2011 is here, and that means it’s time for our biennial search engine ranking factors survey to be renewed. This year, we’re planning something much bigger and, we hope, better. Our plan is to offer a report that provides:
- Aggregated expert opinions on the importance of factors individually and ranking influencers overall (as with 2009’s version)
- Correlation numbers for the factors (on as close to a 1:1 basis comparing data against the question asked to those surveyed)
- “Causation” numbers on a relative scale derived from our machine learning-based ranking models (with error margins)
- A representation of the relative chunks of the algorithmic pie by their contribution to the overall algorithm
This is, obviously, a huge undertaking for SEOmoz’s team, and we could use your help. First, we need your help to recruit the right experts. The form below will enable you to submit nominees:
We’ll likely take between 1-200 participants (possibly more), so please send us your best and brightest!
In addition, we’d love your help in defining the factors we’ll be measuring this year.
Rand’s Current List of 224 Potential Factors
The list below represents my first stab at creating a list of datapoints to use in our correlation and ranking model analysis. Your mission (should you choose to accept it) is to add potential factors (not listed here) that we could gather and analyze in the comments below. This means they’d need to be available on the page/domain itself or fetch-able on the web through an API or other request in a scalable fashion.
In addition to adding your own ideas in the comments, please upvote your fellow mozzers if you like the ideas they’re presenting. The comment with the most thumbs up at week’s end will earn a special gift from the mozplex and recognition in the final report.
Obviously, not all of these will be directly translated to ideas/concepts of ranking factors for the survey participants to vote on, but many can help to inform their construction and compare against as a datapoint.
Some additional notes on our plans:
- Use only Google.com US results for this version (but plan to replicate in other geos/countries)
- Retrieve geographically agnostic results using a query structure similar to this one for barber shop and this one for ice cream
- Record the presence of ads and universal results on the page, but don’t count these URLs/references in our analyses
- Segment the data in several variations (by popularity of the query according to Google’s AdWords estimates and number of words in the phrase, for example) to be provided as drill-downs from the main report
If you have other suggestions, feel free to comment below or use the form above. I’m looking forward to a remarkable step forward in the understanding of Google’s ranking system – thanks so much for your help!
p.s. A huge thanks to our many contributors from years past, many/most of whom we’ll be “nominating” for inclusion this year ourselves 🙂
p.p.s. You’ll likely notice lots of factors on my list that are obvious non-factors (meta keywords, for example). I’m including these only to help us show data on their impact (or lack thereof) which will hopefully assist those of us who need further evidence to help convince clients, managers, etc.
