Searching for trust
Note: The following contains a recollection that is hotly disputed byfriends -- including the marketing person and the chief technology guruinvolved -- whose memories of these events are as vivid as mine. The mainissue seems to be whether we ever seriously discussed the first version ofthe plan discussed below. Also, whether I was a pig-headed, sanctimoniousa-hole about it. Actually, there's widespread agreement about the latterpoint.
About four years ago, I should have quit my job.I was working at a company that created one of thefirst search sites on the Web. Although I was VP ofStrategic Marketing (the "strategic" indicating thatI didn't actually know anything), I lost theargument over letting sites pay us to have their URLguaranteed a particular position in the searchresults. For example, if "BigBoy's Texas Getaways,Inc." paid us some thousands of dollars, we'dguarantee that if anyone did a search on the words"cowboy" or "dude ranch," their site would be listedat #6. Pay us more, and you could make it all theway to #1.
Our company ran the numbers on how many search terms we'd have to sell inorder to become early Web gazillionaires. Fortunately, sanity prevailed andwhen we launched we put a gif next to the paid-for ads that flagged them as"Preferred," which to my mind was better than nothing...barely. For acouple of years afterwards, we were still taking lumps in the press.
The less questionable variation on this plan has long been standard forsearch sites: BigBoy's can pay the site to stick a banner ad onto the searchresults page whenever someone searches for "cowboy." Because it's a bannerad, there's no intrinsic deception. And about a year ago, goto.comintroduced a site whose "business model" is based precisely on the idea we'dfirst discussed [or did we? -- see the note at the beginning of thiscolumn]. While their home page doesn't tell you that this is a commercialyellow pages, the search results tells you how much the advertiser paid tobe listed. For example, a search for "cowboy" brings fogdog.com to the topof the list, which cost the advertiser $0.16. The #2 slot went for $0.15,and #3 for $0.12.
A new report by Danny Sullivan at Search Engine Watch makes it clear just how far this trend has gone.
At AskJeeves, for example, a company can ensurethat it will show up in Jeeves' highly-constrained,visible search results list. According to Jeeves'fact sheet, it seems that a company can buy aquestion. Sullivan, however, says that theconnection is less direct than this, ensuring thatBigBoy's listing might show up if someone asks"Where can I go for a rootin' tootin' good time?"even if the BigBoy site didn't buy that precisequestion. Either way, BigBoy is paying forplacement.
Sullivan's example is "What should my bloodpressure be?" which puts a link to onhealth.com as#1 and then has a special section underneath listingother onhealth.com pages. There is no indicationthat any of this is paid advertising. Sullivan saysthat Jeeves says that that's ok because the"editorial staff" at Jeeves maintains qualitystandards. Yeah, and how many heart specialists areon staff to judge which health site is best?
AskJeeves isn't alone in this. At Yahoo, for $199you can be certain that one of their paid "editors"will at least look at your site, although there's noguarantee they'll include it in Yahoo. (Note toYahoo: If you at last index my site,www.hyperorg.com, my $199 will also buy you certain"favors," if you know what I mean.) And Inktomi,one of the leading providers of the technologyinfrastructure for search sites, will guarantee thatyour site will is re-crawled every 48 hours.
On the other hand, my favorite search site (thesedays), www.google.com, handles these issues well. Ifa company buys a word, it gets a text-only ad atthe top of the list, clearly identified as an ad.(Try searching for "Ford" as an example.) Further,ads shown in the right margin display a green barindicating what percentage of people have clicked onthem, enabling us users to "vote with our clicks,"so to speak. (Search for "flowers" to see anexample.) But, Google draws the line at monkeyingwith the rankings. It uses complex algorithms thatassess factors such as how many other sites link toit. This is in the user's interest. Sites that tryto trick a user into going to sites because someonepaid them are working against the user's interest.While there may be (grudgingly admitted) somebenefits to letting companies pay for placements,there is no reason, except avarice and the abrasionof the ethical sense, to hide the fact that ourattention and trust have been sold to the highestbidder.
Danny Sullivan's Search Engine Watch:
http://searchenginewatch.com/AskJeeves brochure to advertisers:
http://corporate.ask.com/pdf/Jeeves_Answer_Link.pdfInktomi FAQ:
http://www.inktomi.com/products/portal/search/submitFAQ.htmlDavid Weinberger is editor of Journal of the Hyperlinked Organization.