Data is the sole method through which one can make any assessment as to the success of an SEO campaign. The kind of in depth data analysis facilitated by Google Analytics and other tools is one of the core components that differentiate digital marketing from offline marketing and TV advertising. It is also both intrinsic to the success of a campaign and the strongest selling point for any honest agency.
To examine this assertion in more detail:
Data is intrinsic to the success of an SEO campaign
It is only through analysing and understanding the data generated by and pertaining to your SEO campaign that you can begin to understand the impact the SEO is having. If you do not know what traffic volumes your website is receiving or what the sources of that traffic are; then you will not be correctly positioned to be able to understand the value that the SEO is having.
This leads me onto the second point…
Data is the strongest selling point for SEO
Data can completely validate the success of a campaign and conversely can prove the ineffectiveness of one. If you understand what the data is telling you, it becomes possible to make an accurate assessment of whether or not the work is working. It is because of this that such high levels of transparency can be provided by an SEO agency.
This is why data is so important at a fundamental level for both a client and an SEO consultant.
Top level data such as total website traffic can be misleading if taken either outside of any context or in the wrong context altogether. To illustrate this point with a real world example; if you had instigated an email marketing campaign in February and this delivered 1,000 website visitors… However your SEO agency provides you with top level traffic stats only, which reveals that traffic went up in February compared to January by 1,050 visitors. This would, using a cursory glance, appear to be a very significant improvement and thus prove the value of your SEO. The reality of this situation is that traffic increased by 50 visits, once the email marketing campaign traffic is deducted.
Furthermore, and continuing the example above, where did those 50 visitors come from? Were they from search engines, were they direct visitors, were they delivered from other websites? Have you excluded your own IP address from the analytics data, if not 30 of those visits could be you! Are you looking at visits or unique visitors? Understanding all of these aspects and many more will shape the very meaning of the data.
Another simple and effective example of misleading data is this; let’s assume that traffic in January reached 1,000, but in February traffic remained at 1,000 visitors. On the face of it there is no obvious improvement month on month… BUT, February has only 28 days and January has 31 days, SO; in January the average visitors per day was 32.2, whereas in February the average visitors per day was 35.7 AND this means that traffic actually increased by 10.7% when comparing a like for like period of time.
What is good data?
This is the very crux of this article, good data is useful data. Good data should always reveal rather than obscure meaning. To some extent all data has the potential to be useful but it must be supported by other data and your analysis of that data should consider all contributing factors.
Breaking down top level data into the following categories is a very useful way of understanding the sources of those data and how they relate to one another.
Total Traffic | Conversions
Organic Traffic | Conversions
Organic Traffic (excluding brand name related searches) | Conversions
Direct Traffic | Conversions
Referral Traffic | Conversions
The above list of metrics through which to analyse the top level data is a great starting point, you will be able to assess how much traffic has come from what source and how many conversions are attributed to it.
The metric “Organic Traffic (excluding brand name related searches)” is acquired by filtering out keywords from the organic traffic that are related to the name of the company. This is a very important metric as it pertains specifically to the traffic to the website coming from completely natural search.
Conversions and the tracking thereof, where possible, should always be implemented. Conversions are nearly always very close to the bottom line for clients and are usually the best measure of a campaign’s success. Conversions (or goals) can be very easily setup in both Google Analytics and AdWords, either by using the existing tracking code or by adding some to a confirmation page.
As mentioned above though, one must understand where these conversions are coming from not just from the generic source but also from what keywords. This can reveal how people find your site prior to converting, and thus make you efforts more targeted. Traffic is vanity compared to conversion data, much the same way turnover is to profit; it is possible to deliver thousands of visitors to a site but if they do not convert then they offer no value to the website owner.
Use analytics to look at conversion funnels, see where people drop out and try to establish why they drop out. Understanding why conversions fail by analysing the data for that can be equally as valuable as seeing what converts and when.
Again be mindful or wary of sweeping general statistics like; “your site converted 60% more visitors in May 2012 compared to May 2011″. As I have illustrated above this kind of data is fine so long as it is accompanied by a more in depth look at what is happening. Were those conversions from organic sources? It is possible that your conversion rate jumped up after a website redesign or rebranding launch, or some other marketing endeavour?
Compare data sets
Comparing data sets is another valuable statistical trick that can reveal hidden treasures in the data. For example, compare your keyword rankings and the traffic for those keywords; this could uncover some gems… Keywords which aren’t ranked well but deliver a lot of traffic and long tail traffic could be worth your attention during your next retainer. Go further still and compare competing websites, traffic and rankings to get a feel for what keywords will deliver the best results in the shortest time.
Comparing PPC data to organic data and ranking data could also yield valuable results; see if you are bidding on keywords that you already rank number 1 for organically. If you are it is possible to run some tests, to see if the additional paid clicks that you receive are worth the money.
The aim of this blog is to stress the importance of understanding your data and I hope I have given some good examples of what is good and what is bad data. Nothing short of clarity is required when dealing with data, in part because data / statistics of any type are easy to misrepresent or obfuscate the true meaning behind them. Equally, it is easy to misunderstand data by not looking at it either logically or thoroughly.
If you have any examples of dodgy data or some good examples of useful metrics please post them in the comments below.
Green matrix background via BigStock