UK SEO Consultant | Just Search's Neil Walker

Well my apologies it’s been a while sine I put a post out, but don’t worry this is a good one, i’m going to talk about SEO forecasting, how to potentially predict what return you can expect when you undertake an SEO campaign. This is following on from the SASCON Manchester Mini SEO Conference held yesterday at “The Hive” with Speakers @kelvinnewman @NicholaStott @badams hosted by @peteyoung.

It was a good mini conference but now on to the post about what I spoke about.

Background:

A couple of my last posts:

1. Organic Click through rate
2. Using Analytics to prove the long tail theory
3. Predicting time scales in SEO

These are articles were simply part of my process to engage looking at how to predict a return on investment from an SEO perspective, it’s a long article so grab a coffee and start reading.

Assumptons:

I’ll make it clear at the start of this article that this is educated guess work and that it is not a definitive set of figures, however it is a useful guide when you see the results.

Click Through Rates

A recap of my earlier article I took that data from an eye tracking study in 2004, AOL search records in 2006 and then GWT data in 2010  I did this to get an average Click through rate (CTR) if you ranked on the first page of Google. I elimated no.1 position as this would create a higher value becuase of the sheer volume of clicks going to position 1.

The average CTR for position 2-10 is approx. 4.85%

I also needed to consider what percentage of traffic clicked on Organic listings as oppose to Paid listings the research showed approximately 35% clicked on paid and 65% clicked on natural listings.

Website Conversion Rates

In terms of looking at website conversion rates I carried out research from various articles saying how conversion rate can be anything from 1% to 10%, this was to inaccurate for me so in my article about the long tail theory I took the data of 166,000+ keywords set up an custom report in analytic’s and then looked at the conversion rate by keyword length.

From all the data I collated I have advised that an average conversion rate is between 2-4%, for the purpose of this article the figure for website conversion rate is 2.9%

(By conversion I’m talking about a sale, a phone call or an enquiry.)

Keyword Analysis

We now need to get some keywords, this is an investigative exercise, think about your industry, what would you type in to find your site, use tools such as:

1. https://adwords.google.com/select/KeywordToolExternal
2. Microsoft-advertising-intelligence
3. https://adwords.google.com/o/Targeting/Explorer

This tell you how much money your making

Well I’ve been invited onto the SEO Panel at T.R.A.F.F.I.C in Dublin next week which is a conference all about domains. With this in mind I thought I would dedicate this week to some domain related blog posts, this is the first post on “Do different TLD’s affect SEO?”

The Prep
This came to mind because recently I bought a number of domains two of which I bought were a .org and a .co.uk. Now I only bough the .org because I didn’t want anyone to purchase it instead of me but my focus was on the .co.uk so I redirected the .org through the hosting control panel as a 301 (but they lied it was a 302!) to the .co.uk.

All I did was put a holding page on the .co.uk I didn’t submit any domain to the search engines, didn’t do any link building, ping any articles out but I noticed the .org appeared in Google SERP’s. A few weeks later the .co.uk appeared but ranked far lower than the .org, this is not an indpeth peice of research but does point the finger that may be google places .org TLD’d with a little more favour than a .co.uk?

Research
So I carried out some research on the subject, personally I’ve never really had a problem with domains but an instinct has always made me be aware of .tv, .biz, .info despite i have seen some ranking well, i just don’t like them! Anyway I digress but I came across an article on the Google cache about TLD’s in SEO showing bias to .org domains, this study was carried out in 2008 so that’s two years ago which in terms of Search is a “long time” however the data seemed to correlate to what i had briefly experienced.

Their Investigation

There are several ranking factors we need to control.

1. domain age (purchase new domains at same time)
2. link profile (use Google sitemaps for indexing)
3. indexing age (randomize ordering of multiple subdomains in sitemaps submissions)
4. on-site factors (identical text, content)

So, for the preliminary examination, we purchased 3 domains, identically named, with different top level extensions (.org,.com,.net). We then created 3 separate subdomains on each of these domains so that we could create some sort of result duplication and randomize the order of submissions to Google sitemaps. Finally, we created identical content on each site and identical sitemaps.

The Results were as follows:

What else is on the web?

Well the introduction of .co domains has had people in a stir but Google did release the following:

We will rank .co domains appropriately if the content is globally targeted. Webmasters will soon have the functionality to be able to specify this by using the geo-targeting options in Google Webmaster Tools.

What about actual listings?

I read a good article on SEOmoz by Randfish about ranking correlation for Google v’s Bing and it does show some interesting information. They basically took the data set for 11,351 search results for various phrases and then compared the ranking elements to show what would be comparably negative or positive in ranking correlation between Google and Bing.

The first image below shows that URL length has a more negative effect i.e. shorter URL’s are more predominately ranked higher in the dataset. it also showed that a domain other than .com was more likely to rank higher. The data does also show that this has more impact in Google than Bing.

Next up is the TLD section this is based on what ranks higher when a TLD has an exact keyword match domain. This clearly shows that .com exact match domain rank higher than other TLD’s.

Finally the post looked at TLD from a generic perspective i.e. not linked to an exact keyword in the URL, this showed that .org domains in general seemed to rank higher than other domains, this is true for both Google and Bing.

What is the percentage spread of domain names on the web?

After further research I came across a domain name report from Verisign http://www.verisign.com/domain-name-services/domain-information-center/domain-name-resources/domain-name-report-june10.pdf this has a lot of good information inside but my particular concern is around the domain percentage on the web. The reason for this is that the SEOmoz data is very good but I wanted to check the data by looking at the percentage spread of domain names.

You can see that .com domains are clearly the biggest share of domain names owned on the web currently and so it’s not surprising that when 44% of domains on the web are .com that you would more often or not see them ranking higher than other domains. It is useful to correlate this information against some test data such as what I have shown above from SEOmoz.

Conclusion

I think it is difficult to gauge the information overall as the SEOmoz dataset only used the 1st page results and dosn’t compare domain further down the listings, however I would agree so far that .org domain seem to have a slight edge on other domains in Google’s index anyway.

I’ve been writing this article over a two day period and the good thing about doing it like this is that you question what your writing, I have revisited the domain that I have purchased which were .org and .co.uk and when doing an exact match text search in Google the .org appears higher even though it 302 redirects to the .co.uk I think in my opinion I would avoid the cheaper domains and always go with either a specific ccTLD or go with a .com or .org

It’s been a good few week since I put a post out on SEO Mad, so I thought I’d give out some of the good stuff

Basically I have been collating data on how to predict how long it will take to achieve 1st page rankings for keywords in Google’s organic search results. For those who are familiar with my blog, will now that I like excel and so that’s what will help me show you what to think about when predicting time-scales in SEO. (Find here a useful resource for On-line marketing excel tips.)

Background

Predicting time-scales for SEO is some what a holy grail, it can make the difference between a business surviving, clinching a good sale or just being able to visualise SEO as a viable marketing method, so why isn’t it just straight forward?

Well simply there are so many factors which go into a website ranking that it is initially impossible to predict time-scales, factors such as:

• age of Site
• Size of site
• No. of links
• Quality of links
• Anchor text distribution
• On page optimisation
• Site speed
• Yadah!
• Yadah!
• Yadah!

To predict time-scales in SEO most marketers use their experience of the types of sites they have dealt with before, what the market is like and then put a potential time-line for when a website will achieve results. This is very difficult as after analysing the competitive landscape it can change very quickly, which of course changes how long it will take to get results.

The Research

So I put together an excel sheet with 13 websites across 60 keywords, I limited to this due to the sheer time-scale involved in collating the data, however I do think that there is some useful information you will find below.

For each website I detailed the following:

• URL
• Category
• Keyword
• Price bracket
• No. of Competing Results
• 12 Months of results
• No. of Months to achieve 1st page
• Increase in Positions (After 12 months)
• No. of links site hard at start
• No. of links site had at end
• No. of links increased
• Site Age

Gathering this data allows us to diagnose a lot of information so lets take a look at what the information says.

Firstly lets look at the graph of all results. (Are tracking tools stopped at 100 so where something was not ranked I labelled it as 101)

(The graph above simply shows all the results over time, the results were taken once each month.)

The above doesn’t really tell us anything apart from how volatile results are, you can see that some results get to the first page and then fall down before finally settling in the top 10 positions.

The Results

So what can we deduce from the results? Let’s take a look at how the increase in links that a website has affects how quickly it will take to get results. This is an “iffy” subject as if you are buying links then producing too many links too quickly can be harmful, however if you carry out a good viral link bait scheme then the number of links you get doesn’t full under the same criteria.

Increase in links

You can see from the graph that the overall trend line shows that the larger the increase in links the quicker you will get results. (Obvious right?) However the graph is quiet erratic and shows you why predicting time-scales is so difficult.

Site Age

You can see by this graph that the trend line is quite straight, I will say that the newest website did take one of the longest times to gain results, but the trend line does show it quite even. In my opinion this is also dependant on the keywords and industry (Visualised later in the post.)

By Category

Above I categorised all keyword by the top level Dmoz categories, so they are a loose fit, but I would agree with the graph in some parts, newer areas such as technology and Internet are highly competitive for natural link portfolio’s although I would say that finance and real estate are very competitive especially for more manipulated link portfolio’s.

Age and Category

This graph simply visualises the categories and site age, to me it shows that the younger sites in less competitive industries get their results quicker. This is backed up below in a graph to show no. of competing pages by category, you can see which categories have less competition.

Competing Pages

The graph above simply shows that on average phrases which have less competition in terms of results per search will achieve first page listings quicker.

Price Bracket

What I mean by price bracket is how much investment each site has put in. You will be able to see in the graph below the average time-scale to achieve listings via price bracket they have allocated to SEO and the number of positions increased.

You can see that the number of months to achieve listings is less per price bracket, this to me is down to simply less spend is required in less competitive markets.

Concluding the data

So if you want to try and predict how long it will take to achieve first page listings you need to think about all the above factors, Category / industry, no. of links, the number of competing pages per search query and how much of an increase in links your site needs.

As an average figure to look at use the following:

It shows the average minimum and maximum values from the data to achieve 1st page listings in Google in terms of the number of months, increase in links required i.e 10 times as many links as your site currently has and how this relates to an increase in positions.

If there are any mathematicians out there who want to add a formula to this I would be grateful @topnotchseo @rikweber