Jim Millot reported for Publishers Weekly today that MediaBay was shutting down after spending millions of dollars working on an Internet based distribution system for audio books. They made what I believe is a classic mistake for any Internet business, diluting their original domain (mediabay.com) with multiple spin-offs: soundsgood.com, radioclassics.com, audiobookclub.com, radiospirits.com and whenradiowas.com. I doubt this dilution had a major effect on the outcome, because none of the sites showed any real legs on Alexa is concerned, but it couldn't have helped.
I've been fortunate enough to talk with some top people from all manner of publishing companies, and if one thing comes through, it's that their Internet efforts are managed by non-Internet savvy people. In some cases, the entire focus is on aesthetics, and the designers who do this work for top dollar would be shocked to hear there even exists such a thing as search engine visibility. They think web traffic is only supposed to come from people typing the name of the publisher into their browser. And the funny thing is, if that's what you design for, that's what you'll get. I recently saw a beautiful literary press site designed 100% with graphics, there weren't even alternative text tags. If the domain name hadn't been the publisher name, it would have been entirely invisible to Google.
Other publishers have a decent sense of text content, what to put on a website, but figure the more websites they put up, the better off they'll be. Creating a new website for every title is a losing proposition, both in terms of maintenance costs, and in terms of drawing visitors. If a domain name for every title is desirable for branding purposes, fine, but redirect them all to the publisher's main site with 301 permanent redirects.
What really drives me nuts, though, are the blank expressions (or those long telephone pauses followed by "come again") in response to basic questions about the website usage statistics. Every web server maintains logs that are stuffed with far more information about visitors and how they arrived than even I have the patience to digest. But for a publisher to say that their website is a part of their overall business strategy, much less a major part, and not to know anything beyond the number of page views a month, is almost criminal. It would be like a publisher with many books in print only paying attention to the cumulative sales total each month, and instructing the printer to print more copies of random titles in hopes they might be the ones that are selling. If you're going to spend money on a website, not setting aside a couple minutes a day to check what's working is no different than handing out book contracts to random authors and hoping it somehow all works out.
The topline number that most publishers are at least familiar with is page views, the gross number of web pages the server coughs up in a given period of time. This number is almost entirely worthless on its own, the vast majority of the traffic may be false impressions or misdirected. All of your traffic may come from your publishing company being voted the worst site on the web five years running. If you get excited about page views, the ones to watch are your order pages. How many views are they getting, and how many of those views translate into book purchases? Simple tweaks to order pages can have more effect on sell-through than all of the aesthetic improvements you can dream of on an infinite budget.
But my two favorite web statistics to watch are referrers and key phrases. Keeping an eye on referrers, the places your visitors are coming from, tells you how much of your traffic is legitimate. I generally check out every new referring site I see show up in the statistics, just to keep current with what it is that people are finding interesting and useful on my site, useful enough to tell somebody else about it. The key phrases tell me how people are finding my site through search engines. I rarely look at key words (single word totals) but they can also be valuable in tracking how site development is affecting traffic over time.
For example, I just checked the year-over-year change in my top five publishing related keywords from the second Monday in July. The main keyword, "book" was up modestly, from 225 search appearances last year to 300 this year. The next most popular, "publishing", was up to 125 appearances this Monday from 85 a year ago Monday. The plural "books" was doubled, from 35 last year to 73 this year, but "published" was down from 42 queries a year ago Monday to 26 three days ago. For the keyword "publish" however, there were 52 visitors this Monday versus 38 a year ago.
This is the first time I've checked the top keywords this year, and it puts me in a good mood because it means this blog hasn't flatlined for search engine traffic yet. I'll check later using a whole month worth of totals, but I figured the small numbers would be easier to grok in a blog post. One of the problems with writing so much about the same subject is that my old blog posts compete with new blog posts for visitors from the search engines. However, as long as the number of visitors using those basic publishing keywords in search terms is growing, I know that the number of appropriate visitors is still growing overall. If I'd checked phrases, on the other hand, and saw that all of my traffic was coming from "know-it-all-jerk", I'd have to rethink my approach.
And if you're a publisher who said "come again" on reading this post, drop me a line and maybe I can walk you through it over the phone:-)