Biznology Blog: June 2005

« May 2005 | Main | July 2005 »

June 28, 2005

Google's Next Step to Personalized Search

Not long ago, I explained how personalized search will send rank checking the way of the dinosaur. Google has now taken the next step in personalized search. If successful, this could be the first step down that road. If every searcher gets different results, then no one is really #1. Andy Beal has another take—will personalized search make spamming much tougher? I hope Andy is right, although spammers have responded to each escalation in the arms race thus far. In our book, Search Engine Marketing, Inc., Bill Hunt and I explain how to avoid the tricks and stick to search marketing that avoids all spam techniques—perhaps personalized search makes that approach even more critical.

Posted by mikemoran at 3:35 PM | Comments (0) | TrackBack

June 23, 2005

Why is Personalization So Difficult?

Everyone has used Amazon, and most like the way Amazon personalizes pages for each customer. By paying attention to the kinds of books you buy (and what others buy), it offers (mostly) helpful suggestions about what you might want to buy next. So where are all the other personalization successes? Why is personalization talked about so much and so rarely done well?

At first glance, personalization seems like that rare win-win proposition—buyers get a better experience while sellers get higher revenue. And Amazon shows that, when done right, it can work beautifully. But it must be hard to do right, because Amazon is not only the most-cited example of successful personalization, sometimes it seems to be the only example ever cited.

What is the secret of Amazon's success? Several related factors are probably the most important:

• The frequency factor.Book buyers return to Amazon on a regular basis, so Amazon gradually collects enough information to make personalization effective. If you are selling cars, it is unlikely that you will ever develop a pattern on your customers that holds up across purchases made years apart.


• The "no hands" factor. Unlike many businesses, Amazon's book business does not require customers to answer any questions or perform any actions solely to inform the personalization engine of their preferences. Books are bought (or examined) frequently enough that no questionnaire need ever be filled out. Amazon expects customers to register, but uses cookies so they don't need to sign in unless they are making a purchase. If customers visit your Web site less frequently (such as for auto sales), they will be more reluctant to register and the cookie you dropped on their last visit was likely stored on their previous computer (and is lost to the current visit).


• The stability factor. People tend to purchase multiple books on the same subject. Folks who buy one seafood cookbook are highly likely to buy another. Similarly, many readers buy every book from a particular author. Most businesses are not so lucky. A consulting firm can scarcely predict a customer's future needs based on past purchases because their needs change as they are fulfilled.

From these factors, you can see how it is in some ways easier for Amazon to personalize its bookstore (and music store and a few others), while your business may not have those advantages. But are these factors really the root of Amazon's success? Hardly.

The lion's share of Amazon's success does not derive from personalization. It is no more easy for Amazon to personalize your experience when you buy an electric shaver from them than it is for you to personalize your Web site. Shavers are bought once every few years and no amount of data could ever be built up to figure out what buyers want next. So how has Amazon become a leading retailer for products where personalization does not work all that well?

Amazon has two bedrock practices that guide its success, which they apply to their personalization features, but also to everything else they do:

• An obsession with the customer experience. Amazon works mostly because it is easy, reliable, and fast. When you shop at Amazon, you can usually find what you want, you can easily complete your purchase, and it shows up when you expect it to. Amazon is not usually the lowest-priced, but it is usually the most convenient. Amazon succeeds in large measure because it constantly tinkers with every aspect of both its Web user experience and its offline customer experience so it can improve every day.


• A relentless focus on metrics. It's been said that Amazon can change a font for some text on a key page and in a few hours know whether more folks are clicking further into the sales experience or not. If they are not, then the font can be changed back. If the clickthrough rate is up, then the change was a success. But Amazon's testing is even more clever—they can change that font for only certain visitors (it's personalized, remember), and roll the change out to everyone if warranted.

The bottom line is that personalization is what everyone notices about Amazon, but it is probably the hardest thing to copy. What every business should learn from Amazon is how to put the customer experience above all else and to measure, measure, measure. Your business may require measurements that are harder to do than online sales (we discussed measuring Web conversions in a previous article), but that should not prevent you from keeping track anyway. So, the next time you visit Amazon, pay attention to what it does beyond personalization—that maybe those are the best practices to emuulate on your site.

Posted by mikemoran at 7:51 AM | Comments (0) | TrackBack

June 22, 2005

Google's New Free Inclusion Program

Amidst the debate among consumer activists over paid inclusion, Google has announced a new free inclusion program, called Google Sitemaps, which offers many of the benefits of paid inclusion programs at no cost to the search marketer.

Posted by mikemoran at 9:23 PM | Comments (0) | TrackBack

June 16, 2005

Who's Watching the Search Engines?

Read June's Biznology newsletter and see what you think about search engines and the way they treat consumers. Check out this opinion piece based on the conference held by Consumer Reports WebWatch just last week.

Posted by mikemoran at 8:20 PM | Comments (0) | TrackBack

June 7, 2005

MSN Search Preparing to Enter Paid Placement Fray

According to Heather Lloyd-Martin, MSN's upcoming entrant into the Google-Yahoo! paid placement battle is more similar to Google's approach than that of Yahoo!

Since MSN has already dropped use of Yahoo!'s organic search technology, observers have been waiting for the other shoe to drop—when will MSN Search launch its own paid placement service?

We don't know the answer to that one yet, although it seems that mid-2006 is the latest that Yahoo!'s former Overture service will be dropped. Obviously, this is bad news for Yahoo!, because MSN Search still controls over 10% of the search market, so Yahoo! is losing a significant chunk of market share for its paid placement offering. Because Ask Jeeves and AOL Search both use Google's AdWords paid placement offering, Yahoo! has no significant paid search partners, quite a turnaround from how Overture started its business when it partnered with everyone (it seemed) but Google.

But Heather made news with a comment by Matt Lydon of MSN Search in a panel discussion, where it was revealed that MSN will use an approach closer to Google's than Yahoo!'s for search ranking. Yahoo! uses the tried and true "highest bid takes the top spot" approach, whereas Google uses the combination of bid and clickthrough rate to pick the winner. This has been very successful for Google for a number of reasons:

• While Yahoo!'s approach maxiizes bids, Google's approach maximizes Google's revenue.
  
• The use of clickthrough rate somewhat increases the relevance of paid placement ads.
  
• Clickthrough rate also reduces the need for editorial oversight, because poor ads will die a quick death from going unclicked, whereas Yahoo! must weed out high bidders with poor response.

So, if Google's approach has been so successful, why hasn't it been copied? Search insiders have long discussed Google's patent on the ranking algorithm as the impediment, but apparently MSN Search believes they can go Google one better. Matt Lydon, as quoted by Heather, says that MSN's algorithm will use bid, clickthrough rate, and audience. Evidently, MSN believes that this approach not only evades Google's patent but also provides better results by introducing a demographic element that would appeal to traditional advertisers and provide more targeted results.

Time will tell if this is a better mousetrap, but search marketers should think about what kinds of demographics they's like to pay for as MSN's offering gets closer to launch.

Posted by mikemoran at 9:47 PM | Comments (0) | TrackBack