Every time I talk to somebody about running a publishing business that gets the majority of its customers from this website, the discussion invariably comes around to SEO - Search Engine Optimization. I don't bring it there myself because I don't believe in SEO in the traditional sense. I do believe in making it as easy as possible for a search engine to index and "understand" my web pages, which is why I began implementing the new canonical link element today. Unfortunately, my website includes a mix of link types in the navigation, both absolute (http://www.fonerbooks.com/cornered.htm) and relative (cornered.htm) due to my stubborn refusal to give up my HTML editor from 1995 which won't let me navigate a local version of the site on my own hard drive if I use absolute links. Hopefully, the new canonical link element will make sure the search engines fully credit all of the links to a page in their ranking algorithms.
But most of the money and effort that publishing companies spend on SEO is wasted for one simple reason - they don't publish any content that can benefit from best practices. At best, SEO can make it clear to the search engines what you or your SEO company wants the search engine to think the page is about. Of course, a lot of people with websites try lying to the search engines about the true content of their pages because they don't think they would get enough visitors otherwise. But search engine engineers are pretty smart, so trying to tell them lies isn't very useful, except for out on the long tail where there aren't any competing pages for a specific phrase. For example, if you bought the domain name whyismorrisrosenthalsuchajerk.com and posted an essay titled "Why is Morris Rosenthal such a jerk?", no doubt you could get the #1 placement in the search engines for the phrase. But good luck earning a living with a site like that, or with a million more like it.
The reason most publishing companies don't publish any meaningful content is that they still don't understand the web. They basically see it as an extension of glossy magazine ads or as a replacement for the NYT Sunday Book Review, and they vacillate between hoping that the Google Books program will either save their bacon or just go away. The one thing they absolutely won't do is adopt any of their premium content to the web, arranging it in an logical architecture that puts the apples with the apples and the fruitcakes with the nuts, and treat it as a permanent part of their web presence. Without doing that, all the intelligent and even well meaning SEO in the world is wasted, because nobody searches for the CV of the company and nobody will link to the corporate responsibility statement.
In the world of search engines, the #1, #2 and #3 things a website needs to draw visitors are content, content and content. Visitors who arrive at a web page without any meaningful content will never link to that page. It doesn't matter if you buy their visit through advertising, steal their visit with dicey SEO tricks, or even if they stumble on it through your navigation after arriving at the front page your famous publishing company based on your brand. Without those organic links, freely given by individuals through discussion groups, blogs, and their own web pages, your web pages will never build search engine traffic beyond the long tail SEO phrase they were engineered to draw, which won't amount to a pair of beans.
There's something to be said for sticking with what you know, but publishers who aren't willing to learn a little about how search works on the web shouldn't even bother with a company website - it will just prove a distraction and a resource drain. For authors, you could do worse than reading the draft of my abandoned guide to building a platform for marketing books and researching the commercial possibilities of your work. If you want people to find your website, you have to provide content that they want to read. And if you're writing the kind of stuff that nobody wants to read, don't be surprised if nobody does.