Digging in to Michigan Bulb

It would be much more effective if the site could break them out. “Your recently viewed” items should have its own unique presentation, as it's one of the most beneficial things you can add to a site. It takes away a lot of the pressure users put on text search when they're trying to find something they just looked at.

There are a lot of other things that Michigan Bulb could do to improve the site, but the biggest impact would most likely come from fixing the shopping cart. Take the billing information page, for example. At first glance, it looks completely harmless. Contact info on the left (usually better if it's on the right, but at least it's there), nice vertical questionnaire going down the middle with only one item per line — great presentation.

But then you look at the bottom and it gets all confusing on you — you have to figure out where the order is shipped, which is okay, but then you have to register, too — passwords, hints, oh my! To top it off, in the right-hand column there's a thing for returning customers and “have a catalog” customers. It's all so overwhelming.

Plus, there's no temperature bar, so you don't really have an idea of how much left you have to do. Not to mention the fact that there are links in the upper left-hand side to get you to become an affiliate? Perhaps Michigan Bulb should try to get the order first.

	BRIAN R. BROWN

Michigan Bulb faces many challenges typical of online retailers when it comes to search engine optimization. Some should be quite simple and painless to resolve; others will require a bit more effort.

I was excited to see that an attempt at canonicalizing the domain had been made. Canonicalization refers to the concept of “one true source,” which in this case simply means when I entered http://www.michiganbulb.com into my browser, I found myself automatically redirected to http://michiganbulb.com — well, sort of. Actually, I found myself at http://michiganbulb.com/Default.asp?.

Couple things wrong here. First and most obvious, the redirection for the domain is to a specific page, “Default.asp?” rather than the root domain “michiganbulb.com.” Either the www or non-www version is perfectly fine, just as long as everything redirects to one or the other.

The second issue can be seen by checking the server HTTP response header. Entering www.michiganbulb.com using the HTTP Viewer at Rex Swain (http://www.rexswain.com/httpview.html) shows us that a temporary (302) redirect was used, rather than a permanent (301) redirect.

Temporary redirects don't pass PageRank and link popularity. Consolidating PageRank makes that page stronger, which may lead to deeper crawling by spiders.

When Michigan Bulb goes in and changes that redirect to a 301, the company should also make sure that the destination is changed to http://michiganbulb.com/ without the “Default.asp?” at the end.

Since spiders see these as two different versions, without canonicalizing, they could potentially see twice as many pages on the site based on www and non-www versions. And no, this doesn't mean that the site has increased the chances to rank or its importance in the eyes of the search engines. Instead, efforts are divided due to duplicate content.

Digging down into the site, past the category pages down to the product pages, I found that the redirects are not in place, or at least not working at that level. This should be a relatively easy fix.

Using the site:michiganbulb.com advanced query in the three leading search engines, we see that Google is reporting 3,460 pages indexed, Yahoo reports 9,893 pages, and MSN reports 4,810 pages. Which one is correct? Actually, all of them in one way or another — and all are having some issues.

MSN might actually be doing a better job indexing the various paginated category pages. Almost too good, though, as there seem to be URLs for pages that no longer exist and a number of different variations of URLs to the same pages, perpetuating even more duplicate content.

Google is hit and miss on the paginated pages, which isn't surprising as 20 pages with nearly the same URL (http://michiganbulb.com/category.asp?start=190&c=2&Perennials=, http://michiganbulb.com/category.asp?start=210&c=2, etc.) and the same title tag and meta description might lead anyone to think they are duplications.

Yahoo isn't without issue either as there are several session ID pages in its index, which might explain why Yahoo is reporting far more than the others. Checking indexation on just this domain (non-www version) in Yahoo's Site Explorer reveals the page count drops down to 7,256. Canonicalization might eliminate 2,000 duplicate pages right there.

Michigan Bulb needs to clean up the URLs and remove as much duplication as possible. There are a variety of URL formats, some appending product titles or keywords to the URL, pagination, etc. While injecting a few relevant keywords into URLs would be a good thing, maybe not in this case. Michigan Bulb has long, cryptic URLs to begin with and appending keywords to them just makes them even longer. In fact, Google has indexed the non-keyword URL in many instances (e.g., indexed: http://michiganbulb.com/product.asp_Q_pn_E_15073, but not this one http://michiganbulb.com/product.asp_Q_pn_E_15073_A_Lilac+Hedge_E_).

Of course, I'd much rather see a URL like this: michiganbulb.com/hedges/lilac-hedge. Not only is that tasty for spiders, it looks more appealing to humans as well.

Duplicate content is often created due to URL issues like these, bloating indexation and decreasing crawling efficiency of the spiders — they may crawl the same “content” on several different URLs. In the end, these pages just cannibalize one another and dilute PageRank.

Some duplication is also created on the category pages from the filtering feature, which allows products to be displayed with or without pictures or descriptions. While this might be nice for human users, it would be better to limit the spiders to just the page with the images and the descriptions by using rel=nofollow on the other filtering links, robots meta tags on those filtered pages, or a robots.txt file.

This will again improve crawl equity and reduce potential duplication, while assuring the most complete pages get served up in the engines during searches.

Michigan Bulb's title tags are pretty good, focusing on the primary keywords first. The meta descriptions, though, could use some TLC. While meta descriptions won't affect search ranking, a good, call-to-action meta description can be the difference that gets one listing clicked on vs. another.

I'd also like to see heading tags used on the pages; I couldn't even find any H1 tags. Heading tags reinforce the signals to the search engines about the topic of a page. It was great that much of the navigation is text-based, rather than images, providing a stronger signal to search engines as well.

Michigan Bulb is off to a good start with a little descriptive text on the category and product pages. The “Help & Info” section has links to various topics under the Planting and Care area. The merchant should consider adding more links to these pages throughout the site.

Writing great how-to and helpful information based on what people are searching for is an easy way to create attractive link bait. These keyword-rich pages will be of interest to the target audience, and may draw more searchers through longtail searches, thereby connecting Michigan Bulb to new and returning customers.