Nội dung

CHAPTER 2: A $21 Billion Afterthought: How Google Entered the Advertising Market expanded to cover “semantic variations” and did not work very well at first. Google is constantly tuning the workings of Expanded Broad Matching, and at times it appears that broad matches, especially when bid high, lead to erratic performance as ads are displayed unpredictably on “semantically related” search phrases. Much advertiser dissatisfaction has been directed at the content targeting option in AdWords (called AdSense when viewed from the publisher’s perspective—that is, from the perspective of website owners displaying ads paid for by AdWords advertisers). Google was uncharacteristically quick in ramping up this program, likely due to their perceived need to race against competitors like Yahoo and DoubleClick for control over online advertising inventory. In the past couple of years, Google has steadily improved the feature set, transparency, and accountability of this ad program. Most of these changes have come in response to a steady stream of advertiser input, and no small degree of advertiser frustration. In other instances I tend to disagree with analyses that peg any change to the AdWords platform as “just another Google cash grab.” The current Quality-Based Bidding formula is certainly doing no harm to Google’s long-term revenue prospects, but the reasons for releasing and refining this formula are complex, and not all tied to pure revenue maximization. The user experience, in my judgment, does remain paramount in the AdWords program. This search engine user satisfaction, in turn, keeps Google in business, and highly profitable. How Google’s DNA Influences the AdWords Game Let’s turn to an overview of idiosyncratic policies and attitudes that will become familiar to you as you play the AdWords game. Many of them stem directly from the values of the founders and their immediate circle. On the whole, though, a certain kind of attitude permeates the company. If I had to boil it down, it might be “never forget the user experience,” which in the case of a search engine company means “don’t intrude, just help people find what they’re looking for.” A generalized wisdom also prevails: “don’t forget why we’re here and AltaVista isn’t—don’t be dot-com road kill.” Editorial Rules and Banned Items Google spent a lot of time in the early going debating advertising policies. Today, not only is there less to debate, but the policies themselves may be less transparent than they once were. But you may find yourself running afoul of certain policies. That shouldn’t be inherently surprising. Any publisher (online or off) is going to have guidelines for the types of products that they accept advertising for. Google is no different. They must ensure, of course, that ads comply with applicable laws. But they also go beyond the law in areas they worry could become controversial and alienate the general public. Google has sometimes reminded advertisers that it does not censor search results. Whereas an ad may be banned for something like hard liquor or a certain type of knife that might commonly be used as a weapon, this does not preclude pages about these items from showing up in the regular search results. If you’re curious, https://adwords.google.com/select/contentpolicy.html offers a list of basic content policies. It wasn’t published until November 2004. It does not give much detail, and 57 58 Winning Results with Google AdWords many gray areas are still left up to editorial discretion. In my experience, the list of prohibited industries and ad subjects provided on this page is nonexhaustive. You may run into other problem areas that have yet to officially make it onto the list. Pop-Ups and Other Poor User Experiences Pop-up ads, and Google’s policy prohibiting sending visitors from your ad to any page containing such ads, is really now just an example of a larger-scale, systematic program Google has implemented to police what it calls “landing page and site quality.” In essence, Google once took strong stances against a few things that it believed led to a negative user experience. Now, that iconoclastic approach has been tempered and extended through the collection of years’ worth of user feedback. A few of the old policies no doubt remain alongside a richer list of illadvised practices, such as customer data collection without adequate disclosure of your business credentials. Since both Yahoo and Google reserve the editorial right to ban any ad just for pointing to a page they deem irrelevant to the ad, it’s not surprising that Google has also taken the initiative in banning ads that point to pages that they deem to provide a poor user experience. Want to show your ad on Google? You will pay a premium to point it to a page that serves users an annoying pop-up ad; if your ad shows up at all, that is. At this point, such specific guidelines shouldn’t be taken too literally, in the sense that you can also run afoul of Google’s landing page and site quality guidelines by doing other things that smack of deceptive or irrelevant advertising. As mentioned, I’ll cover the gamut of ad quality issues in an upcoming chapter. Ironically, this very policy led some entrepreneurs to come up with pop-up-like technologies that were different enough from pop-ups that they passed editorial muster with Google for a time. But “working around” Google’s rules is more difficult now because Google looks for a variety of signals of negative user feedback. So much like the Mom who doesn’t give her child a pass for saying “Oh, Fudge” (the intent was there!), calling something a pop-in, a pop-around, a pop-a-rooni, or a pop-a-doodledoo is now unlikely to “fool” a Google policy specialist if it displeases users. Privacy Policies It will be interesting to see just how far Google The Advertising Company is willing to go to collect demographic data on users as competitors attempt to do the same. Google, for now, has relatively strict privacy policies and does not know much about the individual surfer using the Google Search tool, although it does look at the user’s geographic location (IP address). Google’s history might suggest that it will go slow on offering advertisers advanced demographic targeting, while its competitors forge ahead with more intrusive schemes. Currently, Google does report limited demographic targeting information to its advertisers, under the auspices of partnerships with (for example) social networking sites such as MySpace. Google does not want to be portrayed as a privacy threat, so its approach is to test the waters gently and to lull competitors into making the first invasive moves (so a competitor, not Google, can take the rap for moving the goalposts). CHAPTER 2: A $21 Billion Afterthought: How Google Entered the Advertising Market Because Google The Global Information Powerhouse has now built so many integrated services around its core search offerings, the topic of privacy and Google now seems to exceed the scope of this book! What are we to make of Google Health, a would-be centralized repository for patient medical records? Or Google’s steady rollout of a variety of telecommunications services? The apparent mapping and archiving of pretty much everything, as Google’s stated corporate mission? When we look at specific policies with regard to searcher privacy, Google looks relatively innocuous. But Google’s role in the global information economy could well make it the single greatest potential threat to your privacy. But that’s another book. Are Policies Consistently Enforced? When it comes to the ongoing quest for top rankings in organic search rankings, many business owners have been torn between so-called “black hat” and “white hat” strategies. A broad consensus has emerged in the industry: you have to be at least “gray hat” to do well. The unwillingness of many businesses to strictly follow Google’s webmaster guidelines often comes from observation of competitors bending the rules for years running, apparently benefiting from the deception. “If you can’t beat ’em, join ’em” seems to be a constant temptation in the world of search marketing. It takes a strong will to resist temptation. For the record, it’s usually best to resist, in my opinion, because that competitor might not be doing as well as you think, and may eventually be subject to catastrophic penalties as Google adjusts their search ranking algorithm or makes manual adjustments to catch certain forms of unreasonable “gaming” of their algorithm. It was inevitable that a certain degree of this gamesmanship might gravitate towards the paid search program, too. Google combines automated methods with human discretion to weed out and punish advertisers who don’t play fair. The problem with some of Google’s policies in the past is that they weren’t grounded in any solid principles. Some have been nearly unenforceable; in other cases, Google has chosen not to enforce them—in essence, “looking the other way” on minor violations given the complexity of enforcing the rules to the fullest extent. One quagmire is the quiet but rarely enforced prohibition on “double serving.” If you think about it, an unscrupulous advertiser could open two (or ten) separate AdWords accounts and blanket the page with ads for the same product or service, crowding out competitors. Google prohibits such behavior, but it’s not uncommon for exceptions to slip through. There are too many gray areas where it actually makes sense to have two ads showing on the same page from the same company on the same keywords. A large company like IBM might have separate divisions that are both likely to benefit from rather different ads on keywords that sometimes overlap. More to the point, perhaps, this is yet another policy matter that has now been subsumed under the all-encompassing enforcement mechanism of Quality Score, which is opaque and largely automated. The extent to which human inputs (real policy specialists twiddling the knobs, as it were) affect Quality Scores is not precisely known. Again, I’ll be delving deeply into this shortly. By moving policy enforcement into a “no-tell” zone (into a numerical score that takes myriad factors into account), “bad guys” don’t necessarily find out what they’ve done wrong. Google is under no obligation to tell them, either. 59 60 Winning Results with Google AdWords That’s led to further uncertainty among conscientious advertisers, wondering if they’re being wrongly judged. Where Quality Scores are low, they want to know what factors are causing the low rating. Thus Google is now swinging back towards more transparency. They’re studying ways of giving advertisers some clue as to which policies they’ve violated, or what factors led to their keywords being given low “quality” ratings. Despite the imperfections that have cropped up along the way, Google is moving in an interesting direction when it comes to policy enforcement. The degree of automation they are attempting to achieve would seem to codify their policies and remove excessive discretion from the hands of editors. The analogy would be with making good laws: judges are needed, but legal interpretations in specific cases should not vary wildly depending on which judge you deal with. Tight Control of Information Flow Despite its democratic, fun image, Google is a serious business entity that holds its cards close to the vest. It employs a degree of secrecy that many consider excessive. Some recent political reading that equated undue government secrecy with a deficit in democracy made me sit up and think hard about just what was going on over there at the Googleplex. Google staffers have always told me as much as they possibly can to help me understand AdWords features. But the company’s secrecy often precludes them from telling the whole story. The pressure on Google seems to have abated some now that the nail biting over their IPO is done. (The first trade of Google shares under the ticker symbol GOOG went through at 11:56 A.M. ET, on August 19, 2004, for $100.01, well above the offer price of $85.) In the pre-IPO quiet period, most everyone in the company was terrified of giving away material information or being perceived to promote the stock, since even the suggestion that Google might be a good investment would have violated SEC regulations and led to delays in the IPO. Delayed IPOs, as AltaVista found following 1999, are not good karma for search engine companies. All Search Engine Companies Are Secretive about Algorithms Much of the secrecy employed at Google is absolutely necessary. Search engine companies cannot share much about the “secret sauce” of their methodologies on a month-to-month basis, since millions of website owners are jockeying for high rankings in the free results. In this regard, Google is not alone. Its cryptic commentaries about its search engine ranking methodologies are in keeping with the demands of its ongoing battle with index spammers. Concealing Details of How AdWords Functions, for Competitive Reasons An unusual quirk of AdWords is that many features are a lot more complicated than similar features offered by competitors. Add to this the engineer-speak combined with public relations spin and you’ve got some features that are downright befuddling. Google has opened up their public relations outreach in the past four years, however. Spokespersons such as Nick Fox, a key manager of the “ads quality” team, tirelessly explain new features of the ad ranking formula. Nick has directly answered many of my pointed questions, and has been forthcoming in public conference sessions as well. For example, on the question CHAPTER 2: A $21 Billion Afterthought: How Google Entered the Advertising Market of whether total advertiser spend or length of account history (in terms of time) affect accountwide Quality Scores and performance, Nick was clear in stating that Google “does not believe in perverse incentives, so it doesn’t include time or spend in the Quality Score.”21 I believe him. Google also has more products in the marketplace today than they did two or three years ago. This has evidently led to a more systematic outlook on how to communicate with the public. Google Analytics (a website analytics service) and Google Website Optimizer (a landing page testing tool) are key examples of products used by paid search marketers. The outreach effort has evolved towards a mature dialogue with affected businesses and interested commentators and journalists, with heads of the product development teams making their insights widely available. Google’s recent move towards glasnost has been refreshing. It appears that they have counseled their key public faces to give direct, clear answers to questions about how products work, while maintaining confidentiality only where absolutely necessary. In fact, it’s the substance of their complex, automated systems that leads to most of the apparent obfuscation. No amount of spokesperson explication or number of oversimplified PowerPoint presentations can make up for the fact that the AdWords program has always been complex, and has shifted from generation to generation quite rapidly, with nearly no external actor being competent enough to distinguish between a small feature change or a major new release. For example, in the early days of AdWords, Google invented a sliding scale to measure the exact minimum threshold of clickthrough rate (a relevancy requirement) that advertisers were required to meet to keep keywords enabled. Officially, the cutoff was 0.5%. But Google emphasized that this was actually “0.5% normalized for ad position.” This means that the relevancy policy, as measured by clicks on your ads, is relaxed as your ad moves down the page to a less visible position. (The 0.5% is no longer part of the formula, but the threshold for what counts as a “good” CTR is still normalized for ad position.) Many advertisers wanted to know exact numbers for CTR cutoffs in different ad positions, but Google never disclosed this. In part, this was because this CTR cutoff would vary by keyword (industry norms). Therefore, disclosing all the figures would have disclosed proprietary search behavior information. Google is unlikely to disclose this level of detail on such matters, at least for the time being. While literal-minded advertisers often found this coyness frustrating, pretty much everyone’s gotten used to it by now. In fall 2003, Google claimed to be raising that cutoff to 1.0% on some keywords in some situations, but the explanation for that was so confusing that virtually no one understood it. The 1.0% cutoff formula, whatever it was, was quietly dropped. Google policy gets a lot more complicated than that. Many features have not been amenable to straightforward description because they’re based on proprietary algorithms and predictive formulas. Pricing on content targeting, for example, is subject to a so-called Smart Pricing formula, where Google’s software determines the cost of a click (subject to your stated maximum bid) based on a predictive or actuarial formula that looks at which kinds of pages online are more likely to return a higher conversion rate to sales. Three motivations have underpinned these elaborate feature designs. First, the brilliant Google engineering team always wants to take a stab at solving a problem through software. Second, Google wanted to design AdWords as an elaborate, proprietary system to muddy the waters in its drawn-out patent dispute with Overture. Finally, the more difficult Google made AdWords to copy, the less likely competitors would be to ape it. Certainly, Overture and FindWhat moved 61 62 Winning Results with Google AdWords quickly to duplicate some of the most compelling features of AdWords—particularly matching options. But it takes considerable experimentation and development time to copy the more arcane features. Is feature even the right word for a formula based on complex interrelationships among a host of variables? Google AdWords is not only multifeatured, it’s multiformulaed. Not Disclosing Details of AdSense Program Content targeting has been an ongoing source of concern for advertisers. Like Overture, Google is content to boast of major publishers and certain “poster child” publishers who have participated in its AdSense program. But there is poor disclosure of not only the full list of participating publishers, but many other details of the AdSense program, such as how pricing is determined, what the revenue share is, and more. Advertisers see click costs, and publishers see basic reports and receive checks in the mail, but a lot of detail is missing. Failure to Break Down Reporting of Ad Spend by Country of Origin One thing I always found curious was that Google will let you choose which countries you show your ads in, but the reporting interface doesn’t break down your click costs by country. I’m sure that’s one feature Google has on its to-do list, but it does stand out as an example of an area of nondisclosure that was left to linger too long. There have been numerous others. Google’s Service Revolution, or “First We Take Chelsea” Relative to the program’s popularity, AdWords was administered by a skeleton staff of customer support and editorial staff in the period 2002–2004. Since that time, Google’s headcount has exploded. The new staff complement runs the gamut from engineering talent, to advertising sales execs, to customer support reps. Their Manhattan office space in the trendy Chelsea area, still a novelty to many longtime observers of the Mountain View, CA–based juggernaut, seemed to me considerably more labyrinthine and bewildering in March 2007 when I visited it than it had been only a year previously.22 As always, the food is tasty (organic or vegan if you want), and a fun-looking selection of confections and beverages is always on hand. Mountain View meets Manhattan, in style. Google appears to believe that if you’ve made it into the Village Voice, you’ve really made it. Google added nearly 7,000 employees in the period between May 2007 and the same time in 2008, when quarterly financial reports are published for Q1. This brought the company’s total headcount to over 19,000. Prior to that, headcount grew at least 80% in each of the previous four years. That’s a breakneck pace. While there may be some slowing of that pace due to the consolidation of acquired DoubleClick employees, as Google expands into completely new fields, its ravenous appetite for talent may well continue unabated for some time. It’s likely that this trend will eventually put a damper on profits, in spite of the notes of caution formally offered by company management. It’s also worth noting that much of this growth is international. At the end of 2002, the year AdWords’ pay-per-click version was launched, the company had only 682 employees. None were stationed in a chic Manhattan neighborhood. As a result of the buildup, advertisers have noticed significantly increased service resources. Many of us who run agencies with multiple clients have permanent “agency reps” who assist in expediting troubleshooting and answering difficult questions. Sometimes, however, this has spilled over into meddlesome behavior. Accustomed to “self-serve,” some advertisers find Google’s growing customer service staff “salesy” just by virtue of their very presence. Achieving a consistency of tone and behavior across the board becomes a challenge with such rapid growth. CHAPTER 2: A $21 Billion Afterthought: How Google Entered the Advertising Market In light of all this, the company’s early technical orientation—automate where you can—is actually a strength, buffering advertisers from the most pesky effects of salespeople with too much time on their hands. At the root level of AdWords is a “product” that works consistently (if in complex fashion). Fundamental decisions about the advertising auction are driven by product managers incorporating feedback from advertisers and users, with only minimal input from this “growing headcount.” Google Underestimated Need for Customer Service At first, by using software to facilitate editorial review, Google assumed that it was onto something big: a business model that could reap revenues even greater than Overture’s, while spending far less on human support. As the program grew, it became difficult to ignore the huge gulf in service. Google became aware that advertisers need a lot of hand-holding, and the pace of hiring accelerated. Today’s attitude towards service appears to be nearly a 180-degree reversal from the early “don’t call us, we’ll call you” approach. Because Google can attract good people and is so stringent in its hiring process, their new commitment to service could make it tough on the competition. That being said, new concerns are now being raised about Google’s overhiring cutting into profitability. What about Rewards for Good Customers? There doesn’t seem to be a conclusive policy on how to provide dedicated support for agencies and advertisers who spend more. The overall level of service and attention is now so high that pinpointing exactly what criteria are used becomes less important, but for some literal-minded folk, the process may be murkier than they’d like. It’s important to recognize, in any case, that few advertisers spending any decent amount are denied time and support, because Google has such vast resources. So make use of it, no matter how big or small you are. Certainly the “sales potential” of an agency or large advertiser (how much can Google expect them to spend in the future) appears to be a large part of their internal criteria for how much extra dedicated support to provide. Uncertain Relationships with Advertising and Marketing Agencies Third parties often advise clients on how to use AdWords, or directly manage complex campaigns. (That’s what my firm does, for example.) Observing Google’s progress in dealing with the environment of marketing and advertising agencies, they have never fully given up on the idea that advertisers really should be coming directly to them for advice. However, this situation appears to be improving. A Google Advertising Professionals (GAP) program, launched in November 2004, was an interesting initiative that was supposed to sort out qualified from unqualified individual AdWords campaign management practitioners. A company wide (agency) version of this is also available. This is more of a training and indoctrination program than anything else, however. The reward to the qualified professionals and agencies is minimal at best, though ostensibly it helps advertisers avoid working with “hacks.” Agencies certainly get much less out of Google in terms of financial rewards (such as a commission) than they have in any relationship in the history of advertising. On a variety of fronts, including the Google-agency relationship, observers have asked the question: is Google sucking the proverbial oxygen out of the room? While consultative relationships have improved 63 64 Winning Results with Google AdWords and become more formalized—a key improvement, to be sure—many of the leading AdWords consultants and evangelists must make their living from service fees alone, putting them too close to break-even for comfort, while Google’s extreme profit margins continue to fuel the company’s growth. There are practical hurdles to be addressed before such traditional advertising industry practices can be adopted, particularly in the “geek culture” which has served Google so well. However, the goodwill and indeed survival of the search marketing agency community, in particular, may hinge on a recalibration of their financial relationship with Google. In its formative years, having the right (geeky, iconoclastic, world-beating) attitude at the right time was a big part of what made Google into a global powerhouse. Some critics predict that this same attitude could be its undoing. Experts believe that the degree of Google’s cooperation with the developer community (and I would add, the marketing ecosystem) will determine whether the company has the staying power of a Microsoft.23 Coexisting with “Resellers” and the Ecosystem in General Through the back door, Google may be studying ways of responding to the above analysis. Beyond AdWords, the company has new, highly technical products, like Google Analytics and Google Website Optimizer. It has initiated partner and reseller programs for these products. By instituting criteria for membership, working closely with that community on product development, and figuring out ways of steering valuable consulting business to such resellers and partners, Google can study the ins and outs of forming such productive relationships. Such relationships seem to be founded on classic models common in the software industry, especially in high-ticket enterprise software. What makes this unorthodox (as usual) is that Google’s products are often free, and many of the customers for them are small to midsized businesses. What will it mean for my consulting firm to “resell” Google’s free product to a small customer, I wonder? Like many others, including Google themselves, I can’t wait to unravel that puzzle. Google’s survival may well require it to balance its secretiveness (against increasingly feisty competitors like eBay and Microsoft) with a new openness in its dealings with certain partners. And as it enters adulthood, it might need to shed its laid-back attitude and become more strategic in forcing users and advertisers (and welcoming reseller partners and application developers) into proprietary, but widely shared, information technology architectures. These architectures may allow new uses of Google’s products, but when the goals of third-party applications conflict with Google’s, Google can always block access or raise prices on certain types of usage. An example might be the recent acquisition of aQuantive by Microsoft, making it the owner of Atlas, a popular third-party bid management tool. Microsoft owning a tool that can gather a huge amount of search behavior and economic data about how real businesses are faring on the AdWords platform may have caused some initial alarm at Google, but it doesn’t seem to have helped Microsoft gain market share. Regardless, through their ownership of Google Analytics and now DoubleClick, Google has trumped competitors in the race for dominion over business data pertaining to advertising performance across wide swaths of the online world, regardless of whether that advertising is going through Google AdWords. Another model Google could pursue to ingratiate itself to the developer and business ecosystem is to release more thoroughly open-source products, as it has recently done with its CHAPTER 2: A $21 Billion Afterthought: How Google Entered the Advertising Market new browser, Google Chrome. Such initiatives can be advantageous to the community and even competitors, while solidifying loyalty to Google as a source of innovation. With specific regard to paid search, the introduction of the AdWords Application Program Interface (API), is promising insofar as it signals a heightened commitment to cooperate with third-party developers and agencies. But it’s clear that Google does not view all third parties as cooperative with its own goals and its customers’/users’ needs. When it comes to third-party “layers” such as bid management technologies, Google can price API tokens (the price for any automated access of the AdWords interface) and set the API terms of service in such a way that it is costly to build certain kinds of software overlays. Such overlays may be seen as superfluous annoyances, given that Google is also developing new features to help advertisers directly within the AdWords platform. Google will likely need to create more formal partnerships in the future, and invite more developers and agency types into ongoing dialogues about features and business relationships. They have already begun to formalize this process, putting together new “blue-ribbon” panel groups to assist in ongoing feedback about their products (including the AdWords interface) that will coexist with older means of gathering feedback from forums, users, and webmasters chosen for limited beta tests. In the past, the dialogue with the “affected community” often appeared to be limited to select groups of beta testers and informal chatter mediated by the likes of anonymous Google employees posting on forums, such as GoogleGuy. These means of communication did little to forge long-term adult relationships with Google’s agency advocates, resellers, and technology partners. Google has now begun to reach out to these latter players, which augurs well for Google’s long-term survival because it is more aligned with the lion’s share of online advertising dollars. Geek-speak will never be out of vogue in this medium, but it will now be tempered by business focus. Google’s unique culture was shaped first and foremost by its founders, moderated by technology veteran CEO Eric Schmidt. The company’s ability to focus depends heavily on the ongoing involvement of top management in steering what has become an increasingly diversified enterprise. To paraphrase the “risk factors” sections of the company’s SEC filings: If Google should lose the services of Larry, Sergey, or Eric, it could be in big trouble. Time will tell, but there is no reason to believe that Google’s top people have anything in mind other than overseeing its continued breakneck pace of growth and change. By building a sound and consistent means of interacting with partners, Google will also build allies for the long term, allies who bring more resources and perspective to the table than the first wave of geeky foot soldiers who helped Google cross the chasm to global search supremacy in the first place. Endnotes 1. Josh McHugh, “Google vs. Evil,” Wired (January 2003), archived at www.wired.com/ wired/archive/11.01/google_pr.html. 2. A useful primer on such matters, covering the whole range of contemporary administrative theories, is Charles Perrow, Complex Organizations: A Critical Essay, 3rd ed. (McGrawHill, 1986). Chronicles of dot-com startup desperation, greed, and excess such as 65 66 Winning Results with Google AdWords Po Bronson’s, The Nudist on the Late Shift: And Other True Tales of Silicon Valley (Broadway, 2000), don’t seem appropriate to grokking the Google work culture, which has always seemed relatively settled and self-confident as opposed to chaotic. As nerdy and unconcerned as Google employees may appear to be about the traditional goals and structures of large corporations, keen observers (see David Vise, “Following a Rich Tradition: Under the Avant-Garde Veneer, an Old-Guard Startup Strategy,” Washington Post, June 24, 2004, E01) have argued that this powerhouse is very much a traditional Silicon Valley “insider” company. Key early investors and advisers—including Jeff Bezos, John Doerr, and Michael Moritz—were all seasoned members of the Silicon Valley elite, and the hiring of Eric Schmidt as CEO introduced a degree of settledness to a group that was already arguably mature beyond its years. Of course, some sensationalistic press reports have suggested otherwise. 3. For example, Danny Sullivan, “Where Are They Now? Search Engines We’ve Known & Loved,” Search Engine Report, March 4, 2003, archived at searchenginewatch.com. 4. For the whole story and a detailed how-to guide to the new Yahoo Search Marketing platform, see Mona Elesseily, Mastering Panama (Page Zero Media, 2007). 5. Search Engine Positioning (Webware Publishing, 2001). An earlier iteration, Achieving Top 10 Rankings in Search Engines: Insider Trade Secrets from Positioning Pros, a spiral-bound self-published effort, was released in 1999. Marckini has explained to me that a book distribution partnership with rank-checking software provider WebPosition Gold fueled rapid growth in his business. 6. See Danny Sullivan, “Death of a Meta Tag,” Search Engine Report, October 1, 2002. This is not to say that metadata are unimportant, just that webmasters were still worrying too much about keyword tags in particular, when Google likely ignores them. Description meta tags are still visible in many search results and are therefore worth using. A proper discussion about the future of metadata would fill a book. 7. Claire Woffenden, “AltaVista MD Resigns Over Unmetered Fiasco,” vnunet.com, August 30, 2000. The credibility of AltaVista’s claims had been challenged by a technology “critique” site, The Register. See Kelly Black, “AltaVista’s Unmetered Access Hoax,” InternetNews.com, August 22, 2000. 8. “Why the Open Directory Isn’t Open,” Traffick.com, March 30, 2000. 9. Sergey Brin and Lawrence Page, “Anatomy of a Large-Scale Hypertextual Web Search Engine,” Stanford University Department of Computer Science, 2000. Jon Kleinberg, widely considered to be the leading contributor to this generation of search technology, CHAPTER 2: A $21 Billion Afterthought: How Google Entered the Advertising Market has published many important papers on search, including “Authoritative Sources in a Hyperlinked Environment,” 1998. 10. For those interested in such issues, Danny Sullivan, “The Bumpy Road to Maximum Monetization,” Search Engine Report, May 6, 2002, archived at searchenginewatch.com, is a must-read. 11. Nick Wingfield, “Engine Sells Results, Draws Fire,” CNET News, June 21, 1996, archived at news.com. 12. Jim Hu, “AltaVista to Auction Premier Ad Placement,” CNET News, April 15, 1999. 13. Steve Harmon, “GoTo.com IPO Set to Go This Week,” InternetNews.com, June 16, 1999. 14. “Paid Search Is Here to Stay,” Traffick.com, March 27, 2000. 15. eMarketer offers a convincing summary of relevant online ad market share stats that serves to re-emphasize points made in the first two chapters here. See eMarketer, “Portals Dominate Online Ad Take,” April 24, 2007, archived at www.emarketer.com/ Article.aspx?id=1004838. Stats include Google and Yahoo now accounting for 91.9% of U.S. paid search spending and the same two companies accounting for over half of all online ad spending. 16. Robert Scoble, “Why Mahalo, TechMeme, and Facebook Are Going to Kick Google’s Butt in Four Years,” Scobleizer, Aug. 26, 2007. Rejoinders include Dave Winer, “Google and Search,” Scripting.com, Aug. 27, 2007; Rand Fishkin, “I Used to Respect Robert Scoble’s Opinion,” SEOmoz, Aug. 26, 2007. Archive locations: http://scobleizer. com/2007/08/26/why-mahalo-techmeme-and-facebook-are-going-to-kick-googles-buttin-four-years/; www.scripting.com/stories/2007/08/27/googleAndSearch.html; and www.seomoz.org/blog/i-used-to-respect-robert-scobles-opinion, respectively. 17. Indeed, the formats were similar enough that Overture contended that Google violated key Overture patents. Following Yahoo’s acquisition of Overture, the two companies settled the ongoing dispute out of court. In August 2004, on the eve of its IPO, Google awarded Yahoo 2.7 million shares of Class A Google stock as a lifetime license payment covering any relevant patents for the pay-per-click auction model. 18. Figures from comScore qSearch, as told by James Lamberti in a presentation at SES San Jose in the panel on “The Search Landscape,” Aug. 20, 2007. 19. See Kevin Ryan, “SearchTHIS! Paid Listings Under Fire,” iMedia Connection, June 7, 2005. 67 68 Winning Results with Google AdWords 20. Mike Grehan, “Optimizing for Google Universal,” ClickZ, July 9, 2007. 21. Andrew Goodman, “Holiday Time, Quality Time: Two-and-a-Half Questions About Quality Score for Nick Fox, Google,” Traffick, November 30, 2006, archived at www.traffick.com/2006/11/holiday-time-quality-time-two-and-half.asp. 22. Gothamist, “Google Eats Up More Office Space,” September 1, 2007, archived at http://gothamist.com/2007/09/01/google_eats_up.php. 23. Charles H. Ferguson, “What’s Next for Google,” MIT Technology Review, January 2005. Part II How to Play the AdWords Game This page intentionally left blank Chapter 3 First Principles for Reaching Customers Through AdWords T his chapter is intended to illustrate how small, unobtrusive AdWords ads fit into the typical user’s journey in a given online search session. (Actually, the first thing to realize is that there is no “typical user,” per se. Luckily, advanced forms of segmentation that might have cost oodles of cash in the old days are relatively easy to approximate at little cost in time and money through AdWords.) Even if you’re well versed in online advertising, this chapter is worth reviewing. In the effort to master the finer points, it’s easy to lose sight of the basics of how and why people are using Google Search, and what you can realistically expect them to do when they come across your ad. It’s all too easy to make stereotypical assumptions that there is one way of searching the Web, or one typical model for making sales to online audiences. The examples I’m about to give are intended to convey a sense of the diversity of this medium. There are many different ways to succeed, and also many ways to fail. In addition, I’ll briefly review recent developments and happenstances that have conspired to put Google and its AdWords program in the leadership position they enjoy today. Google’s visionary path towards search market dominance has created special opportunities for advertisers, but the medium also poses some difficult challenges. The following two chapters are also fundamental, as they pick right up on the basics provided here. If you happen to be super-impatient, skip ahead to Chapter 4 for my philosophy on “setting up ad groups,” but recognize that you may be missing important background by skipping ahead. Chapter 5, on how your ad will rank on the page in competition with other advertisers for various keywords, is far from basic. Because Google now ranks ads with a complicated formula, I had to devote an entirely separate chapter to it. I’m afraid you can’t really get away with skipping that one! 72 Winning Results with Google AdWords Through the User’s Eyes: Profit by Understanding Searchers’ Love Affair with Google For millions of Internet users, search is an habitual activity, performed daily. According to recent figures from comScore Networks, the average U.S. user performs 35 searches per month. The average Canadian user is a bit more active, at 40 searches per month. Keep in mind that these are averages; many users search much more frequently than that. The detailed data about user behavior gathered by panel-based web measurement companies like comScore are not actually all that easy to come by. comScore has a proprietary dataset it calls qSearch. To get access to all of the information about the complex daily patterns of search behavior, you would either have to subscribe to comScore’s premium info service or have direct access to the same search data that some employees of companies like Google and Microsoft do. Various web metrics vendors like Hitwise offer increasingly accurate, detailed information. Again, though, for Hitwise’s best stuff, you’ll pay a fee. We don’t always know what is running through a user’s mind when he or she undertakes a web search, so stereotyping or prejudging is unwise. But we do know that the purchase process is largely about navigation and usability. Web usability experts write articles, convene at corporate seminars, and do in-depth studies to assist companies in their efforts to make it easier to navigate through the website to the point of purchase. But what we as search marketers need to do is to extend the thinking about the user experience back one step, to the very first stage of finding your site listed in a search engine. Users can’t use what they can’t find, obviously. And how they encounter your listing makes all the difference in initiating a positive contact with them. To provide a feeling for the process, I’ll offer a couple of straightforward examples of a typical search-to-purchase scenario. The prospect of sitting in the same place long enough to write a book brought back painful memories of graduate school, eyestrain, and lumbar pain. A new monitor took care of the eyestrain. But oh, my aching back! I decided I’d put off my purchase of an Aeron chair long enough, but I also wondered if there were any alternatives to Herman Miller’s pricey design. Surely I could get a chair of the same quality for less with just a few minutes of research and a little shopping around. What to do? As I and millions of others do every day to solve such problems, I went to Google.com and performed a keyword search. The first query I tried—aeron-like—could have been too clever for its own good, but it turned up a fair amount of useful material. I suppose one of the wondrous things about searching the Web for information using a tool like Google is that nearly anyone can do it without a lot of specialized instruction. Google has advanced features, but no one is forced to use them. Arguably that’s a big reason why search has spread so quickly throughout the population. It’s easy. With a few keywords and a couple of clicks of a mouse, ordinary users find even the most unusual information quickly, a scenario that would have been unthinkable a few years ago. As you can see in Figure 3-1, a number of relevant results appear. The left-hand side of the page contains what are commonly known as search results. I often call these web index results because they’re pages drawn from an enormous database of pages created when Google’s spider, CHAPTER 3: FIGURE 3-1 First Principles for Reaching Customers Through AdWords Typing aeron-like into Google Search yields a list of useful and interesting results. Googlebot, crawls the entire Web. These are the pages that Google’s technology deems the most relevant to my query. (Web index results means the same thing as “natural search results” or “organic results.” Some simply call them the “free results.”) Google’s search technology uses an algorithm that uses several weighting factors. Early versions of this algorithm may have drawn on PageRank—a complex analysis of which important pages on the Web link to other pages, about which topics. (But some speculate that Google no longer uses PageRank per se.) None of the web index results are paid for by advertisers. There is no way to pay your way to the top of the regular search results, or even to guarantee that your website is included in them. Google currently maintains an index of over 15 billion web pages, and on any given keyword query, Google will determine how many pages match. On popular keyword queries like new york hotel, there are thousands of matches. In fact, on new york hotel (without quotes), there are about 22.9 million matches. On “new york hotel” (typed with quotes by a web searcher to 73 74 Winning Results with Google AdWords denote a phrase), there are about 3.6 million matches, for what it’s worth. Good luck getting your website onto the first page of results there! Anyway, since my query, aeron-like, was so quirky, it turns out that only 261 pages matched. Over time, Google seems to be pursuing a number of interrelated strategies regarding information demand and user expectations. Google is at once trying to respond crisply to what users expect to see on the page, while playing an active role in shifting those expectations over time. One interesting thing that Google has been doing recently, some believe, is ensuring that a higher proportion of the search results are informational, educational, or governmental in nature. Google may be treating obviously commercially oriented pages differently when it comes to its ranking algorithm. A major reindex in fall 2003, now dubbed “Florida,” elicited howls of protest from webmasters on discussion forums and was covered in the technology press such as CNET as well as in the mainstream press. But this is an oversimplification of what is really going on. As Google made more explicit when it formally announced its approach called Universal Search, the key today is to attempt to show search results that respond well to the user’s intent on any given search query. As search technology experts have half-joked for years, the perfect search engine would guess what you wanted even before you typed in any keywords. With Universal Search, web page results from Google’s index of the Web will remain in the forefront. But, depending on the query, Google may serve up other types of “objects” or modules on the page to show information it thinks you might want: movie showtimes; weather forecasts; local listings with maps; news results; your own desktop contents if you have Google Desktop Search installed; photos; videos; and much more. Meanwhile, on queries with strong commercial intent, there is a high probability that AdWords ads will show up, and as time goes by, users develop expectations about what types of things they should see in the designated advertising area, as well. As I’ll discuss further in Chapter 5 (on how Google ranks—and sometimes rejects—ads), if it doesn’t fit the mould of a retailer, technology service provider, or some other typical and expected type of business, users tend to squawk to Google that they are getting annoyed. Google takes this to heart, and has developed a system to disincentivize certain kinds of unconventional business models from advertising at all. The more ads Google rejects, it seems, the more money they make. Google today continues to work this paradox as much as they ever did. If anything, they’ve become more self-aware of the sources of their success. Many queries are not commercial in intent. Others are highly so. Google doesn’t really want shopping sites to appear high up in the index search results if a query is obviously informational. Where the query is commercial, Google would hope to fill a little more screen real estate with advertising. They conduct subtle experiments all the time, figuring out how to show more ads where appropriate, and less where inappropriate. Want to get a sense of whether search engines guess that a query is commercial? Try Microsoft adLab. For background research purposes, they offer a tool (Detecting Online Commercial Intention is its current name) that will tell you what the Microsoft Live Search engine thinks of a given query. On the term “Cleveland irons” (a golf club), for example, Live Search (via adLab) believes this query has a 92.6% chance of being at least somewhat commercial in nature. The first time I tried this a couple of years ago, that figure was 96%. As more data comes in for a whole range of search queries, these predictions are likely CHAPTER 3: First Principles for Reaching Customers Through AdWords to become more accurate. Among other things, they can help search engines determine how many ads it’s appropriate to show on the page for different kinds of queries, as part of their long-term planning processes. Some time ago, Google’s director of search quality, Peter Norvig, confirmed that for the purposes of organic rankings, Google might attempt to assess not only the degree of commerciality of sites, but different types of commercial intent (a catalog page versus an “About Us” page) and rank pages accordingly. It might even, Norvig told me, look for data such as how long a company has been in business, or how long a website has been operational, for cues to the quality or reputability of a given page. Practically speaking, this may be putting more pressure on commercial sites to use the pay-per-click AdWords program if they want to generate targeted traffic. But, AdWords is not a get-out-of-jail-free card for businesses wanting to promote themselves without being scrutinized by some kind of algorithm. Nowadays, Google applies to the paid listings some of the same criteria of quality, credibility, and relevancy that they apply to the organic search results. Users demand quality control no matter which side of the page a listing shows up on. There has been endless speculation about which factors are most important in Google’s algorithm. That Google uses some form of link analysis seems unquestionable, but like most leading search engines today, Google’s means of ranking websites goes beyond published formulas. Search engine ranking methodologies are a moving target—deliberately so, as marketers try to exploit knowledge of algorithms to achieve the best possible “free” rankings. To put together a strategy for improved organic traffic, today’s marketers are best served by dividing up ranking factors into “action buckets” such as good content, external reputation, site architecture, user experience, basic labeling, and the like, and pursuing the improvements that seem most likely to please not only search engines, but customers or readers. In any case, this book is not about getting your website well ranked on that elusive left-hand side of the Google Search page. It’s about paying to place your advertisements in the right-hand side of the page (and sometimes, at the very top of the page in the highlighted premium sponsor area) near those search results so that you might grab the attention of customers typing in relevant keywords. On my query, aeron-like, you can see (Figure 3-1) unobtrusive-looking text listings in the right-hand margin. These are clearly marked as “Sponsored Links.” The ad from BusinessInteriors. ca caught my attention, as did an ad from Backs, Etc., which has a strong local reputation and a convenient showroom on Eglinton Avenue in Toronto. (Due to search personalization and everevolving search algorithms, the current screen shots may not sync up entirely with your experience, or mine prior to or subsequent to writing these paragraphs. Backs, Etc. is just an example of a company that did catch my eye in real life.) I didn’t click on any ads at first, though. I read a couple of the articles I noticed in the regular search results. I also noticed that the very top link in the search results was to a page describing a new, less expensive, stylish Aeron-like chair from Herman Miller called the Mirra. That’s about as close to the definition of aeron-like as I could come. Is Google clairvoyant? Anyway, more research was still needed. I had three choices: spend all day reading articles about Aerons, zip directly to an online store and order one right away, or something in between. I chose the middle path: do just enough research to decide which chair was more or less right for me from 75 76 Winning Results with Google AdWords the standpoints of form, function, price, reliability of retailer, and convenience of purchase. Once I got those issues settled, I fully intended to make my purchase in a physical store because I feel more comfortable making some purchases in person. I point all of this out to show how easily a businessperson and an advertiser can fall into the trap of making assumptions about how people search and shop. Above all, you can get lulled into believing that everyone is naturally going to do what you want them to do: buy your stuff online at the price you’re asking, seconds after seeing your offer. This might be nice, but it doesn’t always happen. You can’t control how people behave, and often, unless you’re a large company with plenty of reliable data on hand, you can’t even guess very well about what motivates them to act on your offer. Don’t get too narrow in your thinking; just realize that your target audience—as my Aeron-like chair example shows—may behave in a reasonably orderly but ultimately untidy, unpredictable fashion. In particular, it’s safe to say that many purchase decisions take time. Consider that a sales cycle—be this a week or a year—is a normal part of doing business online, and that repetition and consistency, along with targeting, are the bedrock of any online advertising investment. If you’re looking for some confirmation of this latter point, consult Jay Conrad Levinson’s Guerrilla Marketing. It’s the bible for the fundamental principle of repetition in advertising. To succeed with online advertising you must, at a minimum, be bullish about advertising and understand why you advertise. Although you can’t totally control or predict your target customer’s behavior, you can control many aspects of your AdWords campaign, which is what makes an AdWords campaign so different from traditional advertising. Here, you can analyze your data with reasonable degrees of precision, but then again, the data can’t tell you everything. I never traffic in stereotypes like “no one shops at night,” “they’re not buying because the ad copy isn’t exciting enough,” or “everyone is looking for the lowest price.” Relax. Backs, Etc. got my (offline, walk-in) business, even though I probably never explained to them that some of my searches were done at night; even though the ad copy didn’t make me come to any sudden revelations other than that they were in business and eager for mine; even though Backs, Etc. is not a discounter but rather a trusted local fixture that focuses on good service. Back to the Google Search results screen. In addition to the text listings in the right-hand margin, advertisements appear in one other place on Google Search. Depending on the query, how much you bid, and other factors (Google doesn’t completely disclose these, nor many other, details of their formulas), Google AdWords may also appear at the top of the page in bold text on a colored background, as you can see with the “new york hotel” example query in Figure 3-2. These ads are not treated any differently from the ads on the righthand side. Your ad might show up in these spots in the normal course of your AdWords campaign. This position is all part of the same keyword auction. When I originally searched for aeron-like, some of the ads that appeared on my screen applied to Canada only, as shown in this illustration. CHAPTER 3: FIGURE 3-2 First Principles for Reaching Customers Through AdWords The phrase “new york hotel” attracts many advertisers. Google displays ads at the top of the page, above search results, as well as in the right-hand margin. A different set of ads, however, would have appeared for my aeron-like query to a user in the United States (as shown here). Different sets of ads are shown to users in different countries because advertisers have control over which countries their ads show up in. (A Google AdWords feature called geotargeting also allows advertisers to limit their ad display by geography on a variety of dimensions, including state or province, metropolitan area, or even a specific radius or shape on the map.) After my aeron-like search, I followed up on one of the articles I’d run across, a product review on Wired.com for an Aeron-like chair, the Ypsilon. As directed by the article, I went to the vitra.com website to find out more about the Ypsilon, but found it difficult to navigate, and returned to Google to search under the phrase ypsilon chair. The first thing I learned, after poring over a few regular search results, is 77 78 Winning Results with Google AdWords that the “real” Ypsilon chair is an avant-garde wooden chair designed in 1950 by Hans Wegner. A reviewer declared it to be without “faults or mendacious pretences.” The high-tech 2008 type of Ypsilon chair, on the other hand, is widely available and, at a retail price of $900 all the way up to $3,000, must generate fat margins for the designer and retailer alike. But when I typed ypsilon chair (Figure 3-3), only two or three advertisers typically showed up for the U.S. market. At later points in search history, no advertisers showed up at all. This will vary over time, especially depending on how many advertisers can keep their ads enabled on the broad match for the word chair. In Canada, there are still no advertisers on this term. No doubt this search results page will eventually fill in with more advertisers. But the point is, when potential ad spaces go unfilled on such a high-margin product, especially when you consider that advertisers don’t pay for the advertising until the user clicks on their ad, you can bet someone is missing out on an opportunity. And if you’re reading this book, that’s the kind of void you’ll be looking to fill. There are many such opportunities on Google AdWords if you know where to look. FIGURE 3-3 Few advertisers take the trouble to show up on the query ypsilon chair— an opportunity missed, since such an ad would likely attract a high-income customer. CHAPTER 3: First Principles for Reaching Customers Through AdWords Do People Really Look at the Ads on Google? In conversation I frequently hear observations like, “I never look at the ads when I do a Google search.” The observer will then typically generalize his own experience to conclude that search engine advertising is a waste of time and money. Of course, proving that people not only look at the ads, but also click on them, and then move from click to purchase, is not difficult. These days, this stereotype doesn’t carry much weight in an era when Google’s stock price has vaulted over $700. Its megabillions of annual revenues coming from ads tend to be a fairly good answer to this silly type of comment, especially given that the majority of advertisers do track whether the money they spend on AdWords converts directly into sales. The reality is, your eye often avoids the ads when you’re researching something noncommercial. But a higher percentage of people typing Dallas tax attorney would likely be looking to form a business relationship, and would be more open to the sponsored results (possibly even seeking them). The incontrovertible proof of people not only looking at, but also clicking on, ads is found in data from any given Google AdWords campaign. I’ve had the opportunity to access data from more than 300 client campaigns and to peek at briefly at least as many again. I can tell you that users are clicking on these ads, usually at a consistent rate somewhere between 1 and 10%. That is, for every 100 times a client’s ad appears next to the search results for a given phrase, it will typically be clicked between one and ten times. That’s not bad considering that users usually see between 10 and 20 search results on their screen, and as many as ten text ads. For them to click on something that is clearly marked as a sponsored link even once out of a hundred times is something of a small miracle. Google once claimed that its ads are clicked on 10 to 20 times more often than the typical online ad banner, even though such banners are becoming increasingly large and obtrusive. The disparity between search ads and banners might be even greater than that. Recent figures from DoubleClick put average clickthrough rates (CTRs) on banner ads at under 0.5%, whereas at least one of the ten Google ads that appears on a page of search engine results pages (SERPs) might be clicked on something like 15% of the time. That’s 30 times more often than banners. Eye-tracking Studies: What Can They Tell Us? Even the question of where people’s eyes are looking when they’re scanning a page of results can be answered with some degree of certainty. Some major search engines and portals, including Google, maintain usability labs that can graphically display where users’ eyes go on a screen and how long they maintain their gaze at any given place (see Figure 3-4 for an example). What users are typically trying to do is to find whatever most closely matches what they’re looking for. Different eye-tracking studies may show different results, and these are nowhere near as important as the actual clicks on your ads, along with data that tell you whether there is a return on investment on those clicks, at reasonably high volumes. I don’t need to make abstract predictions about whether a fifth- or sixth-position ad on a popular term will be seen or clicked, or whether this translates into a respectable return on investment. I see account data and financial data for clients that clearly show even the fifth or ninth ad position is big business and generates many leads at an identifiable cost per lead every day. 79 80 Winning Results with Google AdWords FIGURE 3-4 A heat map from an eye-tracking lab study by Enquiro Research. This shows where users’ eyes go on a page of search results. The question of whether users pay attention to the ads is no trivial matter and is so important to Google that they’ve paid attention to it with a fervor sometimes approaching paranoia. The history of online advertising seems to follow a pattern: first, there is a high degree of user interest and attention to advertising presented in new formats; then, a phenomenon of unconscious ignoring (sometimes called “banner blindness”) takes hold, and clickthrough rates (the number of times a user decides to click on an ad divided by the number of times an ad is shown) begin to plummet to levels approaching zero. As this happens, advertisers eventually move to devalue the advertising. As demand dries up, ad rates plummet, and publishers—especially large publishers like AOL, Yahoo, and Google who depend heavily on advertising income—find themselves in trouble. This happened following 1999 as the dot-com honeymoon ended. As ad revenues plunged and hopes of growth for online media companies were dashed, stock values plummeted. Google depends heavily on the advertising program, and they are watching closely for signs that users are becoming “blind” to these ads. So far, this hasn’t happened, but it explains why Google puts so much emphasis on relevancy and the user experience. In his Alertbox column (“Will Plain-Text Ads Continue to Rule?” April 21, 2003), Web usability expert Jakob Nielsen wonders if banner blindness will give way to text box blindness. He argues that the low-tech text ads are currently enjoying the same novelty effect that previous advertising formats did, but that they “are not guaranteed a bright future outside of their native search engine habitat.” But Nielsen does admit that the low-tech format that requires advertisers to be highly relevant might be able to hold user interest for the long term. Thus far we are not seeing any major drop-off in clickthrough rates on search ads. In part, this is because of the uniqueness of the search medium—the fact that the ads are often exactly what the user is actually searching for. Probably more importantly, Google’s ad system is ruthless about relevancy; this isn’t left to chance. CHAPTER 3: First Principles for Reaching Customers Through AdWords Why Users Love Google In Chapters 1 and 2, I demonstrated how dominant Google has become in the marketplace. Take a second to reflect, though: amidst all the competition from companies like Yahoo, Microsoft, AOL, Ask, yellow pages companies, and feisty startups that seem to come along every month, why do so many users display such loyalty to Google? Arguably, it’s what Google Search isn’t that has attracted so many users to it. The clean design and singular focus is what people want. Google has done a remarkable job of appearing clean and simple even as it has added a host of new services. The home page today looks remarkably similar to older versions. In the portal era, emboldened by inflated stock market valuations, every search company wanted to become the next Yahoo, which had grown from a small directory service to a large media company with many offerings. Unfortunately, also-ran portals were not compelling to users. They were simply ad-cluttered, unfocused messes. Taking advantage of the Wayback Machine at archive.org, you can look at old screen shots of Excite.com (see Figure 3-5) and AltaVista.com (see Figure 3-6) and understand why these companies didn’t win the search wars. (These screens actually fail to do justice to how banner-cluttered many pages on these networks became, and they predate by a couple of years the desperate shift by online ad agencies and publishers to oversized, page-dominating banners.) FIGURE 3-5 Excite’s home page got too exciting for its own good. 81 82 Winning Results with Google AdWords FIGURE 3-6 Portal clutter got the better of search specialist AltaVista. As other search sites admonished beleaguered users to shop until they dropped, Google quietly entered the fray in 1998 and gained momentum in 1999–2000 with a simple search box and new-generation search technology. AltaVista, one of Google’s chief rivals at the time, lost market share quickly. Evidently, it had changed courses too many times. What a terrible time to give up on search! Google filled the void and soon took over as not only the leading search engine, but one of the world’s best-loved brands. In May 2000, AltaVista tried to copy Google’s simple layout with a new, standalone site for “search enthusiasts” called, of all things, Raging Search (see Figure 3-7), but it was too late. Google had already won too much mindshare. Users flocked to Google’s oasis of simplicity and relevance for good reason: they wanted to search, not be shouted at. If an advertiser shows up near the search results and delivers something that’s at least as relevant, users seem to see that as a fair compromise. But Google has never taken this for granted. The founders of the AdWords program had the vision to forge that compromise from scratch, taking Google from a company with virtually zero revenue to one raking in close to $21 billion a year from advertising without alienating that notoriously fickle user base. Unbelievably, as recently as 2005, media analysts (including Forrester Research CHAPTER 3: FIGURE 3-7 First Principles for Reaching Customers Through AdWords AltaVista copied Google’s look and feel with Raging Search, but no one noticed; they were too busy getting acquainted with Google Search, the new whiz kid on the block. analyst Charlene Li) peddled the faddish view that the “big portal companies” like Yahoo would eat Google’s lunch sometime soon as Google failed to grow beyond its “one-trick pony” status. Obviously, this prediction was wildly off base. More Thoughts on User Intent Even a search that’s informational in nature can ultimately result in a commercial transaction. A computer user hears through the grapevine that Windows XP has a security vulnerability whereby any user with an old Windows 2000 CD can circumvent the admin-password-protection on Windows XP. Is it true? She searches on a specific query such as windows password vulnerability or a broad term such as windows 00 to find out. In the midst of her research, an interesting, free “Windows tips” newsletter, Brian’s Buzz by Brian Livingston, is advertised (see Figure 3-8). She clicks on the ad and signs up for the newsletter. As it happens, Livingston’s subscribers are typically so delighted with the free information that a predictable percentage of them turn into paying subscribers. Advertising near search queries that are largely informational in nature nonetheless produced a measurable return on investment for BriansBuzz.com. Given Google’s experience of winning with clean, uncluttered search, more search engines are erring on the side of conservatism when it comes to how much advertising they show. This can pose a challenge to many advertisers who won’t be getting the volume they’d like to see. 83