Potential Impact of an Ambiguous Ad.

...with high CTR and low conversion rate

[21_blue]

I've seen an ad appearing on my site regularly that seems to be 'doing well', because it appears in the top spot on certain popular but low-paying pages, and has become the sole ad in expanded versions of 2-ad blocks.

In fact, the wording of the ad, which is true, will be very attractive to visitors to my sites. Eg: "we are the sole supplier of blue widgets". Hence, I think it probably attracts a high CTR. Their landing page has plenty of words/content so I suspect they have a high Adwords quality score and are not paying a lot for their clicks.

However, I also think their ad will attract a low conversion rate. I know the company well and it is, in effect, a trade wholesaler, only providing services to the trade and not end-users. Most of the visitors to my site are end-users, not trade. So, although many may click through to their site, very, very few are likely to proceed to register or purchase from this advertiser.

I really don't like banning ads (our competitive filter is empty) because more ads usually create more competition and higher prices. However, in this case my concern is that the high CTR and low conversion rate will be damaging our smartprice and displacing higher paying ads. I've therefore banned the advertiser.

Does this rationale make sense, and is banning the advertiser the right course of action?

Also, does anyone know if the Adwords or Adsense algos do anything about the situation when there is an ad with a high CTR but low conversion rate? It seems to me that such an ad has the potential to really deflate publisher earnings.

[malachite]

Sounds like the advertiser hasn't fully thought out his ad wording, and if you *know* your visitors aren't likely to purchase from them, it makes sense to me to filter the ad.

If I were the advertiser, I wouldn't want to pay for lots of non-converting clicks ;)

<added> and I wouldn't be too chuffed with you for advertising trade only ads!</added>

[21_blue]

malachite wrote:

If I were the advertiser, I wouldn't want to pay for lots of non-converting clicks ;)

This is a big marketing spender. If they get one new client for every 1,000 clicks, they'll be more than happy. Also, although I'm pretty sure they would track conversions, so they may be happy even with no conversions - they are not so interested in making sales as using Adwords to support their branding promotion, which is what I suspect this is.

[hunderdown]

21_blue, to me the question would be, what counts as a conversion? As I understand it, a conversion is not necessarily a sale. Different advertisers can define conversions in different ways, and the trouble is that you don't know how this advertiser is defining a conversion.

I wouldn't block this advertiser unless you somehow find out more.

[jomaxx]

My first reaction when I saw this thread was that you should just ignore it. The risk of an adverse reaction seems pretty slight.

BUT if you are certain that your site visitors will have no interest in what the advertiser has to offer, and in fact might feel the ad is a bit misleading, then you probably should block the ads. Leaving the ads in place simply wastes the advertiser's money and your visitors' time. That's not a winning combination.

[21_blue]

hunderdown wrote:

21_blue, to me the question would be, what counts as a conversion?

Thanks for the various inputs and ideas. It's proving helpful in thinking this through.

There is one more fact, Jomaxx, I should have mentioned that is relevant. Our sites are diverse, and this particular advertiser is appearing on approximately 50% of our pages which account for only 5% of our revenue. That's not an issue in itself, because I know that this particular area of widgets is very low paying. However, with that proportion of impressions the ads could have a big impact on smartprice in other , more profitable areas.

I don't know how they classify a conversion. But as someone who now tracks my own adwords conversions and knows this company well, I can guess at how they would track it. And whichever method I think of, I anticipate their conversion rate being very low but them being happy about it. In fact, I have one adwords channel myself with a low conversion rate that I'm pleased with, for various reasons.

With regards 'finding out more', I'm always in favour of that, though what is there I could find out, do you think? The only evidence I have is that overall average EPC has been declining in the few weeks that I've seen these ads, though that could be a coincidence and due to other factors.

[moTi]

since i am often undecided what to do with the "high ctr low conversion" ads i turn some of them on and off some time.
last week, i have disabled the ebay-affiliate fraction once again. they advertise like crazy on my sites to earn some cents. for example i have a category like "backview" and they simply imitate this keyword like "buy backview now" or "find more about backview" - total crap, you know what i mean. so, what i recognized the next days was, that my ctr dropped by about one third, epc unfortunately steady (wait a week?).

another example. in the german speaking area, there is an omnipresent advertiser since a few weeks, who pretends to give away free sms. in fact (as i see it), people leave their personal information on the landing page for a costly subscription. you can cancel your subscription within two weeks or so and there you have it, your free sms.
this companys ad consists of only two(!) words. seems to be highly effective, eye-catching and lucrative. some other me-too-people have jumped on board in the last few days and copy him.

what i wanted to say is, this kind of ads drive me crazy. i know that they must have an incredibly high ctr because of their appearance, but i suspect, that they trigger smartpricing and drive my epc way down. in addition i actually don't like my users to be fooled or my site being related to this kind of businesses.
haven't come to a conclusion yet. and in this case there's no "leave it as is, the algo will cope with it". i'm having a hard time to kick out "popular" ads. if only smartpricing algo was smart enough to deal with these ads which obviously suck in publishers income one way or another..

[21_blue]

moTi wrote:

there is an omnipresent advertiser since a few weeks, who pretends to give away free sms

Well, you could knock me down with a flimsy feather. Thinking about it, this is such a "clever" spam, and I could see it doing great damage to your earnings. The appeal of the ad would:

• enable the ad to have a very low CPC because of the high CTR rate
  
• make the advertiser money because these types of 'you can cancel' expensive signups tend to work with a small proportion of people
  
• depress your smartprice terribly, because of a low conversion rate
  

This wouldn't be a problem, however, if 'conversion rate' was something that google took into account in the quality score. That is, there are four categories of ads:

1. Low CTR and Low Conversion
   
2. Low CTR and High Conversion
   
3. High CTR and Low Conversion
   
4. High CTR and High Conversion
   

Ads of Type 1 and 3 damage publisher income, because of the depressing effect on smartprice. Ads of Type 1 and 2 damage both publisher income and other advertisers, because they displace other ads that may be better performing (as measure by CTR).

Therefore, the type of ad that should be "rewarded" is type 4: high CTR and high conversion. However, as I understand it, the current Google algo reward ads of both type 3 and 4, particularly type 3 (because there is an extra discount for low-converting ads).

Is this understand right? If it is, it suggests that an important change needs to be made to the smartprice algo.

[Hobbs]

I understand your positions 21_blue & moTi, but excuse me for taking it to a simpler level, maybe your answer is there:

a) On SmartPricing: When and how exactly can it be in the advertiser's interest to report back a high conversion for any site?

b) What does Google gain from repeating that we cannot increase earning by blocking sites that do not compete with our own?

[21_blue]

Hobbs wrote:

On SmartPricing: When and how exactly can it be in the advertiser's interest to report back a high conversion for any site?

Currently, it isn't. It is in the advertiser's interest to report a low conversion rate, because that means they get their clicks cheaper. My concern is that it also happens to reduce publisher sites' smartprices.

What does Google gain from repeating that we cannot increase earning by blocking sites that do not compete with our own?

I suspect they genuinely believe this is the case. But does the intent work out in practice?

In particular, have they looked at the smartpricing implications of high CTR/low conversion ad?

Also, what assumptions are they using? Does their model assume, for example, that all advertisers want a high conversion rate for reasons that override the benefit of cheaper clicks. I can think of plenty of exceptions to that - eg:

1. when the campaign is brand promotion rather than sales;
   
2. when the site is MFA trying to get as many cheap clicks as possible;
   
3. when the conversion tracking is simply for informational purposes (how many people make it across to the other side of our site?);
   
4. when the site has other products or pages that aren't tracked (ie it doesn't matter if the visitor converts to a particular 'page view', because they may surfer through to other pages)
   

We've recently started using conversion tracking on a particular Adwords campaign for a newish site. On one of our channels we get a low conversion rate. This isn't surprising, because it enters the site further away from the tracked pages than the other channels. We tracked it just for interest - to see how many visitors found their way to the particular pages we are concentrating on at the moment. But the unexpected upside is that our PPC is lower on this channel. That means I'm getting cheap visitors who are, in reality, equally valuable to me despite the lower reported conversion rate to Google.

I can see from this experience how Adwords could be manipulated - using attractive ad wording, such as the type moTI cited - to get a high CTR/low conversion rate and buy lots of cheap visitors. The downside, of course, is that the smartprice of sites publishing those ads would take a hit.

[Hobbs]

>>to get a high CTR/low conversion rate and buy lots of cheap visitors.

Isn't CPM designed to do exactly that (branding too)? Why would the advertiser opt to pay per click when they can pay per 1000 views and get the clicks with a good ad copy?

Ok, so SmartPricing has no built in benefit for an advertiser to report any conversion, but rather its an opportunity for advertisers to scam publishers who are sailing blind with no per advertiser data, how come Google designed such a flawed system?

>>I suspect they genuinely believe this is the case

just follow the money.

[21_blue]

Hobbs wrote:

>>to get a high CTR/low conversion rate and buy lots of cheap visitors.
Isn't CPM designed to do exactly that (branding too)?

No. CPM is designed to control where on the content network your ads appear. This is more about raising the quality of clicks, or pursuing just one particular branding strategy, and could, in fact, result in a higher cost per visitor.

Why would the advertiser opt to pay per click when they can pay per 1000 views and get the clicks with a good ad copy?

Because PPC ads appear in a much, much wider range of locations than CPM ads.

how come Google designed such a flawed system?

They haven't designed it to be flawed. But smartpricing has made it a complex system, and this is perhaps an unintended consequence.

just follow the money.

Hmm... It's interesting thinking that through. The net effect of this on money is:

• The advertiser gets more clicks for their budget
  
• Google's revenue goes up because they get a bigger share of the advertiser's budget - because search clicks are not discounted by conversion algos
  
• The publisher gets less revenue because of the smartprice hit
  

[Hobbs]

I really need to signup to Adwords to understand all this better, just been delaying it as developing content has been my number one priority and I don't need one more distraction.

I love simplifying, so allow me to throw in an analogy:

Google is a restaurant
We are catering companies
Advertisers are the diners

This restaurant does not have a kitchen, they get the meals from caterers and split their earnings with those caterers at the end of the month.
This restaurant does not charge diners for what they order, but asks them on their way out to pay whatever they think their meal was worth.
The restaurant is booked years in advance by happy diners that can pay less no questions asked.
Caterers are confused by this system but consider it a blessing due to the bulk of their sales.
Diners are in consumer heaven and keep coming back.

This model appears to be a happy solution for all 3 parties, but in reality it is not fair to the caterers, only new and smaller caterers will be thankful for bulk price sales, the publishers will move on as they grow, the advertisers will hop around and go pollinate another garden and Google will be left standing in a publisher's kindergarten.

I have no proposed solution, but maybe if Google could come around and give us cases where SmartPricing actually benefited the publishers it would help, Google has repeatedly stated that they are making less money from SmartPricing, what they left out is that they make less money / publisher, and it is made up with more business as 21_blue said, more advertisers, bigger budgets / advertiser, and there does not seem to be any shortage in publishers any time soon.

[21_blue]

hobbs wrote:

I love simplifying, so allow me to throw in an analogy:

Google is a restaurant
We are catering companies
Advertisers are the diners

I agree with simplifying, where reasonable to do so, and I like your analogy. In those terms, what's happening is:

1. There is a regular client who consistently undervalues their meals. Not only that, but they are going round telling all the other customers to undervalue their meals as well. As a result, the catering company's income, for all clients, is less than it should be, or less than a truly fair value. This is analogous to the smartpricing impact
   
   
2. There is another regular client who has booked most of the tables, is selling them on to other clients at $100 a time, but then only valuing the meals at $1 a time. This is analogous to how MFA/spam websites can exploit this particular aspect of the smartpricing to make profit for themselves at the expense of the original content publisher
   

In those circumstances do you say "c'est la vie", allow those diners to depress income for the catering company, and just take whatever happens to come your way?

Or do you say, let's recognise the loophole and do something constructive to try and close it? In the short term, that might involve barring those diners; in the long term, it would involve changing the rules of the game.

[danimal]

21_blue, you started this thread with some questions, but most of your posts have been to hand out advice... meanwhile, you've never ever put a garbage url in the filter(!) well, until now anyway.

so it's an enigma wrapped in a riddle :-) i'll hazard a guess that all of your sites are totally untargeted, something like news sites? you could only be in a sector that didn't care about targeting, in which case, yahoo ads will pay more than adsense.

i have lately seen a trend where the garbage adsense mfa's and such are showing up only on pages that have the lowest ctr and lowest ecpm on my sites... i have started replacing that adsense trash with yahoo, because it pays more.

[21_blue]

danimal wrote:

i'll hazard a guess that all of your sites are totally untargeted, something like news sites? you could only be in a sector that didn't care about targeting

It's nice of you to try and guess, but you're not even close, sadly. And even if our sites (& ads) weren't highly targeted, we're outside the US so YPN isn't an option for us at present.

The problem with the 'ambiguous' ad (now banned) is that, to the visitor, it seems to hit the bulls-eye as far as this niche is concerned.

With regards our not banning 'garbage ads', I guess we've probably managed to maintain a high enough CTR for proper ads to avoid being afflicted by them.

[Hobbs]

but most of your posts have been to hand out advice

That would be my fault, and one of the best exchanges I had on SmartPricing, which affects every one in this forum and is on topic.

[21_blue]

hobbs, as I indicated earlier in the thread, I've found yours and others' posts very helpful in thinking these issues through. I don't think you should be feeling that something has gone wrong and it is your fault. It has been an interesting and productive discussion.

Danimal is right to some extent, in that there has been a shift in my mind, having started out 'asking a question', but then coming to firmer conclusions having realised the implications of what is being said in the discussion.

And the major shift - which you may or may not agree with - is over the potential benefits of banning certain advertisers. Up until now we've operated a strong 'no bans' policy, believing in the ability of the Google algos to get the best price for the publisher. I've advocated that approach to others in the past, and Jensense had made the point in her blog as well.

But moTi's experience in particular - a more extreme version of mine - illustrates the potential for an ad to depress earnings, with the double-whammy of taking visitors away with cheap clicks (due to high CTR) and suppressing smart-price (due to low conversion). Hence I've shifted from questioning to more overt concern.

As I've said in an earlier post, I have banned the advertiser now and will report back to the forum on any change in earnings in 2 or 3 weeks.

[danimal]

>>>With regards our not banning 'garbage ads', I guess we've probably managed to maintain a high enough CTR for proper ads to avoid being afflicted by them.<<<

i'll lay some info on you that i guess you haven't experienced yet.

i created a computer software/internet tech type of website several months ago, and it brought in some wonderfully targeted ads, from some of the biggest companies in the world.

then all of a sudden, with no change on my part, google turned it into a toilet bowl for irrelevant mfa trash ads... garbage that was ruining the rep of the site, and not paying well either, because nobody clicked on the ads.

i raised hell with adsense support, and nearly filled up the url filter to get rid of the trash... many times it was the same advertiser, creating throwaway websites on the same irrelevant topics... i got no help from adsense support, so i went to ypn, because it still, even today, pays better.

so now i am seeing an influx of the same type of trash ads into my main sector... and 10 days ago, earnings suddenly dropped by 50%, and have not recovered... it's a repeat of what google did to my computer site.

the point is that i haven't changed anything in weeks, but google sure has... if you think that everything is going to remain the same forever with google, you are in for a big surprise.

[21_blue]

danimal wrote:

i'll lay some info on you that i guess you haven't experienced yet.

Sorry, wrong again. (Maybe 3rd time lucky with your next guess?).

Having experienced the type of problem you describe I got to the bottom of it, and found the cause to be 'poison words' causing the site to be flooded with cheap ads. I wrote about it here, sparked by someone else having noticed the same thing:

[webmasterworld.com...]

I agree that things can go belly up at times. One has to watch out for it, and try and remedy it when it does.

[danimal]

>>>Having experienced the type of problem you describe I got to the bottom of it<<<

no you haven't, and no, you didn't... i'd suggest that you keep re-reading my post until the point of it sinks in for you.

here, i'll try to help you understand... i had two completely different sites on completely different topics... they were working just fine, with the main sector running adsense for over 9 months... no correlation between 'em at all... when all of a sudden, garbage mfa adsense started showing up.

what are "poison words", btw? and how is that weird concept relevant to a site that didn't change either it's coding or it's text?

[21_blue]

danimal wrote:

here, i'll try to help you understand...

Ah, I think you've hit the nail on the head. My tiny brain is unable to grasp concepts that more masterful intellects spend time thinking about. Ah well, time for bed said Zebedee.

[danimal]

after 10 days in the crapper, today everything is back on track... it's up all across the board; ctr, ecpm, and the $$$... the latter could be helped by a new advertiser in the sector?

the big news, however, is that when i provided a screen capture of one of the totally irrelevant garbage mfa ads, google adsense support admitted that there could be a rare instance of "less-targeted" ads appearing every now and then, as the algorithms are being constantly adjusted.

so it wasn't smart pricing, poison keywords, or any of the other mysto brainstorming in this thread... it was 100% a google screwup, which is what i've been saying all along.

of course, had i been foolish enuf to handle the situation like 21_blue and jensense advocated, i would not have known when an untargeted ad actually appeared... because the only way to get really good ad targeting is to keep putting adsense garbage into the url filter... you have to monitor your ads all the time, so you know who the good advertisers in your sector really are... it's called "standards"... magazines limit who qualifies to advertise in their publications, and so should you.

among other things, that means dumping the pond scum adsense that doesn't list a telephone number or street address on their website.

we'll see if this "fix" maintains itself, but i am hoping for the best... just remember to get yourself a screengrab program, they can be downloaded for free... if you can't provide proof, adsense support will gaff you off, even when you tell 'em the exact bogus url that showed up on your site... screen grabs are the only proof they will accept.

[andye]

Diners are in consumer heaven and keep coming back

Don't you believe it. Us 'diners' (advertisers) keep coming back because there's only really two restaurants in town. To overextend an analogy. ;)

best, a.

[Hobbs]

> To overextend an analogy

or the only two fast food chains, there are still great individual gourmet publishers if you walk the extra mile.

[21_blue]

I wrote:

I have banned the advertiser now and will report back to the forum on any change in earnings in 2 or 3 weeks.

For those who are interested, here is my report back (testing a theory that a high CTR/low conversion ad will be suppressing account-wide smartprice).

Since banning the advertiser, my earnings have increased. However, the increase does not seem to be for the reasons expected.

The overall EPC on my account has been stable and, if anything, has dropped slightly. The increase is not due, therefore, to an improvement in account-wide Smartprice.

I've also looked looked at the EPC on the particular pages where this ad appeared - to see if it might have impacted a page-specific component of smartprice. The EPC on these pages has, in fact, dropped. This makes sense if smartpricing is not coming in to play, because banning a high bidder will result in lower bids winning the space.

In reviewing a timeline of my site stats, the increase in earnings seems to coincide with what we thought was a very minor change, aimed at increasing site stickability without losing clicks. As a result, the number of clicks per visitor seems to have increased substantially (by about 10%).

I'm therefore going to un-ban this advertiser, not make any other changes and see what happens. Again, if anyone is interested, I'll report back in 2 or 3 weeks.