdWords. Yet...

...yet since even before AdWords was a sparkle in Omid's eye, keyword management tool vendors existed and their value proposition was/is largely the same. So what I have a hard time understanding is - Why does Google view as competitors firms whose founding principle is to accomplish the very same goal Google has set for itself with Conversion Optimizer? Surely you are honest in saying that Google wants for its advertisers more conversions at a lower cost, right? If that's the case, then how can you view as competitors 20+ SEM technology firms, employing several thousand people and helping the creme de la creme of your customer base increase their SEM efficiencies?

Given that C.O.'s main calling card is 'free', should we assume, then, that you view SEM tools vendors as competitors because - as paid-for products - we in your view take marketing budgets away from AdWords that would otherwise be spent there? Or should I assume that you view SEM tools vendors - and advertisers themselves - as incapable of using the match type, geo, network and other targeting options Google provides in advertisers' own best interests?

Taking either of the above two roads in justifying us SEM tools vendors as competitors presents problems to say the least. If it's the former, then what you're really saying is that we're useless knuckleheads, for if we were providing efficiency gains clearly you'd view us as partners. If it's the latter, then there too it would appear Google deems us incapable of leveraging the targeting options you provide to our advertisers' benefit.

I point that out to lead us collectively down a path of realization, whose endpoint is that Conversion Optimizer isn't really about conversion optimization, but rather about Google revenue optimization. In the C.O. FAQ, Google says:

The Conversion Optimizer takes many factors into account, including your ad's conversion history, the keyword's broad match query, the user's location, and the conversion rates of Google's search and content partner sites.

Given that the clear majority of available anecdotes show that C.O. does precisely the opposite of that which it claims [meaning, it gets fewer conversions at higher cost or the same conversions at higher cost, and I'm happy to link to that body of anecdotes if needed], I humbly submit that C.O. merely gives Google license to broaden matching (bad) and increase content spend (in almost all cases 'bad' from a CPA POV).

No wonder it's free, and no wonder you view the SEM tools vendors as competitors. We really do by & large create efficiencies for our clients, and for lack of growth as you had in years past, your short-term grow-revenue-at-all-costs goals are blinding you to your own long-term interests.

Embrace the SEM tools community. Heck, pay your inefficient clients to use our solutions; that is much more likely to bring you the growth you so sorely need than engaging in a path that leads inexorably to advertiser pain, an alienated partner ecosystem and a revenue shortfall cliff that no knob-turning can overcome.
:(

[Marc P]

Shorebreak - Google's competitors in this case are MSFT adCenter and YSM.

[shorebreak]

Shorebreak - Google's competitors in this case are MSFT adCenter and YSM.

Neither of them have *any* footprint in the SEM management tool space, nor are they likely to get a footprint due to the simple reality of SE market share. It's for AdWordsRep to clarify [insert sound of silence here...], but I'm pretty certain he was referring to the companies that actually compete with them in the SEM mgmt space.

[AdWordsRep]

[...]Google's competitors in this case are MSFT adCenter and YSM.

Exactly correct.

[...] It's for AdWordsRep to clarify [insert sound of silence here...], but I'm pretty certain he was referring to the companies that actually compete with them in the SEM mgmt space.

That was meant as a light-hearted comment about our competitors in cost per click advertising.

AWR, I have to say that I have a big problem with a Google employee saying they consider other SEM tools vendors 'competitors', and I think it speaks volumes to what Google's really trying to accomplish with Conversion Optimizer. [...]

This did not even enter my mind.

Honestly, I am astonished at how this comment was interpreted - and then a rather far-reaching conclusion attached about the 'volumes which have been spoken'. With a real feeling of sadness, I am once again reminded of how careful I must be with my choice of words when posting as a representative of Google - while preferring to just be a human being talking to other human beings when posting on this and other forums.

AWR

[shorebreak]

Exactly correct.



That was meant as a light-hearted comment about our competitors in cost per click advertising.



This did not even enter my mind.

Honestly, I am astonished at how this comment was interpreted - and then a rather far-reaching conclusion attached about the 'volumes which have been spoken'. With a real feeling of sadness, I am once again reminded of how careful I must be with my choice of words when posting as a representative of Google - while preferring to just be a human being talking to other human beings when posting on this and other forums.

AWR

ARW - I feel sorry for your chagrin, but you have to agree with me here that given that Y! & MSFT are *nowhere* in terms of having anything like C.O., and given that others (in Sr Dir positions I might add) have publicly named SEM and analytics firms as competitors, the logical conclusion one should draw from those & your statements is that Google has as a goal to be the ones managing the spend as opposed to there being other non-Google systems managing the spend.

When Google has a free analytics product and lists other analytics vendors as competitors, it's clear they view those vendors as competitors, right? And if the SEM tools and analytics vendor spaces are/have consolidated (that's a bona-fied trend/endgame), then isn't it correct for those of us outside the 'Plex to infer that Google views analytics/SEM tools vendors as competitors? And lastly, if that's true - and to my eyes, it certainly appears to be a factual statement - then why in the world does Google feel it necessary to go to such great lengths to obfuscate that truth? Come out and say "We want to be the ones managing PPC because [insert strategic or tactical goal here.]" The industry already thinks this, advertisers already think this, and IMHO, when you try so consistently to come across as a wholly benevolent force in the world of search, it doesn't enhance your credibility.

So you know that I'm not flippant or anti-G by nature, I was on a panel at LeadsCon in Vegas yesterday ('Mastering PPC', ~200 attendees) and I was the only one of 4 panelists who said 'Quality Score' was a good thing, and I was also the only one who defended Google's unwillingness to be more transparent w/Q.S.

I'd hope, then, that you think about why I'm just as astonished that you're astonished. :o

[Discovery]

Update - 1 Year later.

We have recently tested Conversion Optimizer, about a year after our initial run. We had been told in convincing fashion that the tool had changed and was vastly improved since its first launch.

Folks let me save you some time.
The results are the same. Our ads ran in neighborhoods Mike Tyson would be afraid to enter and the conversion results were horrible.

A nearly 400% increase in CPConversion cost. Even after a number of days running at high volume to let Google's optimizer "work the numbers out".

I have yet to meet or even hear of one SEM that has used it with any success.... any here want to chime in?

Discovery

[shorebreak]

Microsoft just announced (http://www.adcentercommunity.com/blogs/analytics/archive/2009/03/12/adcenter-analytics-beta-to-close.aspx) that they're getting out of the free web analytics market.

[shorebreak]

Yesterday I met with a $500K/mo SEM advertiser, who mentioned that Google/DCLK were in their offices this week saying that Google's working on a higher-end SEM tool that will have predictive pre- and post-click modeling capabilities. Then just now I heard from an SEM buddy in Europe who heard essentially the same thing from a Google rep.

The plot thickens...

[Discovery]

500k/month damn what’s he selling Shamwow?
We can't spend our max budget without getting hammered with low quality clicks and fraud. The number of companies with quadruple Google profiles to dominate the first page is killing quality and driving up costs. Yet Google will not put a hault to it.

So this new modeling system takes into account business objectives passed to it from a back end?

Anything short of that is nothing new. A system that optimizes on CTR and slants all ads to run on low quality high CTR sites.
If your ads are not generic enough they will suffer, thus, the more generic the higher the CTR, and the higher they are promoted, again promoting lower quality ads, higher volume, and lower conversion quality.

CTR is a poor determinant of success.

Discovery

[LittleChips]

Hi,

This is my first post to these forums having found this site searching for some independent feedback on CO. Although with representatives from MSN & Google throwing in their comments I'm not too sure who's independent & who's not. Having just read through the whole thread some observations from me (I manage 4 separate Adwords Campaigns with a total budget of about £20k per month).

i) The largest campaign I run knocks out about 500 conversions a month, and the smallest about 100. My observation here would be apparently good results on the larger campaign but not so good on the smaller. More data/conversions = better results on CO in my opinion, which makes sense to me having a statistical background. The minimum number of conversions per month required to use CO at the time iirc was 200 and I think I was on about 300 and I had at least 4 months data. The smaller campaigns have not yielded such good results and I'm up and down like a bride's nightie as to whether to continue using it on these campaigns or not.
ii) The case studies on this forum seem to be comparing performance on a daily basis. Surely any SEM worth his salt and with sufficient data would be well aware of daily fluctuations? I can say that the CPA on the larger campaign is well under target and the conversions (number of) up to target. Although, of course, I don't know what these would have been had I not used the CO, maybe better. Seasonal variations in my sector make it difficult to compare month on month purely on these two variables. But, if I'm within budget and on target each month (last month I achieved this about 14% under budget) then I'm well happy;o)). Although this doesn't mean that I won't be trying to improve on this. And for what it's worth August is traditionally one of the 2 worst months in my sector (December being the other).

I would say that when I first switched on the CO on the larger campaign (which I think was about Feb 2009) I had pretty much got as far as I thought I could by managing the bids manually. So the only purpose of me using the CO would be for it to improve on this because of all of its clever whizzbangs. This appears to have been achieved although, as previously stated, I don;t know what the results would be were I not using it so just base its usefulness on the fact that it's keeping me below budget and the time saving element allows me to spend time on other things.

Just an initial repsonse & don't want to make my first post too long but have loads more feedback so I'll wait to see how this is received or if anyone interested in more of my thoughts before turning this post into a novel!

Steve

[Mel66]

Great first post Steve - very helpful information. I've found essentially the same thing as you - CO works well on large campaigns with lots of history & high number of conversions; not so well on smaller campaigns with fewer conversions & history. It's great for big campaigns that you've already optimized, as you say it allows you to spend time on other things.

I'm up and down like a bride's nightie as to whether to use it on these campaigns or not.


This made me LOL! Again, great post, welcome to SEW and we look forward to hearing more.

[LittleChips]

Hi Mel,

Thanks, my first reply, how exciting! It strikes me that Google have reduced the minimum number of conversions, to 30 now I believe, presumeably either because (i) they want more people to use it (ii) it's vastly improved. I can't see that it would be much use if it's switched on after the first 30 conversions as (insofar as my campaigns go) there simply isn't anough data and at that stage I would be looking at reducung the CPA by at least 50% before letting loose any automated bid management tool. 200-300 a month seems the optimum/minimum to me.

It seems to have a few little quirks and I don't know if this is just coincidence but it seems to do very well, for a while, if I switch it on and off every day (ie revert to manual bidding overnight) and then kick it back in first thing in the morning and accept the, hopefully, reduced recommended CPA.

Of course, this is exactly what Google tell you not to do which is good reason to give it a whirl to see what happens if there's a margin to play with.

Steve

[LittleChips]

<<Our ads ran in neighborhoods Mike Tyson would be afraid to enter>>

Discovery - can you explain what you mean by this because I can't see what you're saying? The CO, as I understand it, manages the bids based on variables such as geo-location, time of day, day of week etc and can only shake it's booty based on the historic data that is already available. I believe it continually updates based on just the previous 30 days data (which is a flaw in my opinion as I'd like it to look at monthly/seasonal variations).

You seem to be describing what I would expect the 'Content Matching' feature to do, which I've never, and will never, use. I can't see that the CO would do what you're suggesting it has done. Are you using the Content Matching feature as well? Assuming you're running a broad match campaign I can't see how the CO would put you in places you didn't know there were places any differently to if you were managing the campiagn manually. But then I'm a bit of a novice so maybe I don't know what I don't know.

Thanks
Steve

[deanpowel71]

Nice post Steve. This information will definitely help us all in many of our future maneuvers.

[LittleChips]

Hi Dean,

Not sure which post you're referring to but nice to be made welcome anyway :)

Rgds
Steve

[LittleChips]

Can anyone think of a good reason why Google don't allow use of the ad scheduling and position placement with the CO? I appreciate that it would perhaps limit the intended automation of CO but this could offer the best of both worlds. For instance I would like to stop my ads hitting the number one spot (and possibly the number 2 slot as well) as I know, in my sector, this gets me more than my fair share of window shoppers. I can see this happen every so often and I get a nice pattern with the CPA decreasing then suddenly a day when it shoots up because i've hit a top position. Also I'd like to switch the ads off at about 10.00pm ( I can do this manually of course but a bit of a nuisance).

PS - on a totally different subject could one of the moderators explain if there are rules (apart from the obvious ones) about posting. I posted something on one of the Bing threads (which I can't even find now) which was quite critical of Bing/Adcenter but was not at all offensive and the message seems to have been removed unless I'm being totally dumb (not beyond the realms of possibility). I can only assume it was removed because it was regarded as over critical of Bing and I've not received any notification of it being removed for XYZ reason and/or what the rules are.

Thanks
Steve

[rodhull]

Can anyone think of a good reason why Google don't allow use of the ad scheduling and position placement with the CO?

Ad scheduling CAN be used with conversion optimiser. I've just tested this myself and it works.

With position placement, you shouldn't need to use this with CO anyway. The whole idea is that Google bids to a target CPA and works out the optimal positions for you. If in position 1 CPA is too high (as you suggest it will be), then Google will see this and aim for lower positions. It may take a small amount of time for the system to figure this out, but stick with it as it should work. Having said that, I myself haven't seen massive improvements when using CO, but the results have certainly been no worse.

[LittleChips]

Ad scheduling CAN be used with conversion optimiser. I've just tested this myself and it works.


This surprised me and so had to go and check myself, and I stand corrected, although I don't think it can be used in advanced mode which makes sense, but I think that was what I had previously looked at.

I take your point about position preferencing and how CO would react but personally would welcome a cross between CO being totally automated and having some control over it. My view that would be getting the best of both worlds.

Thanks
Steve