SEO Metrics – Search engine optimization metrics that help

Run a query for “SEO metrics” and you’re not likely to find anything useful, although you’ll find quite a bit of content that is relevant. The irony in this particular query runs deep.

Because there are no industry standards, there are no standard metrics. Most SEOs will look at Yahoo! backlinks and Google Toolbar PageRank for the mere value of seeing something. I look at Toolbar PR so little I have no idea of whether any of my sites were hit in the recent Google “here be pages we think sell links so go buy links from other pages” debacle–er, updates.

Of course, I don’t sell links but what might reduce some SEO spam around the Web would be the ability to use Webmaster Central to ask Google to lower our Toolbar PR. At least that way we wouldn’t have to filter out stupid link requests.

But I digress. I was whining about the lack of SEO standards and standard SEO metrics. Most people in the industry have numbingly adopted the mindless mantra of “PageRank is based on links, go check your backlinks at Yahoo!, but PageRank is based on the QUALITY of your links, so go check your backlinks at Yahoo!”.

I can almost hear the knights from “The Court Jester” mumbling “Yea! Verily, Yea!” in the background.

But I digress. I was whining about the lack of SEO industry standards and standard SEO metrics.

If PageRank and Yahoo! backlinks are really not relevant to measuring search engine optimization success, what can we look at? A number of people have proposed concepts like “number of pages indexed” (not bad for a start but that is so 1999), “number of pages that rank” (rank for what? I must ask), “number of pages that rank for targeted expressions versus number of targeted expressions”, “number of pages that generate search traffic” (not bad, but a bit crude), and “number of pages that have their own Google Toolbar value”.

I can almost see some use in that last suggestion, but how do you assign significance to the fact that, say, 100 pages out of 2200 have their own Google Toolbar valuations?

If you want to measure SEO success you have to measure visibility (you got the pages into the SERPs), referrals (traffic), and conversions. But then, you might not want traffic for a specific page. You might actually want to create a page that is indexed but which sends no traffic.

Why? So you can get some link love out of it. Yes, search engines index pages that don’t actually appear for any queries. How do I know this? I know it because Google tells me about pages that get queries — oh, wait. The Webmaster Central data looks nice but it isn’t very informative.

Have you ever noticed that Google’s data is about as complete and useful as a 2-year-old lump of clay in a Kindergarten class? I give the folks at the Googleplex credit for making a worthy effort, but their manic-depressive attitude toward sharing data with Webmasters (that is, SEOs) really stunts the viability of the information they provide.

I’ve been reading server logs and other Web statistics tools since before anyone ever heard of Google. I’ve gone from lusting after hits to capturing page views to looking at visitor counts, hoping for a higher number of unique visitors, and wondering just what the heck constitutes a page view.

Does it count if someone downloads an RSS feed I create? Does it count if they use an SEO tool to look at my code? Does it count if someone grabs my images off my server for a forum post in a private forum?

I don’t look to Google for the answers to these questions. Nor will I look to Hitwise, Compete, Alexa, Quantcast, comScore, Hitbox, or any other self-appointed metrics guru. Sorry, folks, but your numbers don’t impress me.

The search optimizer’s goal is to build search referral traffic that responds to calls to action, even if those calls to action are only to turn the page, click on a link, or otherwise do nothing to participate in an exchange of goods or services for money. Searchers who surf away don’t count. Searchers who do what you ask of them do count.

Surfers who keep coming back to your site don’t count.

Surfers who find your site through social media don’t count (but they do count for social media optimization).

Search engine optimization is all about bringing in new people through the search channel. Except a lot of us utilize site search functions powered by major search engines. Is that really optimization if we use a search engine to help people navigate around our site? I would consider it to be SEO because we’re achieving optimal performance through search — but some people would argue that SEO is really focused on bringing in new traffic.

I don’t think new traffic or unique visitors is really the goal of search engine optimization. I like to use search engines augment my memory (although it’s not always a good idea). I’ve often remembered just enough about a Web site to be able to find it again through a query on a search engine. Sometimes that idea even works.

There are people who find the SEO Theory blog by searching for “seo theory”, “michael martinez seo”, and “seo theory blog”. I like to think I do a pretty good job of ranking for those and hundreds of other expressions but how many of these people do I really want to bring here? I mean, what are they looking for?

A fair number of other entities in this industry are starting to experiment with optimizing for “seo theory”. I’m not sure there is any money in the expression — perhaps one day there will be some money — but I see more and more people creating “seo theory” content, and I don’t mean the bloggers who fell for the link building scam.

The number of competitors you find for a query is an interesting metric. I tell my team to look at how competitive a query is to get an idea of whether we’ll be relying mostly on content or figuring out ways to build useful, value-passing links.

The number of related queries you find for a query is another interesting metric. In my opinion, a query is related to another query if their shared words outnumber the unique words in either query. Hence, “Britney Spears” is related to “Britney Spears news” but “how about those britney spears dance lessons” is not related to “another great britney spears news article”.

Some search engines will guess at whether two sites are related. Everyone is pretty sure the relationships are defined by links but we spend a lot of time scratching our heads. In my opinion, a site may be related to mine if one or both of us links to the other more than once and if we share one or more sources for inbound links. So why does Google think a form on Lycos’ Help Desk is related to one of my domains?

How is MartinezPhoto related to SEO Theory? I never even met the guy (and he probably hates me for dominating our shared name space).

Maybe Google’s “related site” function takes on-page content into consideration as well as those supposed linking relationships. That would almost make their related: query operator a semantic tool.

One metric I have not seen anyone measure is how many queries a page ranks for. Oh, I know, we could spend all week analyzing one hour’s worth of data from a busy server log. But it’s the thought that counts. If one page gets traffic from 20 queries and another page gets traffic from 2 queries, which is the more optimized page?

What if the page with only 2 traffic-earning queries gets more visitors, more conversions, or somehow produces more revenue than the first page? Which page is better optimized: the page with the most visibility or the page with the most converting search traffic?

What if the page with the most converting search traffic does worse with non-search traffic? You should not assume there are just search engines out there sending you traffic.

SEO metrics are no laughing matter. They are the lifeblood of this industry. Without metrics we have no idea of whether we have achieved any sort of success. Now, one reason for the lack of metric standards is that people set out to achieve different goals. Different goals require different metrics, and that means it takes longer for standardization to set in.

Without standardization we have no way to know whose metrics are best. All we can assume is that everyone who thinks they are accomplishing something with SEO probably has something that passes for a metric.

When taking on a new client it’s a good idea to establish early on which metrics you’ll be covering. Clients can have their own funny ideas about what constitutes success and they don’t always say your way is better.

I like to know how many search engines I get traffic from. I like to know how many search engines I can get traffic from. I like to know which search engines are used by the audiences I want to reach. Has anyone ever learned these things by looking at Toolbar PR or Yahoo! backlink reports?

Those are just some of my metrics, and not necessarily my most important ones but they do look a little beyond who is linking to whom and how and why. They look at the Web as a community, as a market, as a segmented encapsulated metaphor for human society.

My metrics seek to measure growth, change, and trends. I’m not much interested in SEO metrics because no one really has a good SEO metric that I can hang on to. Knowing that I did better today than yesterday is a better measure of my success than seeing my page at the top of a specific query. Knowing that I have continually improved or maintained my performance in search results is more informative than looking at backlinks and Toolbar PR.

Which is not to say I don’t look at backlinks. But I tend to look more at the breadth and scope of backlinks, their types, their placements, who is placing them, than the raw numbers reported by any particular tool.

If I assess a Web site’s competitive value, I’ll make note of the fact that it has 100,000 inbound links. But I’ve outranked such sites before with far fewer links. If I know a site has a high Toolbar PR I’ll assume that the number is derived from something but I’ve outranked pages with high Toolbar PR before.

The bottom line is that I do use some metrics but I don’t really share them on a consistent basis. For one thing, the metrics I use this year are very different from the metrics I used in 2006, which were very different from the metrics I used in 2005. My metrics change on the basis of where I find value. I’ve written about metrics on SEO Theory before and I’ll write about them again, but everything I suggest as a possible metric has limited value.

The value we attach to SEO metrics stems from what we hope to achieve with our optimization. I sometimes create sites just to see what search engines will do with them. I sometimes create sites just because I love an idea. You need two different sets of metrics to measure the succes of those two very different types of sites.

I think your best metric will be the one that tells you how well all your other metrics help your search engine optimization. If you don’t have such a metric, you need one. Figure out what the rules of success are for your needs and then start mapping your data accordingly.

That’s the only standard you’ll ever need in this business.