Web analysis: Get the facts

John Kilroy — Wed, 21 May 2008 13:08:42 +0000

It is good to know where your web site visitors come from and what they do after they arrive.
This and other information is available, using special software, from your web site’s access logs.
It can be a good idea to outsource this type of work for reasons of economy
(doing it properly requires investment on several fronts) expertise and impartiality.
If you do then you should expect human analysis as well as tables and charts, and you
should get reports that actually help with the planning, running and assessment of your
web site and marketing campaigns.

Web analysis is how you discover who comes to your web site and what they do there.

It is one of the most important, flexible and useful of all web marketing activities and can help you optimise the design, development and marketing of your website – both in general and in great detail. It is also the quality-control tool par excellence, and provides insights into:

your web site’s large and small-scale development opportunities and priorities;
design flaws in certain pages, or across your whole site, that reduce the number of leads or purchases you should get;
how effective your marketing is;
your customers’ ‘demographics’, overall performance data and lots of other information that helps you build up a picture of who is using your site, when they are using it, where they are coming from and how they are using it.

A web analyst relies fundamentally on your web site’s access logs. These log files come in various formats; some formats are better than others for web analysis, but all hold raw data that can be broken down into an impressive range of useful information. While data mining programs are used to extract the data from the log files, it is the web analyst who decides what to extract and how to interpret that data. The log analysis software can also generate automatic reports, useful for gathering data according to time and date ranges, ‘hits’, the volume of data, geography, etc.

Despite the obvious advantages web analysis gives to a site manager, it is still one of the most under-utilised marketing activities, particularly among smaller businesses who ironically have the most to gain from it. I believe that’s because:

they don’t understand how useful it is;
they don’t have a marketing or development strategy for their web site;
none of their acquaintances do it, so they don’t either (sheep mentality);
it’s not an ‘active’ activity (a bit like financial reporting);

If you are the kind of person who likes to run your business in a haphazard sort of way, ruled by guesswork and hunch, then you probably don’t need web analysis and probably will never see the benefit of it. Yet, for the right kind of business, analysis provides the ‘intelligence’ to help you make good decisions and be a good manager. It is the only available source of website performance data, and people who believe they can manage a business-critical website without it are, in my experience, little shy of deluded.

So what is the right kind of business? Well, if any of the following ring any bells then you should be doing web analysis as a matter of course:

your business relies on your web site, and/or getting people to it;
you are an ecommerce business;
your web site is an integral part of your company and has a development budget;

Web analysis has had some bad press though, and many people have the impression that it’s limited to providing large round numbers of general and semi-useful (except for bragging purposes) information. And unfortunately, they are often right, which is the fault of unscrupulous, if not downright dishonest, web companies who peddle such services. But that is to proper web analysis what Bernard Mathews is to healthy eating.

Web analysis can answer subtle questions like:

how much of my advertising actually leads to sales or enquiries;
which campaigns are the most successful (and which are failing);
do my visitors come back – if not, why not;
what are my visitors interested in;
are my products well-placed, is my shopping cart designed well, what about my enquiry form;
does my web site have any structural problems that prevent people reaching places I’d like them to reach;
which parts of my web site might be most worth expanding?

You can hypothesise use-case scenarios and story-board notional users all you like, but in the end that’s all just a dress-rehearsal. You can live on hunches and guesswork too if you like gambling. But you don’t have to settle for either option because, ultimately, analysis will tell you how good your web site and marketing strategy really are.

Life outside the Google top ten

John Kilroy — Mon, 05 May 2008 13:06:39 +0000

Some if not most businesses undertake web marketing hoping to get a page one listing in Google or whatever search engine they use, which is a bit like aiming to open a shop on Bond Street or the King’s Road.

Sure, companies obviously do get those fancy addresses, but space is limited and there’s only room for so many. And whilst it’s probably a good aspiration, is it actually vital to land that prime spot? Or does missing out mean that the game is over, that it’s impossible to run a business in another location?

Obviously not. If that were true then the Internet would be a whole lot smaller – remember that (conservatively) only about 1 in 100,000 websites¹ can be in the Google top ten at any one time – it’s a rarefied group and in mainstream retail sectors the top ten is a conglomerate of giant brands and directories, neither of whom are realistic competitors for most businesses.

Also, the Internet is a big place getting bigger by the day, and the frenzied drive for page one Google listings is, to a large degree, a case of not seeing the wood for the trees. Of course everyone would love to see their web site listed in the top ten, but it is perfectly possible to do good business with a lower position, via traffic from other search engines (yes, there are other search engines out there), through partnership activities, advertising and most importantly, by providing value (as your search engine position is absolutely worthless unless you provide a service that people want to use).

Then there is the issue of what do users do when the listings on page one are low quality, which (let’s face it) is not unusual. Do they search again for different terms or do they click on to page 2? Research is clear that most users prefer to search again when a quick scan tells them the listings aren’t appropriate, but I don’t know of any research that looks at what happens when users can see they’ve got the search terms right, but that the web sites that the listings lead to are poor. It’s a good question and, speaking personally, if I think I’ve got the terms right, I will keep looking down the listings until I’m sure that Google can’t help me.

This leads to the possibility that the situation might not be so bleak for those web sites that seem mysteriously to remain trapped beneath poor quality competition occupying the spots above them. As I say, this is an educated guess, but no one I know has proved it wrong.

Benefiting from search engine position is simply a numbers game – and you should remember that search engines are most important for generating new customers, not for retaining customers. At the top end, the visitor numbers are so high that web sites can clearly afford to lose customers, and lower down the listings, web sites clearly cannot. The lower down the listings you are, the more important it is that you retain visitors and turn them into customers, by giving them what they want in an easily digestible form.

Providing good content is always the best policy. It is not only an important tactic in raising your search engine position (as other people tend to link to good web sites, and because high-quality text-based information gives the search engine more to chew on), but it ensures that while your web site is making its ascent up the index you are extracting the most from the customers you get. In an environment where resources are scarce, they must be nurtured carefully – providing good content is the best way to nurture your customer figures.

And, if you think about it, a visitor who comes back to your web site is probably better than a new visitor (especially for ecommerce web sites). New visitors are fickle: they are quick to judge, demanding in their expectations, often they’re only there to compare prices or availability, and they know there are usually other places where they can find what you are offering. By comparison, providing you maintain standards, new visitors are worth their weight in gold: visiting times tend to be longer, they will gradually ‘learn’ your site and so discover and use more of it, and though they still visit to make comparisons they are more likely to spend money if your offer seems competitive – having previously earned their trust to some degree. Unfortunately, a great many shopping experiences on the web end in disappointment so trust is very valuable – leading as it does to the holy grail of marketing: personal recommendation.

The point of this article is to remember that no web marketing strategy relies only on a high Google position. Prepare your website for its ascent up the Google charts by making sure you Search Optimize it. Then move on and deal with the reality that Google isn’t going to be swinging you much business for a while yet – through an ‘organic’ search listing anyway (an organic search listing is one that hasn’t been paid for). Now is the time to get on with the other equally important work of finding advertising and marketing opportunities that do not rely on organic listings in Google or other major search engines.

Some people are impatient though. They just can’t wait, and a ‘non organic’ listing courtesy of a sponsored AdWords advert seems so alluring. But however much it may increase your visitor numbers, it is also an expensive and relatively inefficient way to generate traffic, so finding out what your money is achieving should be a priority. This goes for any advertising you undertake, not just a Google AdWords programme. If your web marketing costs you money then, generally speaking, you should be quite interested in whether it is working for you or not.

You can find out where your ad money goes by using some of the other arrows in the well-filled quiver of your marketing programme – you can get regular reports showing where your search engine visitors go and what they do; you can set up target entry pages with user activities in mind (getting them to make an enquiry or buy a product, for example) and because you know what it has cost to get that person where you want them, you can see how much they return to your business. Remember to track the after-effects of a campaign, as intermittent advertising can be very cost-effective; set yardsticks and be brave enough to pause your campaign if it doesn’t measure up.

Many web sites do these things anyway as part of their standard web marketing activities. If your business spends a lot on advertising then you should do them too, you will learn a lot about your web business and become better at managing it.

But getting back to the Google Top Ten. In the shorter term, there are other important and profitable ways to market a web site (and ways a business can use the Internet for marketing). Sure, as a long-term aim, there’s no harm imagining yourself sharing a joke and a glass of bubbly with the boss on the day you hit the top ten, but for now why not explain to him how Google position is more a means than an end (and if you like cliches remind him that there are many ways to skin a cat); impress him with the fact that you do your job well enough to generate visitors under tougher conditions – in the long run, if your boss understands the bottom line, he’ll see you’re actually doing a good job.

1: Based on a notional 1 million results per search; searches I do often return many more than this.
Note.: This article was first published September 2006.

FTSE 100 web site survey, 2006

John Kilroy — Wed, 14 Jun 2006 09:15:27 +0000

This survey assesses the technical quality, usability and accessibility of the homepage on the main web site of each member of the FTSE 100 group of companies.

Kilroy James first surveyed this group, and a much larger group of smaller companies, in 2004 in order to get a picture of the general web design standards promoted and achieved by British companies. Now in its third year the survey has developed greatly, applying up to twenty tests during each homepage examination.

The FTSE 100 survey looks at WCAG Accessibility ratings, W3C technical testing, and the use (and omittance) of numerous usability features. It measures basic good design and technical practice. We view it and encourage others to view it as a benchmark of basic quality, indicating entry-level standards we think all web sites should meet.

Main findings

75% of homepages fail basic technical testing and 86% contain programming faults.
93% do not meet recommended accessibility standards (WCAG priority II). 56% do not meet the industry minimum standard.
Less than a third (30%) provide properly resizeable text.
87% have a search facility on their web site, but only 71% present it on the homepage.
The average score is 38%.

Top of the survey results.

Companies scoring 60% or higher

Rank	Company	Score
1	J Sainsbury	83%
2	Xstrata	78%
3	National Grid	73%
4	Boots PLC	71%
5	British Land	71%
6	Rentokil Initial	70%
7	Imperial Tobacco	68%
8	Royal & SunAlliance	65%
9	Alliance Boots	64%

Bottom of the survey results.

Companies scoring 20% or lower

Rank	Company	Score
87	Carnival	18%
88	Rolls Royce	17%
89	Legal & General	16%
90	Barclays	16%
91	Next	16%
92	Vedanta	16%
93	Kazakhmys	16%
94	International Power	12%
95	Cairn Energy	12%
96	Antofagasta	12%
97	DSG International	10%
98	Persimmon	9%
99	Prudential	4%
100	Enterprise Inns	0%

Interpretation of results

The best homepages meet basic standards and get the important things right. However, 86% do not meet basic standards, which is a poor figure lending a certain distinction to the 14% who do. Beyond standards conformance, the best sites achieve good accessibility ratings and show that the designer has considered Information Architecture and Usability issues.

The survey’s overall finding though is that the majority of tested homepages are poor (some are very poor indeed) and only a few would meet the standards required of a public sector web site, which isn’t very encouraging considering these are the UK’s largest companies.

Methodology

The survey tests the homepage of the main corporate web site for each member of the FTSE 100 and is in two parts: Basic and Extended. The Basic test looks at eleven aspects of a page’s design and implementation including: whether text resizing is properly implemented; W3C markup and CSS validation; and the presence, form and positioning of any search function. The Extended test looks at nine other elements including WCAG ratings.

We do not publish the exact survey methodology but it does account for qualities such as scrolling, title length and text resizeability, and the use of popup windows, splash pages and search functions.

The scoring method reflects the fact that the tests cover a range of faults with varying degrees of scope and seriousness. For example text resizeability is considered to be more important than whether a homepage scrolls or not: because scrolling is only bad if it makes the scrollbar difficult to use (and homepages are rarely that long, plus there are other ways to scroll a page) whereas a web page with unredeemably small text can, to people with less than perfect eyesight, be rendered unusable.

The 2006 survey was conducted during July and August; web pages may have changed since they were surveyed.

Kilroy James » Web Standards

A new look for Romany Granite

XTH – the XHTML to HTML converter for wordpress

Download XTH (v2.01) here

Caveats