Explanation of Study
This study details 24 searches performed at Google and the top 10 sites that were returned in the results. If more than one page was returned form a site, that result is excluded and the next unique site in the results is used. The searches all relate to the commercial real estate finance industry - a unique and lucrative segment of the mortgage industry. Search phrases were chosen based on the number of searches they received last month (according to Overture & WordTracker databases).
Please note that due to the constant shifting of the SERPs, the results of this survey may no longer be 100% accurate. However, the author of this survey (randfish) has closely monitored these search results for the past 9 months and the results have been surprisingly static, only shifting a few positions/sites per month. Most of the websites have held results in the top 10-20 since monitoring began in March of 2004.
Below is a key for the table columns, please note that analysis on SEO factors (links, PR, etc.) was performed on the top-level domain (TLD) rather than for the individual page.
- Pages GG
# of results returned for the search - site:url.com
- Backlinks GG
# of results returned for the search - link:www.url.com
- Backlinks Y!
# of results returned for the search - linkdomain:url.com -site:url.com
- Backlinks MSN Beta
# of results returned for the search - link:www.url.com -site:url.com
- Archive Date
Earliest Month & Year listed - automatically extracted from the Wayback Machine
PageRank in Toolbar for the top-level domain (home page) of the site
The position of the site in a search @ Google using allinanchor:keyword phrase
Personal analysis of the relevance & quality of information provided on the result page by the author (randfish). The results are on a scale of 1-10 with 10 being an exceptionally high quality result and 1 being a completely unrelated/useless page. Quality is measured by both the relevance to the search term and the quality of information provided. Therefore, a page which offers quality information but is not relevant to the search topic will receive a low rating.
Each result was fit into 1 of 5 categories:
1. (C) Commercial page - direct company seeking business
2. (L) Listing Site - offers listings/services for others; an indirect commercial site; i.e. classifieds, referrers, etc.
2. (S) Spam page - offering useless jargon/links or PPC ads without unique content/services
3. (I) Information page - providing information about the search term; i.e. articles, definitions, etc.
4. (H) Hub page - providing primarily links to resources offering information/services; i.e. directory pages, etc.
5. (O) Other page - pages which did not fit the classification system or where classification was unclear
|Most Popular TLDs|| |
Avg Pages GG
Median Pages GG
Avg BLs GG
Avg BLs Y!
Median BLs Y!
Avg BLs MSN
|dmoz.org - 4 |
bankrate.com - 6
businessfinance.com - 10
c-loans.com - 5
hardmoneyfunding.com - 4
help-finance.com - 4
kennedyfunding.com - 10
lendinguniverse.com - 4
loanuniverse.com - 4
mortgage-investments.com - 6
rebuz.com - 7
steelheadcapital.com - 4
# Cs - 84
# Is - 62
# Ls - 41
# Hs - 26
# Ss - 22
# Os - 5
Overall the quality of the results at Google for these searches was positive. There were several outstanding results for a few of the searches, and several searches for which none of the 10 results were acceptable (non conforming loans, rehab loan, commercial mortgage refinance). My grading on the quality was generally harsh - if I thought that an average user searching for this term would not find the information useful, I graded 4 or below. Slightly useful sites received a 5 or 6, while results with a quality grade 7 and above were quite positive.
For several searches the same result was listed twice (or more), in these instances, my grading may vary based on how well the result was attuned to the search. For example, kennedyfunding.com received a grade of 6 for a search for "bridge loan", but a 7 for "hard money bridge loan" as their page deals more specifically with that subject.
The bar chart above shows the correlation (or lack thereof) between PageRank & Quality for the 24 search terms measured. Sadly, PageRank's once useful measure of quality has declined rapidly as webmasters and optimizers used it as a means to manipulate the rankings. As further proof of this, the average spam page's top-level domain (TLD) had a PageRank of 5.5. Measurements of page PR were not used for this report due to the infrequency of PR updates and the relative uselessness of the data that would have resulted.
PageRank is still an interesting measure in this report, however. Of the 240 sites, 153 were unique TLDs. From these, only 14 had a PageRank of 4 or under. From this, I infer that it is necessary to have at least the importance of a PR 5 to be consistently competitive for a search phrase in the commercial real estate financing sector. Indeed, the high average PageRank of 5.75 across the survey indicates that the power of the links to a website still has a strong influence of high rankings at Google.
Each result was fit into 1 of 5 categories:
1. (C) Commercial pages - direct company seeking business
2. (L) Listing pages's - offers listings/services for others; an indirect commercial site; i.e. classifieds, referrers, etc.
2. (S) Spam pages - offering useless jargon/links or PPC ads without unique content/services
3. (I) Informational pages - providing information about the search term; i.e. articles, definitions, etc.
4. (H) Hub pages - providing primarily links to resources offering information/services; i.e. directory pages, etc.
5. (O) Other pages - pages which did not fit the classification system or where classification was unclear
One of the most important and interesting measurements for this report was the category or 'type' of result each URL fit into. Out of 24 search terms and 240 results, 84 were direct commercial websites attempting to make a 'sale' to the visitor. These sites are certainly appropriate for all of these search terms as users searching could reasonably be expected to seek companies who could fulfill their needs. The commercial results were often the most relevant and directly related results, although they often provided biased or skewed information on the subject.
At one time, allinanchor position was practically synonymous with good rankings at Google. It is clear from this study that this trend has fallen off significantly over the past 18-24 months. Many sites positioned in the top ten were nowhere in the top 100 (or even 1000) allinanchor results for the same search at Google. The use of anchor text to boost website rankings into top positions may still be somewhat effective, but it is certainly no longer the over-powering criteria that it once was.
The sponsor of this study (AvatarFinancial.com) is #1 for the allinanchor search for almost half of the search terms above, yet appears nowhere in the top 20 results for any of the searches. Looking at the allinanchor results, many are spam sites, commonly search engine results scraped pages, that are utilizing anchor text to gain a boost. Luckily, Google has identified this issue and dealt with it.
Many forum threads and SEOs have long been questioning the value of allinanchor, and it appears that, at least for this industry segment, only slight value is placed on the use of direct, corresponding anchor text. In particular, these searches did not have a single result that ranked in the top 5 for allinanchor searches: private money lender, real estate bridge loan, commercial mortgage refinance, bridge loan mortgage. Furthermore, many of the top 10 sites were absent from the allinanchor results entirely.
It's difficult to look at the age of top results without raising an eyebrow. Clearly, there is a correlation between older sites and better rankings. Many theories exist as to why this happens, but it is clear that sites launched after 2001 are having an exceptionally difficult time being listed for searches in this industry. The average age is June of 1999 - more than 5 years old. The results show a total of only 34 sites from 2002, 2003 & 2004 out of 240 total sites. The only outlier is adminlabs.com, which despite being listed in Archive.org only since March of 2004, is listed #1 for 'Commercial Mortgage Refinance' and #4 for 'Hard Money Bridge Loan'.
With age playing such a big factor, I thought it appropriate to list some of the most likely explanations for why Google is valuing age so highly.
- Age of Links
This argument supposes that Google values links that have longevity and staying power, rather than short-term links. However, the results above would indicate that the most valued links are very, very old indeed, dating back to the nascent beginnings of Google's index itself.
- Better/More Links are Inherent to Older Sites
The older a site is, the more time it has to gain more links and more valuable links from other sites. This argument, while compelling, can be easily disproven by poorly ranking sites that have far better/more valuable links than sites in the top 10. Several of the searches above, for example hardmoneyfunding.com & kennedyfunding.com, have comparatively very few links to them from external sites, yet rank exceptionally well based on other factors. Many of the competitors to these sites have 10X or more the number of links as well as a greater number of links from relevant, high ranking sites/pages.
- Age Helps Preclude Spam Sites
One of Google's biggest issues in providing quality search results is to reduce the occurrence of so-called search engine spam results in the top listings. Age is an excellent method to prevent spam results from appearing in the search results, because search spam has only become popular over the last 3-4 years. Unfortunately, this method also hurts newer, non-spam sites and dramatically reduces the growth and quality of the web.
The size ranges of sites was incredibly varied, resulting in averages and medians that were nothing alike. Overall, the average number of pages a site had in Google's index was 89,469 while the median sites had only 1560 pages. The standard deviation was almost 400,000, meaning that results were vastly skewed, especially on the top. Several sites, like dmoz.org, yahoo.com and others resulted in a practically meaningless average. We can look at the overall data and the median number, however, and determine that overall, sites with under 500-1000 pages did not fair well, with a few notable exceptions under 100 pages (kennedyfunding.com, hardmoneyfunding.com, steelheadcapital.com).
The search optimization industry often equates large sites with better rankings and this report certainly lends some additional evidence to that theory. However, overwhelming (10,000+) size does not appear necessary to rank well. Several quality, niche sites that lack both size and high numbers of backlinks continue to thrive in the top 10.
Once again, the massive difference between the average and median for the results indicates a high level of skewing on the upward side. Yahoo!'s linkdomain command proved to be the most useful and consistent measure of links The average for all 240 results was 199,693 while the median was 2055. With 120 sites having 2055 or fewer links or fewer, it's clear that Google considers numerical superiority of links to be only a small factor in ranking. As many SEOs have surmised, the quality, theme & relevance of links may be much more influential.
This report was not able to measure the quality of links to sites. However, the SOCEngine development team is working to create more advanced link analysis tools to help analyze the theming and relevance of a particular link.
Four specific pieces of data stood out from this analysis.
- For one reason or another, age of site appeared to be the most dominant factor in top ranked sites vs. non-ranked sites.
- Certain small sites with relatively few links (kennedyfunding.com, steelheadcapital.com, hardmoneyfunding.com) were incredibly successful, eclipsing the dominance of many larger, more linked to sites (reiclub.com, realestateabc.com, realestatepromo.com).
- Pages located on sites (TLDs) with PageRank of 4 or less were virtually absent from the top results.
- Hub sites & pages were the least likely of the major types to be listed in the results, despite often being of the highest quality & relevance.
Google's results suggest that factors beyond the current SEO knowledge exist that are helping to create the results for many of the search terms. A reasonable hypothesis could be made that suggests the use of manual approval of certain sites/pages to appear in the results, although this clearly does not dominate the SERPs in this segment, as evidenced by the nearly 10% of sites that were spam (22/240).
Additionally, the evidence here suggests that a careful probe of the smaller, successful sites is critical, along with a review of possible other factors that could be influencing the results. This report certainly suggests that a monitoring, scraping tool would be highly useful in order to monitor the results on a consistent basis and examine theories about ranking movement.