SEOMatrix LLC Blog: Your Daily Ethical Search Engine Marketing Treats

I assume Scarlett Johannson has an e-mail address. I assume she also may be internet savvy. I assume she just might have a gmail account. So, I e-mailed her - scarlettJohannson@gmail.com and scarlett.Johannson@gmail.com as well as missjohannson@gmail.com (a long shot, i know). So far, nada, zilch, zippo, no response — no “hiya” no “hehe” no avatar smiley face — nothing. What’s going on:

1) Those may not be her actual e-mail addresses
2) She may not like receiving e-mails from me
3) She may have tried to e-mail me back (went right to the spam folder of course)
4) She is on location and doesn’t have WiFi or access to an internet cafe
5) Ryan Reynolds promptly deleted the correspondence

The difficulty in deciphering her lack of engagement is similar to my A/B website testing. For those marketing newbies who aren’t familiar with Robert McNamara’s system of analysis, A/B testing is breaking down a system into individual elements to test which variables are really a decisive factor within the system (remember the grade-school correlation vs. causation logical reasoning, as well). A/B testing or split testing changes one element at a time, so as to isolate single factors or variables.

I mainly use A/B testing for two purposes:

I test web site landing pages to determine which elements are critical in marshaling a website action or conversion. I also test web site pages to discover which on-site or off-site elements may be changed to affect search engine rankings. Either way, breaking the system down, element by element, is the critical theme of my message here. If multiple variables are changed at the same time, the true measure of what caused the change is difficult to discern.

Testing for website conversions is a relative no-brainer — as long as I have a sufficient amount of data or visitor actions to analyze in a somewhat controlled environment. Testing Google’s ranking system is a bit more difficult, because the formula for ranking websites changes often (Mr. Brin doesn’t like to remain status quo for too long). Anyway, A/B testing for dynamic systems (like Google’s algorithm) can give you some information; but a significant amount of guesswork still enters the arena. I can empirically change factors one at a time and record my results… but, in the back of my mind, I will still be wondering: Was it my change that resulted in the ranking change, or was it Google’s algorithm change? The more control I have over the experiment, the more reliable and predictable (and profitable) my results become.

Google may agree with my first application of A/B testing as they created their own website optimizer www.google.com/websiteoptimizer chock full of A/B testing features. They haven’t gotten around to creating a Google organic ranking optimizer just yet, only general guidelines.
In a similar vein, Scarlett hasn’t gotten around to e-mailing me back as yet; but, before I consider any other factors involved in her lack of communication, I will check my spam folder now… and will keep everyone updated.