The Problem of Conflicting Google Data
Posted: Wednesday, August 25, 2010
by Kevin Morley
SearchPath
With the advent of the new Google Webmaster Tools data (see blog post from earlier in the month), the problem of conflicting Google data comes up. There are now various places where you can get facts and figures from Google: Google Analytics, Adwords, Webmaster Tools, and Google Trends all offer a multitude of data for the user.
As good as Google's tools and data seem to be, they seem to have a fairly serious potential flaw: the fact that the various data sources can conflict.
foster care - 49,500
fostering - 135,000
So its clear that according to Google's Keyword Tool, the number of people searching for "fostering" is almost 3 times that that are searching for "foster care".
Compare this with the data from Google Trends - if you type "fostering, foster care" into the search box for Google Trends, the resulting graph says the traffic numbers for "foster care" is over twice that of "fostering".
This clear discrepancy is puzzling - the error is not 10 or 15%, its over 100%. Considering that both sets of data are coming from a company that is renowned for its data accuracy and the size of its server farms producing billions of accurate search results every day, it seems odd that in the above example the results would be so diverging.
The best way to guard against these sorts of Google data conflicts (especially when doing keyword research, which is so crucial in SEO) and limit discrepancies when taking data based decisions is to cross reference your data, to use a third or fourth source, if possible. For example, using Wordtracker and/or Keyword Discovery in addition to the Keywords Tool and Google Trends when doing keyword research would be good idea, in order to reduce the chance of errors.
Internet World 2010
I went to the trade show Internet World at Earls Court 2 in London yesterday, and saw some interesting seminars. Here are the main takeways I got from the seminars I saw:
1. I saw one guy give a good talk on forecasting for SEO: this can often be a problem when pitching to clients. After all, it can be hard to predict conversion rates and CTR on organic results, plus it can be hard to give definite gurantees on traffic
The guy I saw had drawn up a spreadsheet from various data sets. in order to forecast how much profit a customer is likely to make. First in the left hand column he had the search term, then the estimated traffic from the Keyword Tool in the next column. He then had the estimted CTR, based (I think) on the new Webmaster Tools data.
The next column was an estimated conversion rate, based on an internet average of 3%. The remaining columns in the spreadsheet were the order value, than the profit in the last column.
Laying the spreadsheet out in this way enables internet marketers to make (admittedly estimate) predictions and forecasts for clients' possible ROI, if they were to come and work with you.
2. People who search for long tail terms are much more likely to click on position 1. The fact that these phrases convert well makes it all the more important to rank highly.
3. Search is about the past, social media about the present.
4. Keyword clouds are a useful tool for determining what a given page is about.
5. When you've got a very big site (with thousands of products) it is not realistic to SEO each page manually. This is where automated tools come in.
6. At Dell, long tail terms gives 83% of traffic and 67% of sales.
7. To SEO a site properly, you should make sure you have unique descriptions on every page.
These were the main points I took from my visit to Internet World. Its a free show, and runs for the next two days at Earls Court 2 in London if you are interested in attending.
For more info visit http://www.searchpath.co.uk
This Article has been viewed 204 times. (Not updated in real-time.)
No comments yet.We want your comments! If you can read this, you don't have javascript enabled, so you can't use this comment system. Please enable javascript.