Home of the Spider

Auto-stemming #NAME?

This is really a confluence of annoyances. For a variety of reasons, some of which make no sense, we need to have formatted descriptions for our products in the csv files imported into our new inventory management system. This system will only allow formatted descriptions of up to 35 characters per line. The system wraps unformatted descriptions automatically, but formatted descriptions would get truncated if you give it more than 35 characters.

Manually creating multiple cells of less than 35 characters each would take forever, so my first foray into VBA was automating this process. I'm not that familiar with Excel or VBA, so when I encountered my first error cell, I had no idea what it was. I was getting a type-mismatch error, and when I looked at the data in the cell, I saw this "#NAME?". I didn't even know that this indicated an error!

So I Google "#NAME?" and I get 2 billion results. Now, Excel is popular, but that couldn't be right. And I couldn't convince Google to search for the exact string I entered. They insisted on searching for what they, in their infinite genius, decided I wanted to search for.

Google, here's a free hint. If I typed in "#NAME?", I want to find web-pages with the exact term "#NAME?". Sounds kinda obvious, I know, but there it is. Had I wanted to search for "NAME", I would not have typed the "#" and the "?"!

When I finally realize that "#NAME?" indicates an error in the cell, I search for "excel error cell". The first result gives me exactly what I need, a table of error messages and what the error is. Included in this table is my "#NAME?" error, so why can't I find that directly with Google?

See Error Messages in Excel

---

Item of the Day: Jacuzzi Morphosis BR70 - Jacuzzi Air Tubs

GoogleBot & Javascript

Does GoogleBot read and attempt to understand links in Javascript? Yes.

Google's Webmaster Tools, which I love (and obsess over,) provides a table with all crawl errors GoogleBot encountered on your site. When viewing the Not Found (404) errors, I noticed some interesting URLs listed. GoogleBot reported a missing page with the URL http://www.homeandstone.com/brand:wet. The only place I used that type of construct in my URLs (/brand:wet) is in the Bathroom Vanity Finder, where I'll use a this URL - http://www.homeandstone.com/shop/bathroom-vanities.cfm/all:vanities/brand:wet - to show only vanities by the brand Wet Style.

I combed through every link in the website, and there was no link to just /brand:wet. So where did GoogleBot get the link from?

When we built the vanity finder, our goal was to give our customers the ability to shop for vanities any way they wanted, rather than forcing them to use our category hierarchy. To do this, we used Ajax technology (built by me!) to update the product listing based on the current selections in the finder. So as you browse, you are always seeing a list that is in sync with your selections, rather than having to submit a form to update the product listing.

The only exception is the vanities homepage, which has a graphic along with a few featured vanities, and a links to shop by brand. When you click on a link on this page, you are taken to a products listing page. Same for the finder on the homepage, you do have to submit the form to see the results. On subsequent pages, it will update as you select.

Say you choose a brand. The product listing page will show all vanities from that brand. As you make selections in the finder, these selections apply on top of the brand, so for example, if you choose the brand Wet, and then select '31" - 36"' for Size, the product listing updates with all matching vanities only from Wet. There is a drop-down of brands in the finder, so you can update the brand independently of other selections.

In order to keep that brand setting, there is a variable "brandURL" that saves the current brand selected. When you click on a link that includes a brand selection (the "/brand:" part,) the server sets the brandURL variable to... thats right, "/brand:wet"!

Being that this is the only place such a link appears, I believe that the GoogleBot reads the Javascript, sees the variable with "URL" as part of its name, and assumes (wrongly,) that its a URL used somewhere.

---

Item of the Day: Suneli Taranto 8033-CH - Complete Bathroom Vanity Set

Webmaster Tools

I've been using Google's Webmaster Tools for a few months now, and overall it's a great product. Particularly, it gives an insight into how Google sees your site, something that isn't available from other search engines.

But there are three things they could improve to make it really spectacular. They are little things, but I find them awfully frustrating.

1. Ranking (Stats > Crawl Stats)
So the official stand is that Google is telling you more about your sites ranking than what the toolbar gives you. I don't buy it. The toolbar tells you the ranking for the page you are looking at, and it gives you concrete number - something I can use to compare to other pages and sites. Webmaster Tools gives you a graph that gives you no information about individual pages, just a general site overview. How is that helpful? Also, say you improve your rank from a 4 to a 5 or 6. No doubt, I'd consider that an accomplishment. Well, you would never know from the graph that your rank has improved!

2. Search Queries (Stats > Query Stats)
This is just useless. I don't care an ounce for how Google's datacenters are set up. But when query stats tells me that I show up number 8 for search term, I should not show up number 58! Thats just foolishness. It's a wonderful idea terribly implemented.

3. Links (Stats > Index Stats)
Again, excellent idea, terrible implementation. It just doesn't give you all your links! Google's answer is they are considering adding that ability to webmaster tools, as the standard "link" keyword doesn't show all the results. Whatever. There is no reason that it isn't shown to me for my site. As it is, they are the only search engine that hides the link data, and the stated reason for this limitation does not apply to Webmaster Tools. This limitation means that I am forced to use Yahoo's and MSN's "linkdomain" tool, which is more powerful, and actually works. Good move, Google.

---

Item of the Day: Bathroom Vanity - Prism III 165-36 by Fairmont Designs

Google's Performance Monitor, Part II

Well, I don't think its related to server performance at all. I'm talking about the crawl rate graph I see in Google Webmaster Tools. The last time I posted about this, I assumed that the jump in the number of pages crawled per day was related to the time spent downloading a page.

Now, a month later, I see that jump in pages crawled, and I can assure you that my server was not performing better. In fact, my server has been crashing repeatedly and when up, was very sluggish.

So how to explain this?

Has Google recently changed the way they crawl sites? Are others seeing this type of jump?

It does appear that for about a week at the start of October and November, Google does a more thorough re-index of the site, while continuing to update a little at a time on a continuous basis.

I'm still very intrigued by the Time Spent Downloading a Page graph. See how it almost perfectly mirrors the Number of Pages Crawled per Day graph?

---

Item of the Day: Bathroom Vanity: Ibiza I36 by Empire Industries

Rights & Responsibilities

As a webmaster, I have a strong interest in making sure my pages are listed in the major search engines. Of course, there are many other webmasters, all attempting to ensure the same. This isn't a bad thing, it's the purpose of a search engine - to allow users to find new resources on the web.

To do this, a search engine must decide which search terms "keywords" apply to your site, and how relevant it is - the "rank". Here is where things start to get fun. Say you search for Bathroom Vanities [Google search], and a site doesn't come up. Does the owner of the site have a right to demand to be included, or does the Search Engine have the right to independently determine who is relevant and who isn't. And if the SE has that right, do they have the responsibility to inform the webmaster why his site is not relevant?

Google gives you no information about how and why a site is ranked or not. In fact, they won't even give the full results for a "link" search! Its a little like security by obscurity, which doesn't work. If you ask Google why you aren't included, all they'll tell you is that if you aren't banned (something they also don't notify you about,) just continue to build quality content, and you'll get ranked.

This all sounds good if the SE is a trusted third party, with no bias. But Google has a very large bias, specifically its $6.1 billion in revenue for 2005. This is mostly from search advertising. But if I, as an advertiser, show up high for my target keywords, then I don't spend as much on advertising for that keyword. If I drop in the rankings, I spend more. So to increase revenue, all Google has to do is make sure that sites that would be willing to advertise don't do well in the search results. That doesn't sound unbiased or trusted to me.

As well as the fact that Google made almost none of that money from content it created. It's all made off of content created by others. So the argument (made frequently) that Google doesn't owe a website anything is wrong. They owe everything to the owners of the sites listed in the index. After all, if all websites selling bathroom vanities removed themselves from the Google index, there would be no revenue from that sector at all.

Therefore, I believe that search engines have the responsibility to be more responsive to the people paying its bills. They should be required to explain a site's ranking, at least to the owner of the site, and be held accountable for the way the rank websites.

---

Item of the Day: Kitchen Faucet Pot Filler: Franke PF 2000

Supplemental Results

Recently, I noticed that a large percentage of the product pages on my site became supplemental results, at least according to Google. I take issue with that classification, but thats beside the point. I wondered what caused that. I did some research (using Google, of course), but didn't find a convincing answer. So I figured I'd add my theory to the fun.

I think its a combination of two things.

1: most of the pages that are "supplements" haven't been indexed recently. Why that is is anyone's guess. I think that the age of the page in combination with -

2: an overhauled design. All the category pages that link to products were completely updated, including design and header tags, though the URL's didn't change.

What I'm guessing happens is Google sees that all the pages on the site have changed, so it assumes that the old pages are no longer valid until it re-indexed them. As the new versions of the pages show up in Google's index, most seem to lose their status as "supplemental".

---

Item of the Day: Jacuzzi Hot Tub: Allusion Pure Air 7236

Google's Performance Monitor

Well, not exactly, at least I don't think they meant it to be a performance monitor. Matt Cutts has a nice post on Google's new tools for a webmaster.

But I noticed something else. My chart for "number of pages crawled per day" showed something interesting, with a big bump at the end of September, beginning of October:

Then I noticed this, in the "time spent downloading a page" chart:

Looks almost the same, just inverted. Seems that the number of pages Google indexes per day is an indicator of how well my site is performing.

---

Item of the day: Le Bijou KS-111 Colossal, Apron Front Kitchen Sink

Google vs. Yahoo

Recently, Google doubled its profits, while profits at Yahoo dropped 37%. Google also continued its dominance in the search market itself, with 50% of all searches. So how does Google maintain its momentum?

My personal experience has been that Google's results are no better or worse than Yahoo, and a clean interface will only take you so far. And what is PageRank, anyway?

Thats the point. PageRank is an illusion created by marketing. Its all about the buzz, getting SEO and SEM professionals talking about their PageRank. What does Yahoo offer to compete with that? Nothing. Thats not to say that PageRank doesn't exist, it does. But Yahoo also has a ranking algorithm, it just doesn't include a "tm" after it.

Its the ultimate version of organic marketing, taken to a $143 billion dollar market value.

Google has not shown the same success in other area's as it has in search, because in other areas it has no buzz factor, its just another player.

---

Item of the Day: Bath Air Tub: Serenity II by Aquatic