Forum Moderators: open
Any other tips that a CMS could utilise to serve them up ? or cloaking tips for this content? risks of dupe content at all?
Thanks.
Musing off the top of my head on the other aspects, I'm rather skeptical of the approach in general. I think I could put it this way in brief: demographic targetting is what they used to do back in the 1900s before data capture was sufficient to do individual targetting. Obviously, I'm exaggerating for effect, but think about how the Amazon shopping experience has changed over the years. They do not show me pages/products that other men liked, they show me products purchased by that other people who have previously bought the same things that I have. That trumps demographics any day.
It's not clear to me whether you're selling these leisure products, or just discussing/reviewing them, so the Amazon model may only take you so far, but Amazon does make a lot of assumptions based on which pages I view as well.
So here's a bit on one way to approach it along the lines that you're thinking, but with a twist, and a suggestion for what I think would be nicer if you could pull it off (i.e. if you can gather the data). Again, all just theoretical musing, but at least something to get the discussion rolling.
In terms of the technical aspects of managing the different articles, the first question that comes to mind is how you plan to differentiate men from women, cookies or no. Are you just wanting to assume that someone who starts on a "female" page will henceforth always want female pages? Maybe I track stereotypically with females in some cases (I'm shopping for cut flowers), but not in others (I'm not shopping for tampons).
And even if you pull it off successfully, won't you effectively be cutting your content in half in the sense that though you have 1,000 pages, there are only 500 for men, so once you decide you have a man on the line, 500 pages of the site become offline for him? Also, don't you risk offending people because you serve them the "wrong" pages?
You could modify this somewhat and maybe be a bit more effective (though I still don't think this is a great approach). Don't do any cloaking and don't create any content that is the same article "pinkified" for women (and thus really just dupe content, whether Google sees it that way or not, and probably insulting to a fair number of women). Rather, take a page from usability testing and user interface design and come up with "roles", hypothetical users who have a pretty complete backstory. As you go through the writing and design process, you'll get to know them better.
"Joan is a 57-year-old divorcee with a little extra cash now that her daughter has finished college and is out on her own. She's enjoying being able to indulge herself a bit, but after years of living close to the bone, she likes to mull over purchases for a long time before she hits the BUY button...."
Come up with as many of these personalities as you need. Now start writing content for each personality. What would Joan want to know about this product? What would stop Joan from purchasing?
Then on the CMS end, you could tag pages as "joan" pages and use some sort of "related pages" function to find similar "joan" pages. In just drupal, for example, you have the related links module [drupal.org], the faceted search [drupal.org], Content Recommendation Engine [drupal.org] and many more [drupal.org]. I presume that you could find similar support either natively or with plugins for many CMS.
At the same time, you don't exclude them from any part of the site, you don't cloak, and you shouldn't end up with anything truly duplicate. You're just guiding them.
Having said all that, it seems to me that if there's any lesson from the so-called web 2.0 revolution, it's that the user knows best, that pull is more powerful than push. So I think you could still start with the "user roles" strategy for building out your content, but then let the user decide which pages are really, truly related, much like Amazon does. AFAIK, Amazon does not know my gender, but they do know my interests and *that* is way more powerful than knowing my demographics. Furthermore, they know the interests of a huge number of people, many of who looked at the exact page that I did, and then bought something else. So....
- give people incentive to join and track what each user has looked at and ultimately bought. People who buy X also buy Y. And then show those links.
- track user paths and see which pages have high correlations for readership and suggest that people who like X also like Y.
Those are more complicated, of course, because they require lots more data and I don't know a CMS that will give you those features by default, but it seems to make a lot more sense than writing a male page and a female page.
