Your phone rings. A good friend is calling, more excited than you’ve heard her in months. “My book is on the home page of Amazon! I can’t believe it. My book was just published last week and already it’s on Amazon’s home page!” Exciting? Maybe not to someone who knows how Amazon works. Your friend has seen her book on her version of Amazon’s home page, but a closer look shows it under “Items Recently Viewed.” She views her book’s page each day to check any new reviews. Based on what books she looks at, Amazon thinks she’s very interested in buying this book and places it on her home page. But it may not be on anyone else’s Amazon home page.
That’s how personalization works. Each person sees something different even though everyone is looking at the same page. We’re used to seeing personalization on Amazon, but now it’s coming to a search engine near you.
Inside Personalized Search
So how can search be personalized? By showing each searcher different results. Just as Amazon shows different content on its home page for different people, a search engine can show different content on its search results page, even when two searchers look for the same keyword.
Why do this? Money. Personalization can be lucrative for search engines. If personalized results are more relevant results, then searchers are happier and search more. And if search marketers can target ads to the right people, they’ll pay higher per-click prices.
Search engines have personalized paid results for years, using geographic targeting. With geographic targeting, a furniture store, for example, that delivers within 25 miles of its location can purchase the keyword furniture but ask the search engine to show its ad only to searchers within the delivery zone. Search engines check the geographic location of the IP address for each searcher’s computer to decide whether it is within the geographic zone or not. Paid search marketers can set geographic limits on city or zip code boundaries, or sometimes even by longitude and latitude coordinates.
Search engines are now extending personalized search beyond geography. MSN Search pioneered personalization using searchers’ demographics; all search engines will eventually offer similar programs. With demographic targeting, search marketers can raise their bids for searchers based on gender, age or other characteristics. So, you can raise your normal per-click bid 3 percent for women over 65, if that’s your highest-converting market.
But how do search engines know which searchers are women over 65? They need searchers to tell them. That’s why Google, MSN, Yahoo and other search engines are racing to provide services that entice searchers to identify themselves. Whenever people register with one of these companies, they provide demographic information that the search engines can use to personalize searches.
In personalization parlance, demographic targeting is called explicit personalization, because it’s based on information explicitly provided by the Web user, such as age. Expect search engines to also engage in implicit personalization, changing search results based on searchers’ behavior, such as what kinds of Web pages they look at.
Implicit personalization may eventually prove more valuable than explicit personalization, because so much more information can be gathered. Search engines can observe which results people click when they search, discerning patterns that allow them to rank their favorite kinds of pages higher for all their searches. If a particular searcher regularly clicks product reviews rather than manufacturers’ specs, Google could begin to rank product reviews higher when he’s searching for product information.
Search engines have other ways of observing implicit behavior. Google can analyze the message text of its Gmail users to see what subjects they write and read about. Yahoo can look at the keywords used by its search toolbar users, and even see what pages they look at. You should expect search engines to use this information to personalize search, both for paid and organic. Search engines are always looking for ways to improve relevance – the match between searchers and content. High relevance means the search results correlate closely with what the searcher has in mind. For 40 years, search engines have improved content analysis to increase relevance. Personalized search concentrates on the people side of the relevance challenge instead.
Inside Personalized Search Marketing
Now that you understand the basics of personalized search, you may want to know how search marketing will change.
One change is obvious. Personalized paid search bidding is more complex, because search marketers must consider geographic location, age, gender and other demographics when they make their per-click bids. Instead of different bids for every keyword, now you need different bids for the same keyword.
You should raise your bids for targeted demographics only because they convert at higher rates. To bid effectively, your Web metrics system must track conversions by geography and by demographics, not just by keyword, and your bidding software must adjust based on those metrics.
Less obvious changes will confront us when search engines begin applying personalization to organic search. Search marketers have always wanted to achieve the No. 1 ranking for their favorite keywords. But what does it mean to be No. 1 in a personalized world? If the organic results are personalized, then different searchers get different No. 1 results. Your excitement at being No. 1 will be no more warranted than your author friend’s glee at making Amazon’s home page. In a personalized search world, every site can be No. 1 with some searcher sometime.
Widespread personalization will doom traditional rank checking. The question won’t be, “Does my site rank No. 1?” but rather “For what percentage of searchers does my site rank No. 1?” or “What was my average ranking yesterday?”
And who can answer those new questions? Only the search engines themselves. Only MSN will know where your pages ranked for every search performed with their search engine, so only MSN can tell you. Will the search engines provide that information for free or will they charge you for that analysis? Will search engines tell you the demographics of the referrals that come to your site? Time will tell.
Optimists also believe personalization will reduce the problem of search spam. The thinking is that spam is all about content, so that personalizing results based on searchers makes it exponentially harder to spam the search results (because spammers must then fake their content for many kinds of searchers). By increasing spammers’ costs for the same number of searchers, it takes part of the profit out of these unethical techniques.
No matter its effect on spam, savvy search marketers must stay on top of the personalized search trend – it’s the biggest change in search marketing since paid search. If you focus on who your best customers are, and you craft your content to match, you’ll be ready when the personalized search revolution breaks out.
MIKE MORAN is an IBM Distinguished Engineer and product manager for IBM’s OmniFind search product. Mike is also the co-author of the book Search Engine Marketing, Inc. and can be reached through his website (mikemoran.comTaking It Offline If baseball is the thinking person's game, then online advertising...