To block or not to block?

That is the question (MFAs)

[ari11210]

I have one page on my site that brings in most of my AS revenue. I noticed via my AS tracking that the #1 clicked ad on that page is a MFA site. This ad is almost always showing in the #1 slot in the Google ad unit. People have mentioned blocking MFA sites as a way to increase revenue. But, isnt the fact that Google puts this in the #1 spot an indicator that it will bring me the highest price per click? Or is that wrong? If I block it, and Google shows the #2 ad, will that pay less?
Any thoughts?

Thanks
Aaron

[John Carpenter]

Who says they dont convert.

See the post I quoted.

[ari11210]

What I am in wonder about is why the drastic drop if CTR? Are the MFA ads written so much better? Anyone have an opinion on that?
Also, what is the feeling on eBay ads? Should I block those as well?

Anyone have any theories why MFA ads get so many more clicks?
Also, what to do about eBay ads?

Thanks again

[Khensu]

The reason they get more clicks is that they promise an wanton result to the user. "Get MS PowerPoint Free"
When you click the ad they ask for your email so they can spam the daylights out of you or they have a laundry list of Adsense links straight out or woven in and around some stock content.

I had one the other day that said "Free PowerPoint Templates" Oh, they had the free templates but the page before the download was riddled with ads(not that mine arent) and then linked you to the microsoft site to download their generic templates. Crafty to say the least.

Disclaimer: arent is the worse my english will get!

i prms u ppl dat (sick)

[david_uk]

If MFA sites cannot report their AdWords conversions to Google, then your AdSense earnings can't be affected by SmartPricing caused by MFA sites.

The vast majority of advserisers don't use conversion tracking - including me.

If an advertiser chooses to use this facilty or not has no bearing whatsoever on smart pricing. It might be included in the smartpricing algo, but then again it may not be. Google aren't saying. They do use many factors in smartpricing, and many feel that conversion tracking does not play a big part in it. It may be an idea to look at some of EFV's posts on how Google can operate smartpricing without using conversion tracking - he explains it better than most.

The bottom line, blocking ads that link to pure MFA sites can't affect SmartPricing on your site in any way (maybe only if the MFA sites edged out advertisers who do track conversions -- then you could actually earn more thanks to MFA sites).

That can only be true if conversion tracking is the only thing that Google uses in the smartpricing algo. As that is not true, then nor can the above quote be true either.

[John Carpenter]

That can only be true if conversion tracking is the only thing that Google uses in the smartpricing algo. As that is not true

This is nothing but speculation without any evidence. How could Google know whether a person purchased the goods that you sell (or made any other kind of conversion)? The only way for them to find out is you telling them about the purchase (via the conversion tracking code).

If Google were able to find out about the purchase without you telling them, then they'd be better than CIA/NSA/etc. Don't let the conspiracy theories overcome your paranoid mind (sorry).

[edited by: John_Carpenter at 6:07 pm (utc) on Mar. 23, 2006]

[Play_Bach]

Maybe it's just the niche, but I do better NOT blocking. Also, I suspect Google may have begun to set the ad price just high enough for my best paying site that lately I'm only seeing legit advertisers there. Nice.

[hunderdown]

John, with all due respect I don't think you can reasonably call the idea that Smart Pricing isn't based just on conversions a "conspiracy theory."

Since so few advertisers use conversion tracking, it wouldn't make sense for them to only use conversions. And to my mind, Google has confirmed that they don't just use conversions, just by what ASA and others HAVEN'T said--they haven't ever, to the best of my knowledge, stated that conversions are the only factor.....

[david_uk]

This is nothing but speculation without any evidence. How could Google know whether a person purchased the goods that you sell (or made any other kind of conversion)? The only way for them to find out is you telling them about the purchase (via the conversion tracking code).
If Google were able to find out about the purchase without you telling them, then they'd be better than CIA/NSA/etc. Don't let the conspiracy theories overcome your paranoid mind (sorry).

I only get paranoid when the kids come to me saying "Er, Dad.....". :)

As I said, Google use a variety of factors and do not rely on conversion tracking to work smart pricing.

[adwords.google.com...]

We take into account many factors such as what keywords or concepts triggered the ad, as well as the type of site on which the ad was served. For example, a click on an ad for digital cameras on a web page about photography tips may be worth less than a click on the same ad appearing next to a review of digital cameras.

There are a few mentions of this also in the official adsense blog.

[John Carpenter]

a click on an ad for digital cameras on a web page about photography tips may be worth less than a click on the same ad appearing next to a review of digital cameras.

This actually can't affect you, but it can affect the MFA publisher (reduce his AdSense earnings). You have your content and the MFA publisher has his content. If the AdSense bot finds out that the MFA site has no real content, it may reduce the minimum required bid for ads on the MFA site (not on your site) via SmartPricing. Again, this does not affect you, but only the owner of the MFA site.

The point was that the MFA site could "SmartPrice" you only if they reported false conversion stats via the SmartPricing tracking code. But MFA publishers can't report any conversion stats by definition.

It is in Google's best interest to serve ads that are expected to provide the highest eCPM and conversion ratio (if tracked) based on the ad performance history (if available). And, yes, I've heard the opinion that their algorithms are flawed (which I find highly unlikely).

[21_blue]

martinibuster wrote:

I'm not saying that some ads shouldn't be banned. What I am saying is that these so-called tests are so flawed that any reasonable person would reject them out of hand.

martinibuster, I think you have a point about the tests being flawed, but as a (hopefully) "reasonable person" myself I'm not so quick to reject them 'out of hand'.

Whilst the tests fail to provide statistically significant evidence to support the hypothesis that banning MFAs increases earnings, the fact that there are flaws in the test does not provide evidence that not banning MFAs increases earnings.

The only conclusion that we can draw in respect of the flawedness of the tests is that we don't have strong evidence one way or the other. MFAs could increase earnings, depress them, or have no impact at all - we simply don't know with any level of certainty.

So, in the absence of any 'firm' data, as webmasters we have to make decisions on the basis of what's left - ie: inadequate data. Whilst david_uk's and nitrous' studies count as inadequate from a strict statistical point of view, it may actually be the best data that we have to go on at this point in time.

[21_blue]

John Carpenter wrote:

MFA publishers can't report any conversion stats by definition

John, if I were so minded to publish an MFA site, I can think of ways of manipulating conversion stats to help lower my CPC. I don't want to go into details, to avoid giving ideas to those members of the MFA-publisher/WW-reader group who haven't sussed it out. But I suspect there are a lot of MFA webmasters out there who have worked out how to do it.

[John Carpenter]

John, if I were so minded to publish an MFA site, I can think of ways of manipulating conversion stats to help lower my CPC.

You are definitely right. The SmartPricing system is without a doubt open to abuse. I believe Google can detect abuse of their conversion tracking system at least as well as they can detect click fraud. If not, then blocking MFA ads and other ads where such abuse could be involved might really have some merit. (I hope this won't start a "witch hunt".)

[Khensu]

Breaking out the rope and big rock right now.

I here there is a barbeque starter fluid sale at the Home Depot!

(Sorry, I can never pass up an open line)

[Hobbs]

To block or not to block?
That is the question

All the world .. is a web page
And people are mere visitors

sorry tried to resist but failed.
--Hobbs

[david_uk]

So, in the absence of any 'firm' data, as webmasters we have to make decisions on the basis of what's left - ie: inadequate data. Whilst david_uk's and nitrous' studies count as inadequate from a strict statistical point of view, it may actually be the best data that we have to go on at this point in time.

I didn't set out to prove anything financially - I know that the time I could afford to do the test is inadequate to prove anything statistically. I set out to look at what happens to ad selection if a 7 month old, well maintained blocklist was removed. I didn't think I needed long to asses this.

Apart from this, I get the feeling there is a growing tide of publishers here complaining about MFA's - it's not just Nitrous and me :)

[ari11210]

...and to repeat what I said earlier, you can argue with the test method, but all I know is that after a few months of consistent CPCs, I started blocking MFAs 2 days ago and now my CPC has doubled since then (from about 50 cents to over a dollar). It is tough to dispute that it is not due to the blocking of MFAs.

[Thez]

Or am I reading implications that aren't there?

My rep simply implicated that any changes in your pages and/or cpc will be visible usually within 1-2 weeks.

I think this is a common misconception. SmartPricing can only reduce CPC (not increase). If the advertiser does not report her conversion stats to Google, then SmartPricing is not applied.

I don't know where you get this, but Smart Pricing works both ways. High number of conversions increase cpc for the advertiser, and vice versa.

The bottom line, blocking ads that link to pure MFA sites can't affect SmartPricing on your site in any way

What makes you think that? Have you inside knowledge how G's pricing algos work? Because what I believe is that their smart pricing relies purely on click/conversion rate and doesn't track individual campaigns' conversions.

Otherwise, how would you explain significant increase in cpc and overall earnings when people block MFAs even though ctr might drop as much as 40-50%? This also proves that smart pricing works both ways.

[John Carpenter]

I don't know where you get this, but Smart Pricing works both ways. High number of conversions increase cpc for the advertiser, and vice versa.

If you have a link to any page where Google says that SmartPricing can increase CPC, then post it please. As an AdSense publisher, I'd be happy to hear that SmartPricing can also work for us (not just against us).

Based on information provided by Google, it appears that SmartPricing can only reduce CPC (not increase).

Reference: [adwords.google.com...]

Quote:
Google's smart pricing model [...] now, with no change in how you bid, Google may reduce the cost for a click if that better reflects the value it brings to advertisers like you.

What makes you think that?

Pure logic. See my previous posts for substantiation. However, I'm relying on my assumption that Google is able to detect abuse of their conversion tracking system. If they are not, then I was wrong and blocking MFA ads might have some merit in regard to SmartPricing.

[21_blue]

John Carpenter wrote:

I'm relying on my assumption that Google is able to detect abuse of their conversion tracking system.

I think the evidence suggests not: we see MFA ads appearing with high CTR and low CPC. And are there any signs or visible consequences of Google taking action against those who do abuse conversion tracking?

[swa66]

Experiments show that blocking the ads is good for your bottom line dollar amount. You might get less clicks, but the price for the clicks will be significantly higher.

Remember that those who get the click make money out of giving your visitors a second chance to click.

All GOOG needs to do is to make sure that outgoing clicks always bring in less money for the scammer than the incoming click cost them. That way it doesn't pay to buy traffic to sell it back to the one you bought it from. [This latter is also why I call it a scam].

[edited to correct spelling]

[John Carpenter]

I think the evidence suggests not: we see MFA ads appearing with high CTR and low CPC.

Why do you think this is caused by them reporting false conversion stats? Also, MFA advertisers still have to compete with other advertisers in the auction for your site. If a MFA advertiser reported low conversion ratio (falsely or not), then Google would IMO select ads/advertisers that convert better, which would be in Google's best interest. Offering the SmartPricing "discount" to advertisers causes Google to earn less money (they keep about 20% of our EPC, so Google and publishers are in the same boat).

[21_blue]

we see MFA ads appearing with high CTR and low CPC.

Why do you think this is caused by them reporting false conversion stats?

This might only be a semantic point, but I didn't say that MFA conversion stats were 'false'. Falsification is only one method of manipulating the conversion stats.

But to answer the question why I think the low stats are the result of manipulation, it is a conclusion I reach based on the fact that:

1. Why wouldn't they? The wording of some MFA ads shows that they are prepared and intelligent enough to employ dishonest means to attract clicks, and there are dishonest ways to suppress CPC
2. When MFAs appeared on my site(s), the EPC was substantially below the previous norm, and rates returned to that norm immediately after my banning them came into effect.

I can't 'prove' that low MFA rates are the result of conversion manipulation - and high (dishonestly-obtained) CTR will also get a discount - but it is the explanation Occam's Razor suggests.

If a MFA advertiser reported low conversion ratio (falsely or not), then Google would IMO select ads/advertisers that convert better

Is that what Google do, or what you hope they would do? Could you point me to a URL where it says or implies Google do this?

[JohnKelly]

Perhaps we should have a "sticky" thread, where people could post their blocklists?

It would help, since the list of MFA's is ever-growing.

[Nitrous]

But you cant its not allowed!

But...

I could post 2 or 3 hundred for a start.

Or I could post a web page with them on, and host it free. And the URL would be a free webspace one linked from my profile?

[John Carpenter]

I can't 'prove'

First you wrote: "the evidence suggests not". Now you say you can't prove it. There needs to be some consistency.

Is that what Google do

Well, it is in their best (financial) interest. So I suppose it is only reasonable to think they do.

When MFAs appeared on my site(s), the EPC was substantially below the previous norm, and rates returned to that norm immediately after my banning them came into effect.

EPC is not important for publishers -- eCPM is.

[hunderdown]

JohnKelly, it wouldn't help anyone to exchange blocklists. We each need to block the MFAs on our own site(s). The ones showing up on your site might not show up on mine, and vice versa.

[Hobbs]

EPC is not important for publishers -- eCPM is

Neither are important, earnings total is.

But if I get say a consistant daily 10k visitors I would measure daily success by EPC.

If I get a reliable 500 daily clicks a day, I would look at eCPM for how good I am progressing.
--Hobbs

[John Carpenter]

Neither are important, earnings total is.

Earnings depend on traffic. You could earn more money on Mondays than on Sundays, just because you get more traffic on Mondays. To be able to estimate your real success, you need to divide earnings by impressions. That's roughly what eCPM is.

But if I get say a consistant daily 10k visitors I would measure daily success by EPC.

Judging one's success based on his/her EPC is incorrect. If you, for example, had EPC of $10 and I had EPC of $0.5, then I could still earn more even if both of us got constantly the same amount of traffic every day.

eCPM is determined by all the factors: EPC, CTR, and traffic. That's why it is the most important parameter for publishers.

[Hobbs]

Earnings depend on traffic

Earnings depend on many things, of which traffic is among the most important, but there are othr factors involved.

That is why I said if your traffic is fixed, your progress is seen by earnngs and EPC, if your clicks are stable and te traffic is variable (as you said) then eCPM is the one to look at after earnings.

--Hobbs

[John Carpenter]

Earnings depend on traffic

Earnings depend on many things

What a red herring comment. I did not say they depend only on traffic.

That is why I said if your traffic is fixed, your progress is seen by earnngs and EPC,

Traffic is never fixed for any site. The whole internet gets less traffic on weekends. That's why earnings can not be used to measure success of AdSense experiments/tests, etc. EPC can not be used either, because CTR is a variable, not a constant.