Big difference is CTR of two ads which isn't being reflected in CPC

[snoopy1122]

Hi guys,

I've got a query about how Google prices clicks, here is some data on two ads on the exact same keyword (both run at the same time), one ad mentions the price, one doesn't hence the big CTR difference,

Ad 1. (with price in ad copy)

CTR - Av. CPC
14.90% - $0.30

Ad 2. (without price in ad copy)

CTR - Av CPC
24.67% - $0.28

So the question is, given the CTR on ad 2 is 65% higher (24.67% versus 14.90%) than ad 1 why is CPC only 7% lower?

[LucidSW]

CPCs will depend on your QS as well as the QS and bid of the ad below you. As all these variables are very dynamic, you can't make assumptions based on your CTRs. You certainly cannot say that since one ad has a 65% higher rate that your CPC will be 65% lower.

FYI, the basic formula is:

CPC = QS(b) * bid(b) / QS(a)

where QS(b) is QS of ad below you, bid(b) is their bid and QS(a) is your own QS.

You don't say how many clicks for each ad but let's assume you have enough that it is statistically significant. I'm sure your higher CTR ad is a few positions higher which right there means you have different competitors below you most of the time, each with different bids and QS than the other ad.

Also, the first position gets a "discount" for being there. If your second ad is at or near the top position, lower the bid to get a position or two lower. Chances are your CPC will be $0.30 as well.

[snoopy1122]

Hi Lucid,

Both ads have an average rank of 1.

Ad 1. 247 clicks
Ad 2. 304 clicks

[LucidSW]

Since you say both ads show in first position (exactly 1.0? all the time?), it's easy to explain. This supposed that both ads have a better adrank than your next competitor(s). Your QS x bid results in the higher adrank.

However, one must remember that QS is calculated not just for the keyword, where the QS is shown, but the ad-keyword combination. The lower CTR ad MUST have a lower QS, whatever it is, and the cost calculated using it. Since it has a lower QS, your cost is higher.

Let's assume that the next lower competitor is always the same and his bid does not change, say $0.30. If you both have the same QS, they cancel each other and therefore pay his bid, if that is not higher than your own of course. In reality, QS is calculated to many decimal places so having two QS exactly the same is virtually impossible. Now, your better CTR ad would normally have a higher QS thus reducing your costs. In this example, your QS would be 6.7% higher than his (his 7.00 to your 7.47 for example or 9.00 vs 9.60) and that explains the difference.

It seems you have two great ads and your competitors probably have poorer ones probably giving you QS of 10 for both. Would be hard to improve on that almost 25% CTR but if you do, you may be able to shave a few tenths of a cent off that CPC. Can really add up in the long run and stay ahead of competitors. The nice thing when you're in that situation is you can reduce bids and see how a reduced position affects ROI. First place is usually where you want to be but if your conversion rate is the same in second or third and paying a few cents less per click, so much the better.

[snoopy1122]

Hi Lucid,

Thanks for the analysis,

Yes exactly 1st place all the time, the ads are a lot better than the ads in other places (I've tested their ads) and the domain name is an exact match for the keyword.

Are you sure quality score gets calculated on the a per ad basis? Just reading around over the last 24 hours I got the feeling it is done on the basis of keyword?

The thing about all this is, if the CPC is correct then the ad with the lower CTR is the better performer. (that is because for each ad CTR x conversion rate is almost identical). around 0.012. ie the ad with the lower CTR has a higher conversion rate (because price is mentioned) which almost exactly offsets the lower conversion rate on the ad without a price. So the lower CTR ad is getting me the same level of sales with a lower spend.

What I was thinking about this morning, if QS were calculated on the basis of keyword then is the higher CTR ad effectively subsidising the lower CTR ad in terms of quality score and hence average CPC? ie if I stopped running the higher CTR ad would the average cost shown of the lower CTR ad rise? (hope you can follow all that)

Another thing I was wondering,

If I have an ad with a 14% CTR that gets a QS of 10 (say the competition's CTR is 5%), if I then run an ad that has a 24% CTR do I get any further QS boost? (in google's backend) If not then the higher CTR ad can potentially result in a lower ROI like the situation above.

[LucidSW]

> I've tested their ads

You should never do this. You should be different. Besides, doesn't mean you'd get the same results. Why would I click on one ad or another if they are both the same?

> the domain name is an exact match for the keyword

Irrelevant.

> Are you sure quality score gets calculated on the a per ad basis?

Positive. Remember I said the ad-keyword, not just the ad. Yes, QS is shown at keyword level. But think about it logically. Which has the most influence, the keyword or the ad? You can test this out.

> the ad with the lower CTR is the better performer

It can happen but in your case, no. I also don't understand your logic and calculation.

The CTRs have a difference of almost twice the rate. Unless the lower CTR ad has a conversion rate of twice that the first, then they are similar ROI-wise. But you say they are identical. Therefore, you win twice with the higher CTR ad. It gets a higher QS which keeps costs down and competitors at bay. You also get more sales per visitor and therefore more revenues and profits.

> hope you can follow all that

I think I do. There is no subsidizing. There's no such thing in PPC.

You have two active ads. Each ad has a QS related to the keyword in question.

Ad(1) with Kw(1) = QS(1)
Ad(2) with Kw(1) = QS(2)

So what QS is used at the keyword level if each ad has it's own QS (which could be very different) for that keyword? Seems obvious it's an average but the actual calculations for ranking and costs use the QS for that ad-keyword.

Your situation being in first position all the time makes it simpler. Your CPC is the clue: the lower CTR has a higher CPC. Look at the formula. It means your QS for that ad is lower.

If you stopped the higher CTR ad from showing, the CPC for the other ad would be the same. Of course, that's assuming everything else stays the same such as competitors using the same ads and bidding the same.

> if I then run an ad that has a 24% CTR do I get any further QS boost?

I believe you do. You can test that yourself. Is the QS of the 14% ad indeed 10? If so, then because your CPC is lower, the QS of the other ad must be higher. Ten is just the upper display limit and only shows as an integer. The actual QS used in calculations as mentioned has many decimal places and I'm convinced goes higher than ten. Therefore, even if you do achieve a ten, you can still improve. I realized this myself a few years back when a client with mostly QS of 10 wanted to see if I can reduce his costs further. I did with ads having higher CTR. Conclusion: QS in the backend can be higher than ten.

You can post links here at WebmasterWorld but add a dot com to Lucid Web Marketing and check out my free ebook which goes into more details on how QS is calculated. It's basically a standard deviation calculation.

[snoopy1122]

> I've tested their ads

You should never do this. You should be different. Besides, doesn't mean you'd get the same results. Why would I click on one ad or another if they are both the same?

Why not? Is it not the norm in business to copy a competitors marketing methods?

If two ads contain the same text I'd imagine the ad most likely to be clicked on is the more prominant ad.

> the domain name is an exact match for the keyword

Irrelevant.

It results in higher CTR's and hence lower costs.

> Are you sure quality score gets calculated on the a per ad basis?

Positive. Remember I said the ad-keyword, not just the ad. Yes, QS is shown at keyword level. But think about it logically. Which has the most influence, the keyword or the ad? You can test this out.

I know it is would be logical to base it on the ad, but do Google actually do that? Have they said it anywhere?

the ad with the lower CTR is the better performer

It can happen but in your case, no. I also don't understand your logic and calculation.

The ad with the lower CTR has a higher conversion rate (because it mentions the price, approx 70% better convesions). So revenue is much the same with both ads, but the lower CPC ad produces that similar revenue at a lower cost. (have posted some more data below)

The CTRs have a difference of almost twice the rate. Unless the lower CTR ad has a conversion rate of twice that the first, then they are similar ROI-wise. But you say they are identical. Therefore, you win twice with the higher CTR ad. It gets a higher QS which keeps costs down and competitors at bay. You also get more sales per visitor and therefore more revenues and profits.

Yes, the conversion rate difference is much the same as the CTR difference,

Here is the ctr and conversion rates currently

CTR - Conversion Rate - CPC

Ad 1. 14.72% - 8.43% - 30 cents
Ad 2. 24.34% - 4.79% - 28 cents

So the cost per conversion on the lower CTR ad is $3.55 and $5.74 on the higher CTR ad. Hence the lower CTR revenue produces a higher ROI.

You can post links here at WebmasterWorld but add a dot com to Lucid Web Marketing and check out my free ebook which goes into more details on how QS is calculated. It's basically a standard deviation calculation.

Thanks, will check it out.

[LucidSW]

> Why not?

As I said, you want to be different, differentiate yourself. You may be right that the most prominent (ad on top) may be most likely to be clicked. But that means your ad must be more prominent. Then again, if they are next to each other, the tactic may backfire. Readers may think WTH and not click on any or click the second one. The point is, be smarter than your competitor and you have a greater chance at winning.

As for having an exact match domain to the keyword, yes, it may result in higher CTR as the domain will be bolded. The message however is not or should not be in your domain but your ad.

I remember reading something about one of the first thing people see or look at in an ad is the title (it's bigger) and then the domain. Not sure if that's true. I guess if you're Walmart or some other well-known brand it can influence click rate. But brownmenshoes.com?

> Have they said it anywhere?

Not that I can recall. They won't tell us everything and want to keep us guessing and judging from posts in multiple forums, they are doing a good job at doing that. It's not just logic but also observation.

Conversions are normally not 70% difference between ads. I multiply both rates and the higher the better. Your situation is an exception. Yes, Ad 1 despite lower much CTR is the better producer with its much higher conversion rate, but not by that much. Using rough figures of 15%, 25% and 70%, I came up with just a 2% difference. Your actual numbers show a more significant 6%.

I don't think the number of sales so far gives a confidence level high enough to choose one over another. But it's close to getting there.

Your case is the exception because of the much higher conversion. I guess I'd go with Ad 1. But I'd also test more ads, combining the best elements of both into one and testing other variations. That's really the name of the game. You may get a 20% CTR but with a conversion rate of 10% which would be better. Even 7% conversion at 20% CTR would be better. FWIW, I hardly ever use price in the ad but if it works for you.

[snoopy1122]

As I said, you want to be different, differentiate yourself. You may be right that the most prominent (ad on top) may be most likely to be clicked. But that means your ad must be more prominent. Then again, if they are next to each other, the tactic may backfire. Readers may think WTH and not click on any or click the second one. The point is, be smarter than your competitor and you have a greater chance at winning.

Personally I think to beat the competition it is important to now how they operate, ie how good their ads are.

The message however is not or should not be in your domain but your ad.

The domain name *is* part of the ad, whether people use a good domain or a bad domain. If they use a bad domain they are wasting a line (or a part of a line).

Having an exact match domain results in a higher CTR in my limited experience. There is no rules about how things "should" be done, if one domain name performs better than another I'm going to use that domain name.

Not sure if that's true. I guess if you're Walmart or some other well-known brand it can influence click rate. But brownmenshoes.com?

Let me me ask you this; Have you tried it?

Yes, Ad 1 despite lower much CTR is the better producer with its much higher conversion rate, but not by that much.

So far it is significantly cheaper.

Your case is the exception because of the much higher conversion. I guess I'd go with Ad 1. But I'd also test more ads, combining the best elements of both into one and testing other variations. That's really the name of the game. You may get a 20% CTR but with a conversion rate of 10% which would be better. Even 7% conversion at 20% CTR would be better. FWIW, I hardly ever use price in the ad but if it works for you.

Have done a lot of testing, perhaps 100 different ads over the last couple of years. The two ads I am at now are the two best in terms of maximising CTR x conversion rate though obvious a big difference in cost per sale. I think I'm down to fairly incremental changes in terms of balancing CTR and conversion now, well until I can think up some new ideas.

[LucidSW]

> it is important to now how they operate, ie how good their ads are.

Yes, but not copy their ads.

> If they use a bad domain they are wasting a line (or a part of a line).

What you're saying is that Walmart is a bad domain and that they should not use it. Same for Intel, Barns & Noble and thousands of others. That Cosco should buy and use portableairconditioners.com for example because using cosco.com is a waste of a line.

Yes, the domain is part of the ad and could have an effect. But I don't think it *needs* to contains the keyword.

> Let me me ask you this; Have you tried it?

No. None of my clients ever had something like that. Many have related words in their business/domain name such as XYZRoofing.com or BikerHelmets.com (they sell much more than helmets) or even bargainchildrensclothing.com who doesn't bid on "children's clothing". All are doing well.

[snoopy1122]

> it is important to now how they operate, ie how good their ads are.

Yes, but not copy their ads.

I don't quite follow, why not copy their ads? Do you think they have a right over those 4 lines of text or is there something else?

Personally when I look at how companies market they generally copy each other, Apple releases a smartphone, it does well, everyone copies with something that looks much the same and functions much the same. Goto.com thinks up a great way of serving text ads and does well with it, then Google releases something almost exactly identical.

> If they use a bad domain they are wasting a line (or a part of a line).

What you're saying is that Walmart is a bad domain and that they should not use it. Same for Intel, Barns & Noble and thousands of others. That Cosco should buy and use portableairconditioners.com for example because using cosco.com is a waste of a line.

Big brand like that have great domains, they are great because they are well known....famous. They probably get a higher CTR than others because of that (just a guess).

Now I'm not Barnes and Noble or Walmart, I'm unknown, so either I can choose a brand type names that won't mean much much to most visitors or I can choose a domain which represents exactly what I am selling, the site will match exactly what they are after and it will get bolding.

Now it is all well and good for me to argue that and you to say it does not matter, but when I'be tried both methods the exact keyword domain gets a major CTR boost. At the end of the day it is about numbers.

> Let me me ask you this; Have you tried it?

No. None of my clients ever had something like that. Many have related words in their business/domain name such as XYZRoofing.com or BikerHelmets.com (they sell much more than helmets) or even bargainchildrensclothing.com who doesn't bid on "children's clothing". All are doing well.

Maybe you should try it? I don't think you'll be in a position to know until you do.

[LucidSW]

It's not a question of right. It's a question that you need to differentiate yourself. If you simply copy, you are not different. You also are not testing. What makes you think that the competitor's ad is the best way to go?

I don't think exact domains get a major boost as you say. I know the real boost will be mainly from what your ad says, regardless of the domain.

Why don't you build yourself a brand? I think that's much better. When asked "where did you get those fabulous shoes?", would it not be better to say at Hush Puppies dot com than brown men's shoes dot com?

Maybe it's just me but an exact domain name seems very artificial, forced, fishy and thus something to try and avoid. Would you rather want to do business with Lucid Web Marketing dot com or ppc management services dot com? I think I'm building my brand with the first domain, building trust and recognition, not just a domain that is the service I offer and not even the only service offered.

[Channel01]

Also, the first position gets a "discount" for being there.

Huh? I'm almost certain this is not true. In fact, there's been a lot of speculation over the past 5 years or so that you actually need to pay a premium to get the #1, yellow box spot.

[smallcompany]

So the question is, given the CTR on ad 2 is 65% higher (24.67% versus 14.90%) than ad 1 why is CPC only 7% lower?

If you want to keep this technical, and do some testing in order to get more "meat" for your further thinking and possible discussion, you may want to pause the ad with lower CTR and see what happens with the other one in regards of cost and other measurements. You also watch (note down) the QS for given keyword (or keywords). Then, you may want to do the same thing otherwise (pause better performing ad) just for the sake of testing, and again, note down QS and other numbers. Then compare all what you have (1 against 2), plus all to what you already have.

If you want an opinion which is why you came here, I can tell you that Google AdWords' algorithm (oh my, how scientific that sounds) is there to make a profit for Google Inc. This Inc. is very important to keep in mind as many think about nice times when university guys finally did something good which overwhelmingly received support from (knowledgeable) web community. That was the time when they were not driven by profit like today.
In other words, if that beast (algo) decided to make so much money on that particular keyword, you're going to pay it if you want to be there and get the traffic.
The principle of setting the price based on the history of the keyword in overall and forcing you to pay $1 per click even if your ad is the only one there is not a marketing thing, but actually a setup for whatever they planned to earn. Some call it a rip off.

Interestingly, there is some talk going on about "setup" on quite high level:

[webmasterworld.com...]

P.S.
Please don't get me wrong that I dispute Google's paid search concept about comparing ads and favoring them based on results. It is just that I don't trust them, and that I don't believe software can do something like that in a fair way these days.

[LucidSW]

> Also, the first position gets a "discount" for being there.
> Huh? I'm almost certain this is not true. In fact, there's been a
> lot of speculation over the past 5 years or so that you actually
> need to pay a premium to get the #1, yellow box spot.

First, the first position is not necessarily the yellow box. Not every search results in a "yellow boxing" of ads.

Second, I was not talking about being above the SERPs (yellow box). I was saying that my data shows a lower CPC when an ad is in first position, above the SERPs or not. That doesn't prove there is a discount of course but it's strong evidence. And my evidence is, backed by lots of data, that ads in the first position pay less. Not a huge amount, half a cent most times but other times a cent or two less than the second and lower positions.

The algorithm is there to maximize profit for Google, there's no question about it. But most people think of it the wrong way. It's not there to _force_ you to pay $1 or whatever amount. The algo doesn't say "I'll increase the minimum first page bid to make more money". It's all driven by the the advertisers.

The system is trying to make you increase quality, that's what QS is all about. Google would rather you have higher quality than bidding more. Make the calculation yourself: they make more money with a higher quality ad getting a 5% CTR and paying $0.20 then one having a 2% CTR (lower quality) and paying $0.40. The first one makes them more money for each 100 impressions even though they pay less so which do you think the algo chooses.

[RhinoFish]

i assume the relationship between QS and CTR is not linear, but distributive, because of its relevant-to-others nature. once your CTR is in the upper stratosphere, or scraping the bottom of the barrel, the delta of moves / differences within those regions, shouldn't matter.

[Edwin]

There is definitely a positive correlation between an exact match domain and an improved CTR, because we did a widely reported test on this a couple of years ago and reported our findings in a white paper.

The test backed up what other sources had indicated, mainly anecdotally, i.e. that the right domain name really does boost CTR (when you keep all other factors the same i.e. same bids, same keywords, same ad text etc.)

You'll find commentary on our test and of course the test data itself if you Google exact match domain name ppc

[snoopy1122]

If you want to keep this technical, and do some testing in order to get more "meat" for your further thinking and possible discussion, you may want to pause the ad with lower CTR and see what happens with the other one in regards of cost and other measurements. You also watch (note down) the QS for given keyword (or keywords). Then, you may want to do the same thing otherwise (pause better performing ad) just for the sake of testing, and again, note down QS and other numbers. Then compare all what you have (1 against 2), plus all to what you already have.

Yes that will probably be my next step. I'll probably pause the one with the higher CTR since that is the ones that seems to be producing a lower ROI.

[smallcompany]

lower ROI

It looks like your ad with the price mentioned "scares" away people set not to pay that much. That's why your other ad has a higher ROI as people clicking on it are fine with the price.

Also I see your both ads get served closed to equally. What are the settings on your campaign level for how ads are delivered: standard or accelerated?

That may help you get more info through the performance data for your fine tuning and chase for better price per click when about ads.
In other words, if you're on standard now, accelerated may serve the better ad more which may finally reduce price per click more than now. Not a rule but my guess something like that could happen. Then you check your ROI and figure what works better for you.

P.S.
Oh yeah... if you ever think about changing your ads, don;t actually change them, but pause and create new ones. That will always give you a chance for one step back if you screw up your QS shown on a keyword level.

[enKonversations]

The CPC is always depend on CTR. I don't know any formula which is used for calculating CPC on the basis of CTR but i have noticed that if the CTR is higher then the CPC is also higher.

[LucidSW]

Rhino said:
> i assume the relationship between QS and CTR is not linear, but distributive

I had to think about this for a minute. Since QS is derived from a standard deviation calculation, it would be linear. So if you estimate your QS goes up a point when there is a certain change in CTR, you could estimate the CTR needed for any QS. Problem is, QS is shown only as an integer 1 to 10, it's not enough resolution. And although the mean CTR and SD surely don't change that much because of lots of data, I'm sure it does change over longer periods of time, further complicating things.

I believe any increase in CTR is worthwhile. The QS you see is an integer but the one calculated and used has many decimal places. Seeing your QS remain at 7 may make you think it's not worthwhile. But if that means going from 7.18 to 7.56 in the background, that have an effect on your CPC and average position.

> There is definitely a positive correlation between an exact match domain and an improved CTR

I'm sure it does most times. But for what Snoopy was asking and commenting about, it's irrelevant.

> It looks like your ad with the price mentioned "scares" away people

I agree. It also prevents people from seeing your sales message. There's a reason price is always mentioned last. You want to make your pitch and persuade the potential buyer that your offer is better than others before springing the price on them, not make a decision based on price but features and benefits. Test other ads without price.

> The CPC is always depend on CTR.

Indirectly through the QS.

> I don't know any formula which is used for calculating CPC on the basis of CTR

There isn't any. Look at my second post for the formula. There is no direct correlation of "if CTR is X and CPC is Y, then if CTR is doubled, CPC will be half". It doesn't work that way.

> but i have noticed that if the CTR is higher then the CPC is also higher.

You meant to say, "then the CPC will be lower". Not always and there are many factors, each of which is constantly changing, but in general if CTR goes up, CPC comes down.

[snoopy1122]

> It looks like your ad with the price mentioned "scares" away people

I agree. It also prevents people from seeing your sales message. There's a reason price is always mentioned last. You want to make your pitch and persuade the potential buyer that your offer is better than others before springing the price on them, not make a decision based on price but features and benefits. Test other ads without price.

The thing is though, the ad with the price is the one performing better (in terms of lower cost per sale in adwords). You could say the ad with price scares some people away or your could say the price weeds out the people will to spend versus those not willing to spend.

Given both ads produce similar levels of sales I think having the price is just weeding people out who would not have bought anyway if price wasn't mentioned. Those people don't bother clicking the ad and for some (still unknown) reason Google is only charging a bit more per click.

I've priced it so that it is the most expensive source of the product in the market and it is being promoted on the basis of quality, fitness for purpose etc.

[snoopy1122]

Also I see your both ads get served closed to equally. What are the settings on your campaign level for how ads are delivered: standard or accelerated?

That may help you get more info through the performance data for your fine tuning and chase for better price per click when about ads.
In other words, if you're on standard now, accelerated may serve the better ad more which may finally reduce price per click more than now. Not a rule but my guess something like that could happen. Then you check your ROI and figure what works better for you.

It is on standard, the ad is always in the no.1 spot though.

[exmoorbeast]

have u considered Google and Search Partner Network causing the different cpcs?

[LucidSW]

> The thing is though, the ad with the price is the one performing better (in terms of lower cost per sale in adwords)

We were not talking about ROI but CTR. As for CPCs, I thought it was explained.

> I've priced it so that it is the most expensive source of the product in the market and it is being promoted on the basis of quality, fitness for purpose etc.

Just my point. You are preventing people from knowing about quality because many, probably knowing they can get it cheaper, don't click your ad and read your copy. In my mind, you are putting the cart before the horse. Try different ad copies without mentioning price.

[snoopy1122]

Just my point. You are preventing people from knowing about quality because many, probably knowing they can get it cheaper, don't click your ad and read your copy. In my mind, you are putting the cart before the horse. Try different ad copies without mentioning price.

That is what the whole thread is about. One ad has no price, the other has a price. The one with no price generates much the same level of sales but at a higher cost.

[snoopy1122]

have u considered Google and Search Partner Network causing the different cpcs?

Had a look at that just now, the figures are much the same if I just take the Google search traffic. (which is about 85% of traffic)

[snoopy1122]

I've paused the ad with no price today, so I'll see if this change effects the CPC on the remaining ad or the quality score on the keyword.

[LucidSW]

Why expect the CPC to change for the same ad?

You should see a change in QS unless both ads have the same one. I doubt that's the case since one ad is so much superior to the other, unless your QS was 10 to begin with.

[snoopy1122]

Why expect the CPC to change for the same ad?

You should see a change in QS unless both ads have the same one. I doubt that's the case since one ad is so much superior to the other, unless your QS was 10 to begin with.

Because I think the high CTR ad is subsidizing the lower CTR ad in terms of quality score.

Only a few days in but the CPC of the lower CTR ad has jumped quite a lot after suspending the high CTR ad, from 30 cents to 39 cents.