Google Quality Score - It Could Be a Bad Thing . . .

As reported in Search Engine Roundtable, Google recently announced that they are ready to start spidering AdWords landing pages to assess "quality score." In Google's own words: "If a landing page has informative content related to its AdWords ads and keywords, these keywords will receive higher Quality Scores and potentially lower minimum cost-per-click bid (CPC bid) requirements. Poor quality landing pages or those that restrict visits by our system are likely to experience a decrease in quality scores (and a potential increase in CPC bid requirements)." This is actually something Google announced about six months ago. Apparently they are now ready to actually start implementing this system.

My sense is that this is potentially innocuous, or potentially very dangerous, depending on the path Google is taking with this.

The "Do No Evil" Path

Path one can basically be described as the "stop AdWords arbitrage" path. In this scenario, Google is trying to crack down on companies that buy low-priced CPC ads on Google, only to send users to an AdSense page with higher priced ads on them, effectively playing the margins between high and low-priced keywords. Google doesn't like this for two reasons: first, it's bad user experience. If a user clicks on an ad, the user expects to be sent to a relevant Web site. If users repeatedly have bad experiences clicking on ads, they will eventually stop clicking on ads altogether, thereby reducing Google's revenue.

Second, AdWords arbitrage is also bad advertiser experience. Oftentimes, companies that engage in this practice do everything they can to conceal the fact that links are advertisements to the end user. Since the model can't work without a high clickthrough-rate after the initial click, the company needs to make sure that almost everyone who comes in the door clicks on an ad. As a result, I'm pretty confident that the actual quality of clicks on these AdSense keywords is very low. Over time, this will reduce the amount that advertisers are willing to pay for AdSense keywords.

When you combine blatantly bad user experience with equally awful advertiser experience, it's hard to fault Google for cracking down on AdWords arbitrage. And if this is the point of quality score, I support the initiative.

The "Do Evil" Path

There is, however, a scarier - Big Brother - scenario. In "path two", Google decides to play God. The focus is no longer AdWords arbitrage sites, but any site that Google concludes may not be relevant (defined exclusively by Google). Consider this scenario: you have created a site dedicated to comparing online flower vendors (ProFlowers, 1800Flowers, etc). Every time someone buys flowers from a vendor on your site, you get a revenue share - a classic affiliate model. Moreover, you are also providing a valuable service to users, in that they can compare prices from multiple vendors on one page.

Google, however, disagrees. In Google's eyes, you are a leech on the system - an unnecessary step for consumers who just want to order some flowers. Surely it can't be good user experience to send someone to a comparison shopping site to buy from 1800Flowers if you could just send them directly to the vendor? As you and your 'cronies' continually optimize your user experiences and conversion rates, Google's ad space is increasingly flooded with 'middlemen' while the ProFlowers of the world are pushed lower and lower in the AdWords rankings.

Internally, Google is getting a lot of pressure to clamp down on your flower comparison site. The sales force - empowered with the task of going after the "G1000" - the top 1000 brands in the US - is having a hard time meeting their quotas, because ProFlowers can't bid high enough to get a lot of clicks. Wall Street is also concerned - after all, if GM, AT&T, and Coke aren't among Google's top advertisers, can Google really be a viable play long-term?

So, over time, the Google "quality score algorithm" slowly optimizes Google based on Google's preferences. No more AdWords arbitrage, no more lead generation companies, perhaps even no more comparison shopping engines (Shopping.com, Nextag, etc). In the new model, you type in "order flowers online" and 1800Flowers and ProFlowers compete for first position, while your little comparison site has to struggle for anything on the first page of results (where it matters).

A Logical Step for Paid Search?

When you look at the history of PPC advertising, "path two" seems like a logical - if dismal - path. When GoTo (now YSM) first launched in 1998, the model was close to a pure pay-per-click model. Barring editorial guidelines (i.e., a flower shop couldn't buy the word "Sony Playstation"), every advertiser on a particular keyword had the same opportunity to show up #1; whoever paid the highest cost per click (CPC) was the winner. I have always liked this model - both from a fairness model and from a user experience perspective (after all, an advertiser with a totally irrelevant product won't be able to afford to stay at #1 for very long, since they won't get any conversions).

Google iterated on the GoTo model with their "yield management" system, which combines CPC with clickthrough rate (CTR). In a yield management model, an advertiser is also rewarded for the quality of their ad text. This is a reasonable change to the system, since it was fairly easy in the GoTo system to create ultra-restrictive ad text as a way of bidding #1 but reducing errant click-throughs. Moreover, yield management is fair because it gives all advertisers the same opportunity to be #1 - if you have bad ad text, you can change it and eventually move to the top.

You could argue that "quality score" adds a third element to the equation - landing page relevance. Now, to show up #1, you need a combination of high CPCs, high CTR, and high landing page relevance.

The Dangerous Path Ahead?

The problem with the quality score model is this: it is not fair. In a quality score system, your ads can be irrevocably downgraded simply because Google says so. No amount of bid adjustments and ad text changes can change your fate. In many ways, it is similar to organic search (SEO). Google decides which site is the most relevant and that's final (it also follows that this sort of system could create an entirely new industry - SEMO - search engine marketing optimization!).

Think of SEO and SEM like two parts of your evening newscast - SEO is the journalistic content, and SEM is the commerical breaks. Ultimately, the decision to broadcast a particular news story is up to the television station. For example, if Fox News wants to only cover stories that paint conservative politicians in a positive light, they can do that. Fox's decision to run some stories and not others will determine their viewership and financial success.

But what if Fox decided it only wanted to run advertisements from certain advertisers? Well, on its face, that's OK - it basically means that Fox is leaving money on the table at their own peril. Let's make this scenario a bit more relevant though; what if 56% percent of all television ads ran through Fox? If Fox decided that they simply didn't want to do business with some advertisers, and set up rules that basically made it impossible for these businesses to advertise with Fox, could they? At some point, if a company wields dominant control over a form of media distribution, a different set of rules apply.

Much Ado About Nothing?

At the end of the day, this post may be totally irrelevant (it wouldn't be the first). If Google is just cracking down on AdWords arbitrage, I don't have too many complaints. But when you combine an internal focus on brand advertisers, Google's increasing requests for more and more data about advertisers' businesses, and an overwhelmingly dominant market share that enables Google to do whatever it wants, a press release about "quality score" can sound downright scary. Here's to hoping I'm wrong!

Tags: quality score, arbitrage, adwords, adsense, google, ysm, goto, overture, search engine market share

An Update on the Google Click Fraud Settlement

Many of you were likely bombarded this weekend by multiple emails from Google AdWords announcing the details of the $90 million click fraud settlement in Arkansas. To be clear, this is not a new settlement - it is the same one that was announced last March. The emails you are now getting are part of Google's obligations in the settlement.

As I noted when this settlement was originally announced, I believe that the only people who truly win from this settlement are Google and the plaintiffs' lawyers who filed the class action. I am not going to rehash my arguments in this post, but I do think there additional points worth commenting on.

It's Worse Than I Thought
First, I was wrong about the value of this suit to the law firm filing it. Perhaps naively, I wrote in March: "My guess is that a $90 million lawsuit will probably return at least $1-2 million to the law firm that filed the suit." When you look at the actual terms of the settlement, you'll see this whammy: "Class counsel [the law firm] intends to seek a maximum of 33 and 1/3 percent of the settlement fund, or $30 million, in attorneys fees and expenses in this case. Under the settlement, Google has agreed that it will not oppose an award of up to $30 million to class counsel." Translation: this $90 million settlement could quickly be whittled down to $60 million if the court agrees with the plaintiff lawyers' demands.

You may be asking yourself: why would Google agree terms that on their face seem outlandish? Well, consider this potential scenario: Google and the plaintiffs' lawyers are discussing a settlement. The plaintiffs lawyers give Google a choice: you can fight us over attorneys' fees and we'll demand that you actually pay out $90 million to plaintiffs, or you can let us collect our $30 million and we'll craft a settlement that makes it difficult if not impossible for plaintiffs to actually collect meaningful damages from you (for example, by giving Google sole determination of click fraud, but putting the onus on the plaintiffs to research potential click fraud, by giving advertising credits instead of cash, and by limiting claims to a two month window). If you were Google, which route would you take?

Should You Opt Out?
As with any class action lawsuit, you have the option to "opt out" of the class. This basically means that you are rejecting the terms of lawsuit and you are opting to individually sue Google instead. Generally speaking, this is an option in name only. After all, how many marketers actually have the resources to sue Google (it will likely cost six figures to try such a case).

So my sense is that unless you believe you have millions of dollars of unresolved click fraud claims, the risk-return ratio for opting-out isn't in your favor. And if you do have millions of dollars of click fraud, you are probably a big enough Google advertiser such that you should leverage your internal Google contacts first to try to get a refund before resorting to a legal action.

How Can You File a Good Claim?
OK, so let's say that you aren't going to opt-out of the suit and you do believe you were victimized by click fraud between 2002 and today. What can you do to make sure you maximize the value of your potential claim. Here are a few steps:

1. File during the filing period. Your claim must be received between June 19 and August 4th - if its late, you get nothing, no exceptions.

2. Consider retaining a click fraud firm that is able to analyze your log files. My sense is that most click fraud occurs in one of two ways: an irregular number of clicks on a particular keyword, or an irregular number of clicks from a particular IP address or referring URL. A company like Clicklab or Authenticlick can do this for you, although if you have a lot of clicks to analyze, the prices can get somewhat steep. Remember this point though: the key here is that you are analyzing historical clicks, not future clicks.

3. When in doubt, include questionable clicks in your claim. If you have the slightest inkling that some clicks could be click fraud, make sure you add these into your claim. There is no advantage (at least as far as I can tell) to being selective about your claim.

4. Plan ahead for the next lawsuit. It's possible that this settlement might not even end up being validated by the court. There is another class action that was filed in California, and the Arkansas court is holding a hearing in July "to consider whether the settlement is fair, reasonable and adequate." In the meantime, you owe it to yourself to sign up for a click fraud service (like the ones mentioned above) and to start collecting evidence for the future. There is currently at least one free service (The Click Fraud Network) that may be worth using while you deciding whether to make a purchase of an advanced service or not (note, however, that this service is for future clicks only, not historical clicks).

Wanted: A Search Engine Optimization Expert

It's not often that I use the massive audience of this blog (all 12 of you) for work-related purposes, but the time has come! Adteractive is looking for a search engine optimization expert to lead our efforts in, surprisingly, SEO. Some of the incredible benefits of this job include:

• Reporting to me (you can even get sneak peek of up and coming Blogation articles)!;
• Joining a leader in performance-based marketing (over $100 million of profitable revenue, two years in a row and counting . . .);
• Driving an SEO strategy from start to finish (does SEO ever really finish?);
• Working at a fun place with Google-like perks (well, free sodas, breakfast food, and dinner at any rate).

What are we looking for:

• Demonstrated expertise in SEO. Preferably at least one year of experience, proven results;
• The ability to "talk the talk" when it comes to SEO technology. You don't need to be able to execute an Apache re-write, but if you don't know what it is, and how to tell an engineer what you want, this probably isn't the job for you;
• A team player. Someone who works well with others, loves to be coached, and loves to coach others.

So, does this sound like you? Does it sound like someone you know? Does it sound like one hand clapping? If so, please drop me a line here: drodnitzky@adteractive.com. A resume is also helpful.

Oh, and if you want to see the official job posting, it's here:

http://www.craigslist.org/sfc/mar/156120356.html

Grading the Search Engines

While the ultimate measurement of a search engine is whether it is profitable for your business, there are plenty of intangibles that also need to be considered. In particular, I consider four key metrics:

1. Traffic volume: does the search engine have the ability to deliver a lot of volume to you?
2. Traffic quality: what percentage of the traffic converts, and what percentage is click fraud?
3. Customer service: how responsive is the search engine to your queries. Do they make you feel like you are a valued customer, or just an invoice?
4. Technology and tools: Does the user interface make it easy for you to quickly make changes, or is it like pulling teeth. Is there a readily accessible API?

Google
Traffic Volume: A+. The granddaddy of them all.
Traffic Quality: A. Pretty good, though there could be a little more transparency on the click fraud issue.
Customer Service: C-. Nice people, but I've never gotten the feeling that anyone really understands the significance of the millions of dollars I spend with them.
Technology and Tools: A. Leader in UI tools, slightly buggy API, but getting better.
Overall Grade: 3.66 (A-)

Yahoo Search Marketing
Traffic Volume: A-. Losing MSN hurts.
Traffic Quality: B. Pretty good, but it seems like there have been some click fraud issues as of late.
Customer Service: A. Excellent. Totally on top of their game.
Technology and Tools: C-. Terrible user interface. Promising it will get better . . . in August?
Overall Grade: 3.33 (B+)

MSN
Traffic Volume: B+
Traffic Quality: A+.
Customer Service: B. Nice people but they need to fix some tech issues (like their testing bot speed, which can crash servers due to the lack of a limiter).
Technology and Tools: B. The first version of the UI is pretty good, but I expect improvement.
Overall Grade: 3.33 (B+)

Ask
Traffic Volume: C
Traffic Quality: B - I think the value of Ask.com is high, but the distribution network is questionable.
Customer Service: A
Technology and Tools: C - Using the Looksmart UI wouldn't be my first choice.
Overall Grade: 2.75 (B-)

Miva
Traffic Volume: C
Traffic Quality: C
Customer Service: A
Technology and Tools: C
Overall Grade: 2.5 (C)

Looksmart
Traffic Volume: B
Traffic Quality: F
Customer Service: D
Technology and Tools: C
Overall Grade: 1. 5 (D)

No one's got a 4.0 yet, but hey, we haven't even reached "mid-terms" for search engines, so there's a long ways to go!

Trust and Distrust in SEM

These days there is a lot of advice floating out there about SEM. Here's my recommendations on who to trust and who to take with many grains of salt.

Who to Trust:

1. Forum Posters: For whatever reason, people who post on forums like WebMasterWorld or Search Engine Watch seem to be very forthcoming with accurate information. Perhaps its because they enjoy feeling like an expert, or perhaps its the egalitarian nature of forums, but I find that the best way to get an answer to a hard question is to throw it up there on a forum and wait for the responses to come in.

2. In the Trenches Search Engine Workers: What do I mean by "in the trenches"? Anyone who works at a search engine but not in a sales role. For example, at Google, there are tons of rank-and-file employees who are in charge of creative maximization, editorial review, algorithmic design, etc. If you can connect with any of these folks, you'll find that they are often willing to share "insider information" that can give you a major edge on the competition. I once had a connection with a guy on the Google algorithm team. Let me tell you, that paid some major dividends!

3. Your Colleagues at Other Companies: We SEMers gotta stay together! When I run into friends at trade shows - perhaps even folks from direct competitors to Adteractive - it's always refreshing to share ideas and observations about SEM. Granted, I don't give away the Crown Jewels of my SEM techniques (nor do I expect them to), but there are always some good nuggets that improve performance.

4. Numbers: Ultimately, SEM is a numbers game. When you measure everything you do (tracking URLs, Web analytics, multivariate testing of usability, etc), you can verify whether someone is telling you the truth or not about a particular campaign or search engine.

Who Not to Trust:

1. Speaker Circuit Speakers: This means anyone who always appears at every trade show, writes for ClickZ, or is on the board of SEMPO. These folks are great at self-promotion but not great at actually guiding you toward accurate information.

2. Most Bloggers: Sadly, most bloggers regurgitate information from other blogs or Google press releases. They add no value and actually have no expertise to impart to you. This blog, of course, is an exception to the rule!

3. Sales Reps from Search Engines: OK, this is a bit unfair. In truth, almost every sales rep I work with is a straight-shooter that will always tell me where to buy and not buy on their engines (kudos to Kevin, Jung, Jean, Coco, Ben, Michael, Mike, Kirsten, John, Kristen, and my other great reps!). The problem really is with the reps that bombard me with calls from second-tier sites that want a quick $1000 for one week campaign. As I noted in my article in The Search Marketing Standard, when you hear someone promising "50 billion clicks a month" it's time to run away as fast as you can.

4. Anyone with a Web site with More than 10 Pages of Content on One Page: Increasingly, I'm seeing the "Cory Rudl" types moving into SEM advice. These are the Web sites that seem to scroll forever with lots of free bonuses, testimonials from full-time testimonial experts, and a strange affinity to yellow highlighting. I'm amazed that these sell, but I'm certain that they won't work for you!

Who did I miss?

Wedding Mania!

As some of you know, I am in the final stretch of preparations for my impending wedding in early June. As such, I haven't had much time to eat, sleep, or blog. So please excuse the infrequency of postings, and please excuse the non-structured nature of this post in particular.
Here's a laundry list of interesting observations I've had over the last few weeks.

1. Google Trends (available here: http://google.com/trends): OK, this is a decent idea, and what's more, Google may be able to monetize it at some point (unlike so many other Google releases as of late). This is basically no different than the Yahoo Buzz Index, but since it's a Google product, it's getting a lot of press. By the way, the "Project Panama" release of Yahoo's new Overture user interface will have a similar feature built into it (and that is all I am going to say about the new Overture product, for fear of violating any agreements I once signed).

2. A Blackberry Massage? While looking for a spa in Kauai (for the honeymoon, of course), I came across a place that offers a BlackBerry massage - that's right, a hand massage for people who use their BlackBerry cell phones too much.

3. Yahoo Click Fraud. Is it me, or has there been a sudden increase in click fraud on Yahoo Search Marketing (Overture)?

4. Ameriquest and University of Phoenix. Now that Ameriquest has closed all its mortgage offices and University of Phoenix has had several quarters of disappointing earnings (and significant layoffs as well), it will be interesting to see how the online lead generation business reacts. At a minimum, I suspect that this will be the beginning of the end for small "mom and pop" affiliate marketers who are trying to make a buck at mortgage or education lead generation. Consolidation here we come!

5. Almost Famous. For those of you who subscribed to the Search Marketing Standard - the first magazine dedicated to SEM - you'll see that the cover story is written by yours truly. Note that it was a pretty basic article (as requested by the editor of the magazine). Its suppose to appeal to a wide audience, unlike this blog which is focused on search uber-nerds!

6. eCompany 2001. While away for the weekend at my Bachelor Party (yes, the wedding rears its ugly head again), a friend of mine handed me an eCompany magazine from 2001. Talk about a time capsule. In addition to a huge story about the inevitable success of Enron Broadband, there was also a great read on "the Dumbest VCs," with Michael Moritz of Sequoia Capital heading the list (Mr. Moritz later went on the fund such companies as . . . Google). Too bad there wasn't a story on the dumbest magazines.

7. Larry and I. So it turns out that the Rabbi (for my wedding, of course) happened to go to high school with my Mom in Chicago. This reminded me of another person in my Mom's high school class - Larry Ellison, founder of Oracle. I wonder whether I could get any blackmail money from Oracle by threatening to post high school pictures on my blog?

8. TagStreet. It may be the first SEM parody blog - check it out at tagstreet.blogspot.com. I didn't write it, but it seems that the author "DrewZPerogineFalcon" is a homage to DavidZHawk, my nom de plume.

OK, that's it. I'm hoping that once I am married (and settled down) I can write a little more frequently and eloquently. Until then, make sure to keep reading Jay Weintraub and the few other bloggers that actually have something to say!

What Google Has That You Don't: Information Asymmetry and Search Engines

Google asks you to share a lot of information about yourself and your business with it. For example, depending on the Google products you use, you share:

• All of the sites you surf (if you use the Google Toolbar, Google places a cookie on your computer that won't expire for 50 years!)
• The contents of your emails (if you use GMail)
• The files on your computer (Google Desktop)
• Your best keywords (AdWords)
• The behavior of people going to your Web site (Urchin)
• The structure of your Web site (Google SiteMaps)
• Which people actually convert for you (Conversion Counter)
• How you intend to use Google's API (Google now wants to visit you three times a year to make sure you are using the API properly

On the other hand, Google doesn't like to share with you. Among the things Google refuses to share:

• How they determine whether your site is relevant for organic results
• How much you have to pay for top position on AdWords
• How the content network works (how do you know when you are going to show up)
• How quota for the AdWords API is determined
• How Google determines whether a click is click fraud
• What your overall clickthrough rate is (Unbeknownst to most people, the CTR that shows up on the AdWords API is Google-only, it does not factor in the CTR on the Google distribution network.
• Why certain ads get rejected
• Financial projections (if you are shareholder)
• If you live in China, Google has decided not to share all relevant information with you at the behest of the Chinese government
• Originally, Google didn't even tell consumers that PPC ads were ads - until the FTC threatened to sue, hence the "sponsored listing" tag you see today.

In a nutshell, Google doesn't like to share, but they are more than happy to demand a lot of personal and financial information from you.

That doesn't sound very fair, does it? After all, every since we were kids, we've learned that the Golden Rule is "do unto others and you would have done onto your self" or "share and share alike" or (perhaps in a different vein) "show me yours and I'll show you mine."

Right now, though, Google doesn't have to share, for two reasons. One, consumers don't realize how much information they are freely giving to Google and how little Google is giving back. I was out to dinner last night with a friend and he posed the following hypothetical: tell a consumer that they will need to share their email contents, searching behavior, Web usage, desktop files and small business marketing strategy with Microsoft. How would they respond? Of course they would go crazy. But tell them that Google already has that information and they would probably shrug.

After all, Google still has the reputation of "doing no evil." If you asked a consumer (or small business professional) to use adjectives to describe Google, you'd get responses like "innovative", "fun", "relevant" and "best of breed." Ask the same person to describe Microsoft and you might here "evil", "monopoly" and "big brother."

As long as consumers and advertisers have it in their head that Google is their friend, the outrage against Google's asymmetry of information will be limited if any.

The second reason Google doesn't have to share is because they simply don't have to. Especially in the case of search engine marketers, SEM media buyers have no choice but to use Google and follow Google rules. For the average SEM campaign, Google probably accounts for 50-70% of cost and profit. Failure to follow Google's rules will get you kicked out of AdWords, which will destroy your profit stream overnight.

So if Google wants to, they can continue to ratchet up the demands on their advertisers. If they want to charge a $50 annual "maintenance fee", they could do that. They could start charging for API usage (too late!), they could categorically ban types of sites they don't like (for example, lead generation companies could someday be banned), they could set arbitrary minimums for keyword prices (too late there too!), and on and on.

In essence, because there is a lack of public or business outcry, because Google still has a great reputation as an ethical company, and because Google has a virtual monopoly on search engine marketing, Google has absolutely no incentive to share with you, but absolutely no barriers to ask you for more and more data.

Again, this may not concern you that much, but ask yourself how you would feel if you replaced the word "Google" with "Microsoft."

Ultimately, I do believe that Google's aura will start to wear off over time. People forget that Microsoft was once associated with words like "innovative" and "revolutionary." But when Microsoft got too big and started throwing its weight around, the tide changed.

And that actually begs an interesting question: if Google has a 70% or greater share of the SEM industry, if SEM is becoming an increasingly important foundation of every marketer's marketing budget, and if Google uses its market position to make asymmetric demands on marketers (share with us, pay us for additional services, we can ban whoever we want without explanation), will that someday put Google on a collision course with the Department of Justice's Antitrust division? Will Google someday have the same 'evil' brand that Microsoft has today?

As I see it, that's not such an unreasonable scenario (well, assuming that the DOJ still enforces the Sherman Antitrust Act, that is). Again, when you have a company that is gradually collecting reams of personal information about its users, has a monopoly position in a growing and lucrative industry, shares as little information as possible with anyone outside the company, and has even started putting up competitive barriers (i.e., with the AdWords API - you can't display Yahoo and Google info on the same page of a commercial bid application), it doesn't seem like a pipe dream anymore.

These days, I know a lot of marketers who are actually rooting for MSN's AdCenter to be a resounding success. They're hoping that MSN can take some of the pie away from Google (Yahoo clearly hasn't been doing that lately). Any battle in which Microsoft is portrayed as David instead of Goliath is one where there's already a formidable Goliath.