Increasing your traffic from Google
and other search engines
Dean Blobaum
Online Marketing II panel
AAUP Annual Meeting
St. Louis, MO, June 24, 2003
How do most of your website visitors actually get to information on your website and what, if anything, does that imply about the work of electronic marketing?
Nailing down a solidly reliable number for how people get to our site is difficult, but the evidence suggests that at least 70 to 75% of the Chicago Books Division website traffic comes from search engines and of those visitors referred by search engines about 90% come either from one of the Google search engines or from a web search done on another site that utilizes Google's results. What are these other sites use Google search results? Several, including most importantly Yahoo and AOL.
In other words, there is nothing more important for traffic than search engines, and there is no more important search engine than Google. So, I will talk today about strategies for increasing the visibility of university press pages on the Google search engine, which will also include strategies that help with others.
So, first things first, is your website in Google?
Probably at least one page of your site is spidered and indexed by Google—your homepage if nothing else. But every page on your website should be in Google. Have you checked? At Chicago we knew for quite some time that Google didn't index all our webpages. Specifically, our book product pages were missing. Why was that? We assumed it was because those pages are dynamically generated and search engines often avoid dynamically-generated webpages. But then we heard that Google spiders and indexes lots of dynamic pages—so why not ours? Back in December the Google robot finally visited all our book product pages and we began to see traffic from Google searches to those pages.
I have a chart here to show the difference this made in our site traffic.
Why did the Google robot finally come calling to those pages? I did three things and I don't know which if any of them actually had an effect:
To find out how many of your pages are indexed by Google just use the syntax allinurl: and the relevant part of your URL. For instance, allinurl:www.press.uchicago.edu finds all the Chicago Books Division webpages in the Google index. Adding the folders just for our book product pages, allinurl:www.press.uchicago.edu/cgi-bin/hfs.cgi/00/ yields our book product pages indexed by Google.
How does Google work?
Every search engine has a strategy—an algorithm—for trying to serve up the most relevant webpages first for whatever search terms the user enters. Estimates of the number of factors in the Google algorithm range from 25 to 150—and many of these factors are common to all search engines. These include things like the text on the webpage—duh—and the words in the title tag of the page.
In my humble opinion book product pages should have the author and title of the book in the title tag of the webpage. Makes perfect sense to me, but it's not universal among publishers. And interestingly, Amazon and B&N.com have only the book's title—not the subtitle and not the author. Should we view that as a model or as a point of vulnerability? At Chicago, we see a significant number of people arriving from search engines where they have used the name of an author; so, having the author in the page title seems like a good idea. (Here is an example book product page.) Author, title, subtitle—use some combination of those three elements. And don't add extra words within the title tag, like the publisher name. What for? Those extra words will lower that page in the plausible search results. And most definitely you don't want a generic page title for every book product page like "XYZ University Press Book Detail" or any other generic title—that's completely useless for generating search engine traffic. The title of the webpage is very important for search engine results, so make sure it's done right.
The content of the book product page should be as textually rich as you can make it, with promotional copy, review quotes, and table of contents. Google likes content. Give a moment's thought to the important search terms. For example, this book on the architecture of Frank Lloyd Wright. What do people search on when they're looking for stuff on Wright? They're not searching on "Wright" and they're not searching on "Frank Wright" they're searching on "Frank Lloyd Wright." You will get more visitors from search engines if you use the full name more than once on the book page.
The unique thing that Google added to the standard search engine algorithm—and still the key to the quality of Google results and hence its dominance of the search engine market—is Google's analysis of links that are made to webpages. Google's strategy is: people have found and linked to the pages that other people are searching for. The more links to a webpage the greater weight is given to that webpage in search results. And if those links are coming from webpages that have greater weight, that in turn gives more weight.
Every month or so the Google computers analyze the linking structure of all the webpages Google knows about and they assign a ranking number to every webpage that has at least one link pointing to it. The ranking number for a page is called its PageRank.
To climb higher in any given Google search results—everything else being equal—a webpage has to boost its PageRank. It gets higher PageRank by getting linked to by more webpages. A webpage that is itself highly ranked will confer more PageRank on its links than lower-ranked pages. So you want links, especially high-quality links. Links can be from your own site or from outside your site. That's important. It's the external links that confer PageRank on your webpages, but the internal navigation links on your webpages pass it around your site. Good internal navigation is essential for getting PageRank to all your pages.
How do you know the PageRank for each of the webpages on your site? Of course, only Google really knows. But there are two ways you can see a representation of PageRank, which helps get a sense of the landscape.
If you use the Google Tool Bar (which you can get at toolbar.google.com) you can see a graphic representation of the PageRank of any webpage you visit. PageRank is always of a page—not of a website. When you roll your mouse over the green bar you get a numerical ratio, for example 8 of 10. How accurate a representation is that? Literally, not very, since there's a whole lot more than 10 pages that make up the worldwide web. But it's a visual aid.
Another way to see PageRank in action, is to go to any category in the Google directory (directory.google.com) where you will see a graphic representation of the PageRank associated with each webpage listed in that category. As an example, here is a category for academic publishers. You can order the category by PageRank. How accurate is that? Well, it's relative position of these pages, which are the homepages of these publisher sites, many of whom are represented in this room.
Using links to boost PageRank
It's easy to make either too little or too much of PageRank when it's displayed in this popularity contest fashion. So, let's bring it back to real world searching and return to our book with the title The Architecture of Frank Lloyd Wright. When someone goes to Google and searches on "frank lloyd wright" I want our webpage for this book to show up as high as possible in the search results. That's a challenge because there are not only lots of books on Wright but also lots of websites about him and his work. I'd like to see our book higher than other books on Wright, and in particular I'd like to see our book higher than all the online retailer pages for our book—I'd like us to get the first shot at presenting our book to a customer.
It was the author of the book, actually, who recently issued that challenge. He said the book turned up as search result #193 when searching Google with the terms "Frank Lloyd Wright." He assumed we could do something to improve that.
What could we do? We already have "Frank Lloyd Wright" in the title of the book and so in the title of the page. We took a look at the copy on the page and tweaked it a bit—changing a few instances of "Wright" to "Frank Lloyd Wright." That should help a little. How else can we improve our results? The only other option is to get links to this book product webpage and so increase its PageRank, which will raise it in the results.
Is it easy to get links to a book product page? Of course not. But it's easier to get links to an excerpt or feature of some kind. So we created a little excerpt from the book featuring two famous Frank Lloyd Wright homes in the vicinity of the University—the Robie and Heller Houses. A few weeks ago we suggested the URL for that excerpt to Yahoo, for inclusion on the directory page that lists architectural highlights in Chicago. Yahoo accepted it for the directory. A Yahoo link confers significant PageRank, and the excerpt has a link to the book product page, so the PageRank of that page should increase, leading to something higher than the #193 our disgruntled author found.
That's a strategy for one book, one excerpt, and one set of search terms. Use that strategy multiple times and—because of the way PageRank gets passed around a website by its internal navigation—this becomes a strategy that can raise the visibility of each and every one of your webpages when it gets included in any Google search results.
What links help raise PageRank most? Those from higher-ranked websites, which are—most important for us—human-edited directories like Yahoo and the Open Directory Project.
Yahoo first. How many links are there in Yahoo to pages on your website? There should be at least one—to your homepage. But Yahoo will link to other pages in your site if they find them worth linking to. If you have excerpts, interviews, etc. on your site you can get Yahoo links to some of them. At Chicago we've tried to get Yahoo links to many of the book excerpts and other feature content as we add it to the site. Yahoo currently has 31 links to Chicago book division pages. (And a tip of the hat to our Journals Division who have 30 Yahoo links to the journals site.) How many links could we get? Well, Random House—today—has 334 Yahoo links into its website. There's a worthy goal. The syntax for this search is simply u: followed by the relevant piece of your URL.
Second, you should try to get links in the Open Directory Project. What's the ODP? The secret weapon of the web, but easier to find than weapons of mass destruction at www.dmoz.org. It's the open source equivalent of Yahoo and second only to it in scope and quality. The directory can be licensed, for free, by any website that wants to utilize it, so long as they abide by the terms of the license. And many sites, particularly search engines, have licensed it. The Google directory is ODP; the Netscape directory, ODP; the AOL directory, ODP; AskJeeves, ODP. And lots of specialized search engines you and I have never heard of. Definitely, submit your stuff to the ODP. You'll then find it propagates to other sites. The Chicago books website has 36 links in the ODP; Random House about 180.
(But just as important—non-commercial plug here—become an editor for the ODP. You're intelligent and have good judgment, you know how to write a listing, and no doubt you have some special obsession. Become an ODP editor in whatever subject area you know best. Find the subject of your interest in the ODP directory, look at the bottom of the page, and if it says, "Volunteer to edit this category," consider yourself a contender and apply. Your ODP editing experience will give you some insight into how edited directories function, and is certain to improve your ability to get listed in those directories.)
Links in Yahoo and the Open Directory Project will contribute significantly to PageRank, but any credible link out there on the worldwide web will add some PageRank. Long, long ago in the early days of the web, there were no search engines, and many people built webpages of relevant links. They're still doing it. And Google, at any rate, still considers those links relevant.
[The next fifteen paragraphs were not delivered at the AAUP panel; there I went directly to the conclusion.]
I'm going to take a little detour in order to show you a difference in how Chicago has put content out there and how other university presses have done it—Chicago tends to build all its excerpts and editorial content drawn from books as freestanding pieces.
Here is an interview on Princeton http://www.pupress.princeton.edu/chapters/i7442.html
and here is an interview on Chicago: http://www.press.uchicago.edu/Misc/Chicago/562166in.html
When you see the interview on Princeton's site you cannot but be aware that this is the publisher's website. Whereas Chicago downplays that fact. The Princeton interview is (I think) optimized for easy of handling and to support the sale of the book. The Chicago interview takes longer to build and is optimized to get links from outside the website. Princeton can put up more of these interviews than Chicago can in the same amount of time; Chicago can more easily get links. Which is a better strategy? I don't know. Google likes both content and external links.
End of the detour.
Ephemeral links and bloggers
OK, to get back on track: the linking I've talked about so far are what can be called permanent links. Links in directories and many other webpages can be expected to endure indefinitely. I also want to talk briefly about what I'll term ephemeral links. (Truthfully, one would be crazy to call anything on the web permanent, in the usual sense of the term. And most ephemeral links actually but not too obviously become permanent links. But anyway.)
Ephemeral link sites are websites that add new links on a regular, often daily basis, and often are news-oriented sites. Ephemeral link sites add new links and drop off older ones. Arts and Letters Daily is an excellent example of an ephemeral link site. Also see Sci Tech Daily, Human Nature Daily, and Arts Journal.
(Another way to draw the permanent/ephemeral distinction is to say that there are two primary ways that links can be organized—by subject or chronologically.)
A link on ALD can yield thousands of readers—up to ten thousand in our experience.. No other newsy site pulls like that—in our experience. But links on other ephemeral link sites are worth hundreds or maybe a thousand readers.
Also in the ephemeral link category are blogs (or web logs). We have seen enough traffic from blogs to begin to proactively go out and try to snare some of it. In our experience there are some blogs that can send a few hundred or one-to-two thousand readers your way. Many other blogs will send you a dozen readers or so. Bloggers cross-fertilize: once you are on one blog you stand a chance of getting on another.
I'll illustrate this with an example:
A year ago, on June 3, 2002 we put on our website an essay—Paul Robinson's "The Philosophy of Punctuation" drawn from a collection of his essays we published. Within eight days it had been visited over 10,000 times and by the end of the month over 17,000 visitors had seen it. The first and most important link it got was on June 7th on Arts and Letters Daily—that link was responsible for just about exactly half of those 17,000 readers.
Within a week of the essay appearing on ALD, 40 bloggers had added a link to the essay in their blogs. The blog links varied greatly in how many visitors they drove—many of them brought us only a dozen visitors or so or the course of the month, the best brought us over 1,000.
How did we find those 40 blog sites that had listed the link? By looking at the daily traffic logs for one thing. But there are also some tools that help. The best is Blogdex (from the MIT Media Lab). You can search Blogdex for a specific URL; doing so we currently find 55 links to the Robinson essay.
The blogger links to Robinson were probably responsible for somewhere around 5,000 visits in that month of June to the essay. The essay still gets about 1,000 readers each month. It got a link in the ODP and a number of other links on the web, which keep the visits coming. Currently the Robinson essay is #5 on Google for the single word "punctuation"—it would be nowhere near that high except for the links that raised its PageRank
It is reasonable to think that Google and other search engines will make increased use of blogger-created data. The fact that Google bought Pyra Labs—makers the first software to make blogging easy enough for the rank amateur—suggests that Google maybe be preparing to make use of the human intelligence that goes into blogging. So, it's a good bet to keep up with bloggers.
Conclusion: how do you maximize your place on Google?
More information:
Google's information for webmasters
Searchenginewatch.com is a good resource
Think blogs are all self-referential prattle? Take a look at wood s lot
My other AAUP presentations on electronic marketing of university press books:
"Using editorial content to market books on the university press website" from AAUP Web Marketing Workshop, 1999
"Promoting and Selling University Press Books on the WWW" from AAUP Annual Meeting, 1998
"Marketing Your Books on the WWW: Measuring Success" from AAUP Marketing and Sales Managers Meeting, 1998
Copyright note: © 2003 by Dean Blobaum. All rights reserved. This text may be quoted in accordance with the Fair Use provisions of the US Copyright Act. It may be archived and redistributed in electronic form, provided that no fee is charged for access and provided that this entire notice is carried and the author of the review is notified. Archiving, redistribution, or republication of this text on other terms, in any medium, requires the consent of the author.