# Let's calculate SERP

Discussion in 'Black Hat SEO' started by merieke, Sep 2, 2011.

1. ### meriekeRegistered Member

Joined:
Jun 20, 2011
Messages:
54
21
Thought it might be a worthwhile endeavor to collect some raw data and see if we can generate a formula that loosely models the correlation between backlinks and SERP position.

For simplicity each variable will represent the average, estimated monthly value.

Y = Current SERP Position
B = % of "high" value backlinks (ie. 10% means that 10% of your backlinks are PR4+)
B2 = Average age of backlinks in months
C = Age of your domain in months
D = % of usable, quality content (ie. 75% means that 75% of your content would pass human review)

There are hundreds (potentially thousands when compounded) of variables, but let's keep this as a simple thought experiment and solve for Y using the above 5 mentioned variables. In addition you can post optional qualitative or quantitative data to help explain your data.

Y =
A =
B =
B2 =
C =
D =
Optional: Niche =
Optional: Monetization Method =
Optional: ?

My primary objective is to model a formula to avoid sandboxing. Depending on the quality of the data used and robustness of the chosen fields we can perhaps calculate other interesting things.

• Thanks x 1

Joined:
May 16, 2011
Messages:
54
13
Hey mate, great idea, unfortunately I can't help you at this stage. Goodluck with it though

3. ### meriekeRegistered Member

Joined:
Jun 20, 2011
Messages:
54
21
Thanks, any other takers?

4. ### jon_xx_xJr. VIPJr. VIP

Joined:
Nov 15, 2008
Messages:
3,240
1,498
No, it's kinda dumb, sorry.

5. ### machines-in-motionJunior Member

Joined:
Aug 11, 2010
Messages:
129
46
Occupation:
Engineer
Location:
Here and there
It's good to see someone finally applying the scientific method to Google's algorithm. I suggest adding variables for "time since last update" (which has been extremely important for my sites) and keyword competition (also extremely important).

Also, it might be a good idea to standardize how we measure backlinks. Different methods under-report by different amounts.

6. ### meriekeRegistered Member

Joined:
Jun 20, 2011
Messages:
54
21
I'm thinking of this more like the dirty scientific method. I thought it would be easiest in a forum to pull the simple stats. Getting into time of last update and keyword competition becomes much more complex to track, so even though its completely relevant it may not be practical ... at least not in terms of data gathering in a forum.

I was measuring backlinks using google webmaster tools, since I think most people who are trying to rank in google would have this installed. I'm open to suggestions though.

7. ### ingredibleJunior Member

Joined:
Sep 11, 2010
Messages:
183
25
hi, nice thing - but you forgot to measure the Onpage SEO - for this I recommend Using "Traffic Travis" to get the value of your onpage seo.

8. ### meriekeRegistered Member

Joined:
Jun 20, 2011
Messages:
54
21
Decidedly an important factor, but difficult to measure and make sure that everyone is using the same standard. We need to pick out only the easiest variables to estimate. In a more controlled environment we could use this variable along with many others.

9. ### machines-in-motionJunior Member

Joined:
Aug 11, 2010
Messages:
129
46
Occupation:
Engineer
Location:
Here and there
I'll start:
Site 1:
Y=1
B=100
B2=2
C=2
D=100
Onpage SEO should be qualitatively good because a service that was designed for onpage SEO was used to create the site.
Last updated 2 months ago (on creation)
I would rather not say niche, but there are 600,000 other results as of this moment.

10. ### englandrmBANNEDBANNED

Joined:
Nov 11, 2010
Messages:
942
431
I think you need to take into account their current rank, and all those stats of at least the top 10. You'll be gathering data from so many niches of differing competition, so they won't compare without more info.

11. ### jonnymchRegistered Member

Joined:
May 8, 2010
Messages:
50
4
If you google "seomaz factors 2011", they have done something quite similar and it is definitely worth reading.

I would be up for making an econometric-style regression but would need a heck of a lot more data that 5 people on BHW If someone is able to collect say 1000 mid-high competitive keywords and get TT data points, then we could get some interesting results.

On page SEO title, url, description, meta etc could be dummys and off page could consist of no. links at each PR and split between anchor + non-anchor.

12. ### ibmethatswhoibJr. VIPJr. VIPPremium Member

Joined:
Feb 17, 2011
Messages:
1,561
1,164
Occupation:
Staying Informed
Location:
Bay Area, Ca
ya they did a good one, just make sure you g00gle "seomoz factors 2011" cuz seomaz won't get you any results. Just for those that have never heard of seomoz.

13. ### VIC SEOElite Member

Joined:
Feb 19, 2010
Messages:
2,165
365
Gender:
Male
Occupation:
SEO Specialist
This kind of thing is what those guys at the SE developers tend to figure out. We out here cannot sit and figure out something that is kept secret from us. Do you think that if could do it then wouldn't we always be at the top no matter what update they throw at us.

14. ### CodythebestSenior Member

Joined:
May 29, 2009
Messages:
1,138
896
Location:
Fuerteventura
Are the backlinks anything to do with SERP?
I have a site that is #14 for a keyword and 84,000 backlinks.
The #1 has 3, #2 has 0, #3 has 7, #4 has 1, #5 has 0, #6 has 4 and none of the #1 to #13 has more than 10.

Joined:
Sep 17, 2008
Messages:
2,692