Link Trend Analysis made simple

How do you assign value to the pages that link to your own content? If you were to create a scale ranging from 0 to 20, how would you rate your own pages, the pages where people give you links without your interaction, and the pages where you are somehow involved in creating the links?

You should devise a quality test for linking pages. In my opinion, about 70-80% of those pages should rate very poorly on your scale — probably less than 10, maybe less than 5 (but the rating scale is subjective).

If you use your own backlink source pages as a benchmark, you should be able to apply that same test to every page on your own site (if you do the first test right, you’ll have included all your own links anyway). In other words, how well do your outbound links perform for other people as opposed to other people’s links performing for you?

You can use any criteria for your test but I would strongly recommend you use no fewer than 2-3 or items. Be consistent.

This is the kind of research that takes time and link evaluation is a time-sensitive activity. You really only have a few hours in which to capture a snapshot of your natural linking profile.

Most people have very poor link evaluation criteria. Number of backlinks in Yahoo! and Google Toolbar PR value are the two worst possible tests for link value and yet they are undoubtedly the two most popular tests. Still, you could combine those two tests and apply them to all your linking sources and you’d probably see some sort of trend.

If you keep the test simple the first 1 or 2 times you perform it, you’ll start to see the value in this kind of analysis. If you then expand your criteria gradually, you’ll get a feel for which criteria work better for you.

Any criterion will do. Literally.

You could assign a value of 10 out of 20 to any link that is on a page with a white background and a value of 0 out of 20 to any page that is on a black background. It doesn’t matter how bizarre the rule seems at first as long as you stick to it.

If you have fewer than 100 backlinks you won’t see much in the way of trends. If you have more than a few hundred backlinks you’ll see all sorts of trends.

Trend analysis can extend across your link profile snapshot (it tells you something about the character of the sites linking to you), it can extend across weeks or months if you take multiple link profile snapshots (it tells you something about the true nature and quality of your links), and it can extend across substrata of your site or network. You may find your best inbound links (either from your own sites or other people’s) tend to concentrate around your best content.

Now, maybe that isn’t true. There is only one way to find out.

To start your test, sit down and force yourself to write 10 qualities about linking pages that can be measured. Don’t hold yourself to a standard of perfection. After all, you can come back and change the criteria at any time.

Just understand that if you change the criteria you cannot pool snapshots that use different criteria. You can only pool snapshots that use the same criteria. Intermingling link profile snapshots that are made with different criteria invalidates your analysis (unless you’re looking for derivative values — which in itself could be used to validate the reliability of a test).

The more consistency you find between snapshots, the more consistency you find between your rule sets, the more reliable your test results will be.

If you get inconsistent results then one or more of your tests are invalid. This is the sort of thing that takes time and practice.