AdRank, Affiliate Marketers and Max CPC

September 18, 2005

Apologies in advance for the length of this post (I haven’t written it yet, but it’s been brewing for a while), but there are some issues about AdRank and affiliate advertisers that neither Google nor the Perry Marshalls of the world nor anyone else I’ve found can (or is willing to) answer, and IMHO, should scare the living hell out of anyone who thinks his/her ads are safely well-positioned on Google. If AWA or anyone else out there can provide some insight here, I and I’m sure tons of others will be very grateful.

I’ve experienced something fairly since the QBA (is that what we’re calling it?) changeover, and strangely enough, it has nothing to do with minimum bids. The institution of the minimum bids had only good effects — we’re actually paying 1-4 cents for traffic that was profitable at 10-15 cents per click. However:

My company advertises using the GoogleCash method, and for those who don’t know, that simply means we shell out our own money to run Adwords campaigns for other companies’ products, and the only compensation we get is our affiliate commissions. If you know that business, you know that ever since January, Google displays only one affiliate ad per display URL — and the lucky ad that shows will be from the account with the highest AdRank for that particular keyword. Easy enough so far, right? Right.

Fast forward to August. In the seven months since January, my company carves out a very strong position for three very profitable keywords — we’ll just call them KW1, KW2 and KW3. By “very strong position,” I mean seven months of keyword history with average CTR of between 15% and 20%. Because of the AdRank we achieved on those 3 keywords, we were always the one affiliate ad that got served up. And because of Google’s rule about showing only one ad per affiliate program, none of our competitors could ever even get an *impression* on these three keywords from January through August. Even maxing out the keyword bid couldn’t displace us. It was, in essence, the Holy Grail of Adwords — owning three very profitable keywords in a position so strong that we were unassailable by competitors. Life’s good, right?

It was — until the QBA switch. On the *very day* the switchover took place, our traffic for all three keywords died. I didn’t even notice it until two days later (it was basically on autopilot due to the past six months of success). I searched on KW1, KW2 and KW3, and sure enough — there was a competitor who was, by all appearances, outranking us on all three keywords. A brand new competitor, who had no chance of even getting an *impression* on these words for the past seven months, is now outranking our company, with our 7 months of 15-20% CTR per keyword, all alone in the blue strip atop Google.

It made absolutely no sense, but there was one way to regain position — Max CPC. Even if our new competitor had a gargantuan bid, we could match it, and then our performance history would be combined with our new monster bid and we’d get our spot back. Right?

Nope. We raised bids to $20. $50. $75. All the way to $100. And we still didn’t get the position back. And I apologize for being repetitive, but let me summarize this just one final time: **Since the QBA switchover, a brand new advertiser can come in and instantly outrank an established advertiser on each of three keywords with 7 months of 15-20% CTR and a $100 Max CPC.** If that makes absolutely no sense to you, then the line forms right behind me.

I did, of course, ask Google to clarify. I suggested one of two things: an internal error in Adrank calculation, or an Adwords employee (one who precisely understands “other relevancy factors” that comprise Quality Score that the rest of us can only guess at) had simply taken his/her proprietary knowledge and set up a shadow account on the side and is cleaning up at our expense. I was assured that neither is true (although I certainly haven’t ruled out no. 2 myself). Google’s response, boiled down, was this: “Sorry you’re frustrated, but yep, you’re outranked now. Don’t ask why, because we can’t talk about other accounts. Just trust us, you’re outranked.”

I trust ‘em, all right — we’re still outranked and not showing for those keywords. But here’s the catch-22 that Google won’t comment on; in an environment where only one ad per URL will show, how can anyone OTHER than the top-ranked ad make progress toward a higher Adrank, if their ads, by definition, cannot be displayed? The natural answer is, “they can’t”, right? But that’s NOT right, and we’re the poster boy for it — we owned the no. 1 spot for 7 months, and one day, bang — we’re gone.

All I want to know is — how? The discussion, then, MUST return to these “other relevancy factors.” Adwords Advisor has mentioned that AdRank is determined as it always has, based on Max CPC. That may be true — it may be “based on Max CPC” — but everything you’ve just read above (and again, I’m sorry for the length of this thing. I just want some answers), some element(s) of the overall AdRank formula have changed significantly. And the changes are significant enough to do major damage. These were three keywords that were netting us about $100-$150 in daily profit — not retire-in-South-France money, mind you, but a very nice piece of change for logging into Adwords once a day. The worst thing about it is, we earned that money by following Google’s instructions to the letter about tight, relevant advertising — and now they’ve swept the rug from under us, and won’t tell us why.

OK, there it is. Anyone experiencing anything similar? Comments? AWA?

[...] in an environment where only one ad per URL will show, how can anyone OTHER than the top-ranked ad make progress toward a higher Adrank, if their ads, by definition, cannot be displayed?

One reason: The misterious “Quality Score” and its “other relevancy factors”.

Nobody really knows what it is, how its calculated - we only know it’s very very important now. And that could be the reason for you loosing your position.

Jan

in an environment where only one ad per URL will show, how can anyone OTHER than the top-ranked ad make progress toward a higher Adrank, if their ads, by definition, cannot be displayed?

of course this was the position for your competitors before. They had no way to beat you, now there obviously is a way. If anyone knows the answer they probably wont be telling.

…should scare the living hell out of anyone who thinks his/her ads are safely well-positioned on Google
I have to say I never think my ads are safe, especially affiliate urls bid to the top like this. They just arent sustainable, and Im not sure they should be sustainable just by bidding high and building clickthrough. At least now it IS possible to dislodge someone. I cant believe Im going to say this, but this maybe the new changes have merit I just wish I could fathom some of the min bids.

Sorry, that probably didnt help did it.

Iain.

slamthunderhide, I can understand your plight.

I, too, had some nice keywords with very good CTR that got hit. And not by more relevant ads - I was bidding on a brand name and these other ads were either generic keyword ads or ads on some other competing brands.

There is obviously some things we don’t know about the new algo and I don’t think we can get the answers here.

Just wanted to let you know that you weren’t alone in experiencing this bump in the road. From the day the minimum bid update went into effect (Aug 19 by our count), our impression counts dropped 65%. The blue bar went bye-bye for a good week on nearly every term we bid on. We think it has something to do with the re-evaluation of terms concerning the minimum bid process. Mind you, we’ve never had to deal with a minimum bid problem since all of our keywords are in a competitive market and at Max CPCs of $4.00 or higher. But I do believe that the re-evaluation of keywords in and of itself effectively reset the metrics by which Quality Score and Ad Relevancy were determined. Since then (after about a week and a half) things have returned to normal. We did manage to expedite our issue with Google, but then again we also spend about $150,000 a month with them so we have more than a foot in the door to work with. I think that the new evaluation of keywords is taking place one industry at a time (ours is travel) in order to really lock down what is relevant and what is not commercially, at a manual review level, so I wish you good luck and maybe it will be over for you soon.

slamthunderhide

Seeing the same thing here. High ctr history and testing with an outrageous bid is not winning the affiliate auction. Tried everything. It’s almost as though a google employee is setting their own CTR history. Last time meeting with our merchant one of the other affiliates there was a Google employee, and an Adwords employee at that. How is it acceptable to have to compete with someone with insider knowledge that actually works with the system daily and can look at my account. IMHO, Google should really not be allowing this.

AWA. Care to comment on this appearant major conflict of interest at the plex?

Bttmfeed:

Wait a minute — you’ve met with your merchant alongside an affiliate competitor who admits to being a Google Adwords employee? In my case, I just had some strong suspicions, but I’m floored that someone would just come out and admit it (if I interpreted correctly what you’re saying).

Yes, if that’s the case, I would say that’s the ultimate act of treason against Google’s customers. Would love to see what AWA has to say, but I’m not optimistic that he or anyone else from Google will comment.

Wait a minute — you’ve met with your merchant alongside an affiliate competitor who admits to being a Google Adwords employee?

That’s exactly what happened. I was shocked by it as well. Appearantly this Adwords employee has Google’s blessing to have an Adwords account and compete with other affiliates.

Did the employee explicitly say that he had Google’s blessing? And did he also freely admit that he uses Adwords to advertise the merchant? I still can’t believe he came right out and said so.

Most people wouldn’t be so brazen; a potentially much bigger problem are “shadow accounts” set up by a friend, girlfriend, Uncle Bob, or anyone else without an obvious paper trail to a Google employee. Without bttmfeed’s face-to-face experience, this would just be conspiracy-theory talk that’d be easy to brush off. But really, what about the savvier Google employee who cuts a deal with Aunt Edna — “Let me use your credit card and your computer, and I’ll use my inside knowledge to set up dominating affiliate-ad campaigns that will neutralize all competitors and make us a boatload of money.”

Generally I hate conspiracy theories without something to back them up, but your experience indicates that this isn’t just an active imagination talking. If Google is serious about the whole “Do No Evil” thing… what is its plan to deal with situations like these?

I’m sorry, but if you work for Google you should not be allowed to have an adwords or adsense account due to the clear conflict of interest.

JKelly,

Yes, there’s actually more to our story that I didn’t get to because my original post was already so long. We left our $100 bids up for about 5 days, and then all of a sudden, we had our position back — at between $2 and $4 per click, as opposed to 5-7 cents per click before. And these aren’t low volume keywords — about 1,000 clicks daily, so you can do the mat on how that worked out against $150 of daily profit before the switch. I recommend you drop your daily maximum to something that won’t devastate you should those big CPCs kick in.

We ended up dropping bids all the way down to $5 and holding position, and paying 4-5 cents per click. Hooray. Happy days were here again. But about three days later, position was gone again. Raise the bids to $100 again. Position comes back — this time at $3-$5 per click. Ouch. Then, things get really interesting…

We begin dropping bids in $10 increments to see what happens to our rank. When we get down to about $25/click, we lose rank. At $30, we get it back — so there’s the approximate “tipping point,” if you will. But here’s the funny thing — 10 minutes later, with no bid modification at all, we lose the position. Up the bids to $35, we get it back. Another 15 minutes later, we lose it again.

This, of course, makes me think that our competitor is using automated bid-management software like ATLAs One Point — nothing wrong with that on its face, although it does raise an eyebrow toward Google at the remarkable sophistication of this brand-newbie advertiser. Here’s a question for the board — does ATLAs One Point or any other bid management software out there allow you to dynamically scale down your bid based on your most recent CPC? For example, does it allow you to formulate something like this:

“When my average position CPC for Keyword1 is greater than X cents, lower my bid by Y cents”

Or is it possible using the API to a) query whether your ad is showing via the Diagnostic Tool, and then b) raise your bid based on whether it’s showing or not? I am almost completely ignorant about how the API works and what the possibilities are.

I recommend you drop your daily maximum to something that won’t devastate you should those big CPCs kick in

Yeah, I have tried that. The result in my case is usually the ad then gets no impressions. In cases where it continues to get some impressions, they are way less than what it got at $100 max cpc and eventually dwindle to practically nothing. What I don’t understand is why the tipping point as you call it is so far away from the CPC people are actually paying.

About the Tipping Point price — well, it’s just one of who-knows-how-many elements in an algorithm that tips the scales in our favor. Previously, there were all of two factors involved — your Max CPC bid and CTR.

That’s the real story that the tech press, let alone the mainstream press, isn’t picking up on — that the “other relevancy factors” that Google’s been trying so hard to quietly gloss over represent a major change in its relationship with its customers. In essence, they’ve eliminated any transparency or accountability to us that the previous system of (MaxCPC x CTR = AdRank) fostered, and simply closed the process off. Which, as with most processes that were once open to the public and are now hidden from the public, opens the door wide open to corruption.

This also seems like the right time to point out the absurd catch-22 of the new system, illustrated here by my hypothetical conversation with Google:

Client: It’s now a great Quality Score that Google wants to reward us for, right?
Google: Right!
Client: So how do we improve our Quality Score?
Google: Sorry, can’t tell you.
Client: Pardon?
Google: Can’t tell you.
Client: Isn’t this whole thing about improving the Google user experience?
Google: Yes, absolutely!
Client: And you want to reward us for improving your users’ experience, correct?
Google: You bet!
Client: But you won’t tell us how?
Google. Sorry. Can’t do it.

It’d almost be comical if there weren’t billions of dollars at stake here, no?

Well, I finally had time to settle in and read this thread. I have to say that I never got through even the first post till now. It was a little daunting to me, as someone who is posting in-between meetings, and so forth.

slamthunderhide, I don’t have an answer for you, but I will ask around with a few colleagues to see if I can gain any additional insight. I doubt that I will be able to add anything of note, since I can’t see your account, and since you have already consulted with a team who can. But since you seem to be speaking primarily about top placement, I can think of one thing that would pull you down from the top spots in a hot second: editing your ad, even in the most minor way.

Any edit to an ad effectively creates a new ad that must be reviewed and approved before it go back up ‘into the blue’.

All that said, I must admit that the sub-topic to this thread that really caught my eye was the one regarding Googlers advertising on AdWords.

.. Would love to see what AWA has to say, but I’m not optimistic that he or anyone else from Google will comment.

I’d be happy to comment, actually, and I speak as one who has made no secret of my own 3 year old account - having mentioned it on this Forum many times in the past.

By policy, AdWords employees may certainly have AdWords accounts. Please rest assured, however, that they are thoroughly monitored and governed by a list of requirements as long as your arm - designed to ensure no conflict of interest.

It is perhaps worth noting that everyone concerned with AdWords, whether an engineer working behind the scenes, or the person that answers your email, is actually encouraged to have a small AdWords account. The purpose of this is to make sure that everyone who works on AdWords knows exactly what it feels like to be an AdWords advertiser.

Yes, I’ve had my ads disapproved. Yes, I’ve had disabled keywords in the past, and even had my account slowed way back when. Yes, I’ve raised my budget to experiment, and then forgotten to lower it, and yes, I’ve paid more that I expected as a result. And on and on.

And, yes, all of these things irritated me. The good thing about all of this, though, is that I really know how it feels to advertise with AdWords.

Bottom line: Many Googlers have AdWords account. IMO, It’s a good thing, and it certainly is no secret. I can understand why this is concern to some of you, and I can also say with a great deal of confidence that you may rest easy.

AWA