Market Share, Metrics, and SEO

by Michael Martinez on August 24, 2007

I like to analyze things. I suppose that’s obvious by now.

I’ve been watching search engine market share reports for years. They never seem to agree on what the search market should be. You have four major reporting services: Compete, comScore, Hitwise, and Nielsen. Although a couple of these services have begun adjusting their metrics, none of them really look at the search engine market. They limit their analysis to number of queries performed.

Over the past few months I’ve been tracking the monthly reports in tandem at Spider-Food. Each thread in the list below starts out with a recap of the four monthly reports (there may be one or two missing reports because the data is not always made public). Yesterday I added recaps to the various threads that look at the numbers in different ways. For example, I had forgotten that Google lost “market share” after the Christmas holidays.

These metrics are virtually useless for search engine optimization as they are badly contrived estimates that don’t take other important factors into consideration.

Real search market share needs to be measured in terms of search properties used (MSN.com versus Live.com), number of visitors who perform queries versus number of visitors, and number of queries that produce click-throughs.

Although I do most of my searching at Google, the chief reason I perform so many queries at Google is that I usually cannot find what I am looking for. I’ve been railing at Google over the sucky quality of their search results not because I am having trouble manipulating those search results but because they make it so damned hard to find what I am looking for.

So why don’t I use other search engines? I do, actually. But Ask only lists the best Web sites and sometimes I’m not interested in the best Web sites. Yahoo! lists non-existent URLs. I guess they suffer from index-envy or something because they just make up crap and throw it into the index. Have you ever actually CLICKED on those links that Site Explorer shows you?

Live Search is okay for some things but its index is limited and, frankly, I don’t really like the spooky interface. Google gives me a plain white page with black text and big bold titles and I can just sit there and scroll through irrelevant search results and feel like I’m accomplishing something.

Google gives me comfort search results. The other search engines just give me search results.

What I’d really like is to find what I’m looking for. I think other people are frustrated with Google like me, but they probably just continue wearing the floppy old shoes because — like me — they feel comfortable in those floppy old shoes.

A few months ago I compiled a more thorough search engine market share report for March 2007 by combining data about numbers of visitors with numbers of queries. But though I say the report was more thorough it was really an estimate. I estimated how many queries per visit and queries per visitor each search engine might have received.

Google has hire queries-per-visit and queries-per-visitor rates over Yahoo! and Microsoft, but the nearly 3-to-1 ratio does not necessarily mean Google provides a better search experience. It means the data is clouded by non-search visitor counts. Both Yahoo! and Microsoft are major content providers, so the data provided by metrics companies doesn’t really help us measure search engine markets.

Another metric I like to watch is how long people remain on the average Web page. As length of visit is now achieving greater prominence in search market metrics (and other metrics), I am sometimes amused to find that my ugly Web sites consistently outperform beautiful Web sites. Maybe that’s because people spend more time reading Web content than looking at beautiful page layouts.

If the average Web page view lasts around 45 seconds to 1 minute, what does it say about SEO Theory that the average visit lasts around 8 minutes and that the average visitor looks at between 1.6 and 2 pages per visit? That I write some long-winded copy is obvious, but the implicit value the metrics reveal is that copy is more interesting than graphics.

People may insist that your site performs better with less copy, more gimmicks, but I’ve never found that to be the case. Some people spend up to 20 or 30 minutes on SEO Theory — new visitors especially tend to devote 15 to 30 minutes of their time on the Web site. Maybe they take phone calls or get up for a cup of coffee, but these lengthy visits are numerous and have been consistent since I started tracking metrics for the blog.

Xenite.Org also gets some marathon visits. One of the more common (and much appreciated) compliments we hear from time to time is, “You guys have SO MUCH CONTENT!” Yeah, we’ve spent years creating that content.

There is a ton of interesting research on Web metrics that you can find (of course, it takes me quite a few queries to find this stuff). For example, Yahoo! says that 1 third of all queries are looking for opinion. Not in my case, but sometimes I do look for people’s opinions. I look at service and product reviews like many other consumers do.

So what does that research tell you about ecommerce sites? How can you optimize an ecommerce site for 1/3 of all queries on Yahoo!?

Search engines don’t always show you where to find lots of link building tips and techniques. You might want to know about Technique Tip: The 3-second Test, Technique Tip: The Link Anchor Test, and Technique Tip: Who Is Depending On Links?.

Maybe not. Maybe you just want to read blog posts that show you links to five great blog posts on other blogs. Maybe you’re satisified with the search results that search engines give you. I’m not.

And that is really what drives search engine optimization. If you want good on-page optimization tips and techniques you may need to dig a little for them. I mean, how many times can you go down the list of blog posts that tell you to write good title tags? I may just write a bad title tag for this post, just to make it hard for people to find it.

We optimize our sites so that our content is found. We optimize our sites because we’re not happy with what we find in search engines. Part of that dissatisfaction stems from our desire for visibility, but in many cases we would not be moved to create content if it already existed. And the thing is, it may indeed already exist. We just cannot find it.

Analyzing the information that people make available on the Web is about the only means we have of seeing what the competition is up to. Analyzing backlinks on Yahoo! doesn’t tell you whether I get traffic from Microsoft’s network. Looking at the search referral data I share every month at least gives you an indication of where I get my search traffic. Analyzing PageRank tutorials won’t change your search engine positioning, but finding out how other people boil down PageRank to just a few bullet points helps you see where it really helps and where it doesn’t (keep in mind that PageRank’s general usefulness has only recently become significant thanks to Google’s favoritism).

And speaking about Microsoft, what would you learn if — in analyzing your own server logs — you found that only 10% of your pages were getting referrals from Microsoft? If your site is well-indexed in Google and Yahoo!, if it seems well-indexed in Ask and Microsoft, why should so few of your pages earn referrals from Microsoft?

How many SEOs look at their referral data in that way? How often do you look at your content’s relationship with a specific search engine (other than Google) and ask, “Can I improve this?”

If only 10% of your pages get traffic from Microsoft, what would your traffic be like if you could get referrals for all of your pages from Microsoft?

If only 30% of your pages get traffic from Yahoo!, what would your traffic be like if you could get referrals for all of your pages from Yahoo!?

Do your referral statistics compare well to other people’s referral statistics? That would be a great SEO social medai site and a tool well worth building. Right now, there are absolutely no decent SEO tools anywhere on the Web. None. I’ve looked at them. They tell you nothing useful.

But maybe I’m so disappointed in today’s SEO tools only because I can’t find any decent ones when I search for them.

{ 0 comments… add one now }