Measuring SEO success 2 – What really matters

An earlier post looked at why search position reports still have a place; although it should be said that this is changing rapidly as the search landscape changes. This one goes beyond that and looks at why rankings are only the first step; what you should be aiming for with your website and what you should be measuring to assess it. This post and others to come are aimed more at site owners rather than SEOs, who really should already know this stuff if they are looking beyond a very narrow definition of what SEO is. However it’s so easy to get bogged down in technicalities that it never does any harm to refresh your mind on why you’re busy working on projects whether it’s recoding sites or chasing links.

Profit – ultimately the only real measure that counts

At the end of the day any commercial company needs to make money and its website has to contribute to that. There are different ways, some direct and some indirect, in which it can do that. All of these ways depend on traffic.

The basic value of traffic

To some extent traffic alone can give you some revenue via advertising impressions. It can also contribute to your company’s PR and branding presence. How much value you place on that depends on you and your particular market and whether the site is purely standalone or part of a wider operation. (As an aside it always surprises me to hear businesses that spend thousands on Yellow Pages and other conventional print advertising, considering it well spent despite little direct evidence of return, who will dismiss the advertising value of their website and insist that they only count direct sales against it.)


However the real value only starts to kick in when you have some sort of conversion. That could be a small contribution if it’s that the user clicked on a paid advert or affiliate link and left the site. Or it could be a subscription or a sale for a large value item. But it can take many other forms. It could be that it generates a lead that produces a profitable business to business collaboration – do you have measures to assign such a lead against your site?

It’s also perfectly possible for the money for a direct sale to be taken at another stage – your “conversion” may be that the user orders a catalogue or requests a callback which allows your sales team to interact with them as a likely prospect – so you have to define your conversions carefully and assign your successes accordingly or you’ll have no clear idea of what’s working for you and what isn’t.

Sources of Traffic

Assuming you’re running a commercial website (and few non-commercial sites would be spending money on detailed analysis unless perhaps they were a charitable organisation which had to justify it’s web spend) then your traffic will come from one of the following sources:

  • Organic search results
  • Paid search results
  • referrals from others sites either by links and recommendations or through paid advertising
  • direct visits from people who know your site from previous visits
  • direct visits from people who’ve discovered your address via other channels such as TV or printed media.

Measuring Traffic

To get an idea of what is working on your site and to try to improve matters you need to measure your traffic and understand how it behaves in the site. There are many ways of measuring traffic, but unfortunately very few of them are accurate. This is an area I’ve studied extensively and specialised in for a while so I speak from direct experience.

Hit Counters

Back in the “old” days you’d see garishly styled hit-counters on many sites. These were generally based on a very simple bit of JavaScript and were hilariously inaccurate and easily fooled.

Online Stats Packages

Many web hosting packages come with free online traffic recording statistics which attempt to analyse the log files which should be recorded on your server – they’re usually worth exactly what you paid for them because most of them can’t differentiate between search engine spiders, site scrapers, and real human traffic. Even the better ones such as AWStats which give well-formatted multi-page reports are really only guessing at what’s happening. They can give some useful information in some areas but should be approached with caution and used more as an indication to further detailed study.

Logfile Analysis Programs

Then there are a range of reporting tools which run on your local PC to analyse those same server log files once you’ve downloaded them – programs such as Webtrends and SurfStats. Such programs can be much more detailed as there is no problem of load on the web server and they can be configured to some extent and updated when changes occur. However the nature of the logs means that in order to derive meaningful figures the designers need to make certain assumptions about what constitutes a page view or an unique visit, and there are complications due to such things as the caches used by the major ISPs. The result is that the major programs in this field seldom agree with each other and you can only compare reports derived from the same program.

Analytics Programs

These programs are the usual choice for the most detailed information – they all utilise some form of tracking system for each page of your site. This can be an embedded link to a small graphics file or a JavaScript program such as is used by Google Analytics. Commercial examples will usually have a little more detail in their reports but obviously have a direct cost. Free programs such as the one from Google are viewed on-line and thus of course dependent on their servers being available. You also have to consider that while the commercial ones will usually regard your data as confidential, using the free ones means you are sharing your data with the supplier in a much more open-ended manner.

Whichever of these analytics programs you choose there is much more data available than with the other systems and on the whole that data is more reliable, though it should be said that no analytics system is ever completely accurate.

You’ll be able to see the number of visits, unique visits, length of time on site and on individual pages. You’ll be able to see which keywords were used to arrive via organic search or pay per click. what the bounce rate was and you can usually set up “goals”, which can record activities such as making a purchase or requesting a callback or a brochure.

It’s these measures that are far more important than simple search rankings in determining how successfully your website is performing as well as giving clues to how to improve it.

Once you can read and interpret this type of data you’ll be in a much better position to understand how users react to your site and to draw conclusions about how you can improve it. For instance there may be a keyword phrase that you have a very high ranking for in Google. However the traffic that arrives from that source may be spending very little time on the target page and a large percentage may be bouncing off without looking at other pages. Close investigation of the analytics data may allow you to understand why and to take action to improve the page that the users are landing on.

A good SEO company should be able to not only achieve rankings but also improve your overall traffic, your understanding of how your site works, your conversion rate, and ultimately your bottom line.

Comments are closed.