Long time, no post:)

It certainly shows it's summer, and to tell you the truth, I've just been too busy with the Enlightened Salesperson project and my work at the business daily.

As you probably know from other posts about the Enlightened Salesperson project development, we're basing our marketing analysis foremost on split-testing.

While split-testing does tell you what approach or specific element work best, it really gets interesting when you combine it with log file analysis.

There's no way of knowing for certain whether your findings are 100% accurate, but it does provide you with a nice foundation for further split-testing.

Here's what we discovered:

1. Test sites that had less "Table of Contents" (the page with the Table of Contents from the e-book) views and more "About the Authors" views consistently achieved greater conversion rates.

This may be just pure coincidence, but it also might be relevant, so we'll be split-testing it in the near future, because the difference in conversion rates was even more than 3%.

It's also interesting that the "Table of Contents" page has the highest Exit ration among relevant pages.

2. Two test sites with a different design and different web site copy length (weren't part of the same split-test) achieved an exact same conversion rate. This is even more interesting because test site A got 40% more visitors than test site B, but the conversion rates were still nearly the same (16,33% and 16,35%).

It's quite clear than their effectiveness seams the same. But how about if we combine them together?

For instance, we know that the shorter letter pulled almost the same as the longer letter on this test. Test site A used the shorter letter and one variation of the web site design (during another test we proved that this variation works better than the other one).

Test site B used the longer letter and the second design variation. While it seems that both the letters are performing almost equally well (with inadequate statistical difference), the second design variation performed not as well, but was tested only with the longer letter.

What if we can achieve a greater conversion by using the shorter letter from test site A and design variation from test site B?

3. During our second split-test the test site with the highest conversion achieved the least content views. Unfortunately the population sample still isn't large enough to enable us to make any conclusions. But it is possible that a single-page web site might work better than a multi-page web site. Something to test ...

4. We discovered that we actually have customers from 35 different countries. The next step is determining which countries have the highest potential and specifically focus our promotional activities there as well.

5. Page views and visit lengths are quite low. Although we can't say whether more page views per visit mean more sales just yet, we do know that greater visit lengths do.

Right now we're assuming that the opening paragraphs and content "above the fold" do not achieve enough interest in visitors for them to read on.

We'll be able to confirm this after we analyze the current-running split-test campaign when it's finished.

6. By taking a look at our log files we were able to discover exactly how many visitors individual marketing partners are sending our way, which becomes especially useful when combined with "sales per marketing partner" data.

The next step is discovering why certain partners achieve much better conversion rates and then use that knowledge to help others achieve better results.