We’re all familiar with the idea that there is only six degrees of separation between almost any two people. By moving from one person’s circle of friends and acquaintances to the next, six times over, any two people can, theoretically, be associated together.
Of course it didn’t take long for someone to wonder how this might apply in the virtual world. In 2000 the notion was reinvented to assert that there are typically just 19 clicks of separation between any two websites.
Making sense of people’s information-seeking behaviour has received a lot of attention in recent years. Those of us interested in marketing and optimising websites have taken a particular interest.
Online marketing is all about promoting a website, while search-engine optimisation (SEO) is the name given to the fine tuning of a website—improving it in such a way that search engines such as Google will be more likely to find your site—ideally ahead of any competitors (see Courting Google, July T&B).
A well-used axiom in marketing circles says something like: “I know that half of my advertising spending is wasted, I just don’t know which half”. In the online world people have quickly begun to see that by making an adjustment to a website—optimising it (or by running a promotion or ad campaign) and then monitoring the number of hits the website receives—they can actually begin to measure its impact.
Web analytics, as it is called, is the measurement of the behaviour of visitors to a website. Its great advantage is that it is able to provide real data relating to how people are actually using a site, so rather than relying on your sense or guesswork for marketing and promoting your site you can instead operate on an empirical basis.
SEO combined with analytics becomes a very powerful tool. In the online world the question of “which half is wasted” suddenly presents less of a problem when analytics tools are in place. Every interaction, every event on a web page can potentially be tracked.
Half of the firms surveyed by Forrester Research in late 2005 said they expected the number of people using analytics within the company to grow by an average of 23 percent. The research company expects the potential of web analytics to continue expanding, with frontier businesses integrating analytical data with other more general customer analysis platforms—“customer analytics” as Forrester has coined it.
In an August report from Forrester titled The Business Case for Web Analysts, it described web analytics as “must-have technology—a real no brainer”. The case it goes on to argue is that you should actually employ a dedicated web analyst to really try and take advantage of analytics tools. It develops two models showing the range in benefit, one for a retail-based site and one for a business-to-business site.
Forrester’s conservative estimate shows ROI benefits for the retailer of more than 1000 percent, and more than 550 percent for the B2B example—after one year.
Even with some compelling benefits many businesses still lack anything more than the basic tools.
“We see companies using some of the free tools out there that have been around for six or seven years that just give very basic information,” says Tom Petryshen CEO of web consultancy firm Amplify. “Often it is something that their web provider has given them [that] they really don’t do anything with. The percentage is very small that are actively using analytics to measure success online,” he says.
Other analysts agree. Australian businesses do not seem to be really leaning on the technology as much as they could.
“We’ve seen very little published information about how people are using analytics,” says Bruce Arnold, director of Internet research and analytics firm Caslon. He suggests that small- to medium-sized enterprises are probably not collecting statistics themselves.
“They are relying on one of the commercial services or their ISP, and they’re certainly not making sense of the data. It’s almost a sort of an osmosis process. There is this sort of feeling out there that it’s enough to just collect this information. Most businesses are not going beyond that to what should be the real target; understanding our market and who is visiting our site” he says.
This is the key to using analytics with real success. For many businesses, data is collected but not really understood or used. It becomes what Arnold likens to “ego-ware”, or what Bill Gassman, research director at Gartner, describes as “comfort food”. A pie chart that shows basic traffic data is a nice reminder to have but it doesn’t help you to drive effective campaigns, understand your market or show you how to improve your site.
Information can keep you comfortable but it needs to be action oriented according to Gassman. “It’s the same with any business intelligence tool—info by itself is not really worthwhile, you really need to have a reason for it to be there,” he says. “You end up with a very powerful tool that can show you in black and white what’s going on.” Bill Gassman, Gartner
Gassman describes the maturity of web analytics across five levels. At the most basic are operational necessities—virtually every site would have these features.
This includes looking at things such as browser statistics—how many visitors are using Firefox or Internet Explorer—links and who it actually is linking to the site. Organic search keywords that people are using can also be measured, or perhaps where people are entering and leaving the site. This sort of monitoring would tend to evolve naturally as the site matures. The data is often attainable simply from your ISP or by simple counters.
Once you move beyond these basic metrics the potential of web analytics kicks in. The second level Gassman describes as behaviour optimisation. This involves looking at the site itself and investigating where people are going and where they are getting lost, or maybe even looking at multiple sessions—following a user as they browse, leave and return later on.
A/B testing can also become a feature at this point. This testing involves using a load splitter to divide your web traffic in two so that half the traffic will see one version of your site while the other half will be directed to a different version.
A basic example of how this can be used would be to make site A’s background blue, while B’s is red. If blue gives you a better result this change can be adopted. The focus is really on user’s behaviour in order to optimise the site. Gassman roughly estimates just over a third of businesses would be undertaking this sort of analysis.
The third level, emarketing, still ties in with behaviour optimisation but focuses on the best ways of actually bringing visitors to the site.
Practices such as buying appropriate keywords on search engines are well known, but analytic tools need to be used to help make decisions about where the money should actually go. It can tie into email campaigns or banner-ad campaigns—the analytics should be able to reveal which are the best options in your unique case. It also embraces areas such as merchandising—if people are buying dips they’re probably going to need chips as well, so what else can you do with that? Amazon.com is a great example with its “your peers who bought this also bought . . .”
At this level you will also see analytics used to segment visitors. Learning about them enables you to run analytics according to these slices. So, for example, it can begin to reveal if older women are buying my product more than younger men. Or what is your customers’ geographic location, gender, age, income and so on.
With this information you can begin to really understand your market. If you run an ad campaign in the eastern states of Australia, for example, you can measure its impact at your website. According to Gassman, certainly less than a third but probably closer to 20 percent of businesses will operate in this way.
The fourth level and still pretty much the frontier of web analytics, according to Gassman, will see businesses begin to consider integrating analytics across channels. A recent Forrester report shows that: “Firms improve their ROI acumen when they follow customers who visit the website, don’t buy immediately, and then consummate their transaction through other channels.”
It explains that while businesses are most comfortable producing basic reports many are exploring some of these more complex uses. This includes sharing data with other channels—importing information from other sources (such as point of sale), and pooling data to build rich, multi-source customer profiles.
Forrester states that while integrated customer profiles are really only on the horizon for a lot of firms, 60 percent of those surveyed engage occasionally in at least three (of eight) core examples of what it views as this kind of “customer analytics” level activities.
Over the horizon Gartner’s Gassman sees the top level of web analytics simply as part of corporate performance management. This is where the web channel merely shows up in an executive-level report as simply another available channel. The Web becomes integrated with all the other business intelligence systems so you can do activity-based costing, really understanding if it is easier to pursue customers through the Web or another way. Typically only the top one percent of businesses will be at this point, says Gassman.
The maturity of analytics in Australia does—like ecommerce—lag a year or two behind the US. Amplify’s Petryshen suggests that it is due to a combination of a lack of knowledge and the fact that people are just too busy running their existing businesses. “Some businesses that are really successful are still just going with the flow rather than looking at the data they are often already getting and trying to make sense of that,” he adds.
For some businesses, just getting started can be a challenge. Many aren’t sure of what to track and what information to look for. Petryshen says that when he embarks on the process with a client he tries to identify relevant pieces of information that can be used to track trends, really starting with the basics. “Information such as visitor numbers from one month to the next, how many new visitors compared to repeat visitors, how many page views did those visitors look at—that kind of information, over time, will tell a story and you can at least identify if you’re moving forward or taking a steps back,” he says.
Once there is a fundamental understanding of this it becomes much easier to really move into some of the deeper levels.
The analogy made by Caslon’s Arnold is that it’s like having a sports car—you have to know how to drive it. “The problem, certainly for many SMEs, is that in most cases they don’t really go beyond the entry screen of their analytics tool. If they’re enthusiastic they’ll have a look at it once a month and see a couple of pie charts—end of story. They won’t start digging into the data, which is where the real value of this stuff comes in. Otherwise its just ego-ware,” he says.
To move far beyond this, the number one challenge is finding skills. It requires quite a specialised skills set and at the moment there is no standardisation between different vendors approaches. Gartner’s Gassman says that it is often hard to find the people who can really use the tools.
Regardless of what analytics systems you’re using the difference, adds Petryshen, is the people you have. “How can you actually take that data and make sense of it and tie it into your business goals? I don’t believe Australia has a lot of people who are skilled at that,” says Petryshen.
Gassman says the other great challenge is having the right processes in place to actually make changes. “If your analytics show that when the background is blue you get a five percent better conversion than if it’s green, how long does it take you to change that from blue to green? In some organisations it can take months. You end up with a tool that is very powerful and can show you in black and white what’s going on but people can’t react to it.” Gassman says.
The tools themselves have evolved a long way in recent years but issues still exist. Forrester’s January evaluation of analytics products describes the market as a decade old, but yet to settle down. A “proliferation of vendors” has made it difficult for users of analytics tools to focus on one product amidst the “sprawl”.
Gartner details somewhere near 60 vendors in the analytics market, although Gassman believes only 10 actually demonstrate real cultural knowledge and do the hard work to understand people’s needs.
Vendors with hosted application models have embraced web analytics, some of the big players including NetRatings, Omniture or WebSideStory all operate in this fashion. Others such as WebTrends offer both licenced and managed versions, and Google’s hosted package is available completely free.
The main advantage of hosted products is, of course, their low IT resource requirements, while licensed in-house products offer far greater flexibility and capability for integration.
As the market settles some of the confusion may dissipate—Google is forcing the pace but for the business looking at tackling its own website some broader concerns could also become apparent in time [see sidebar].
The bottom line, however, is very positive. With the likes of Google there are some really easy and cheap ways for a business to test the waters using enterprise-class products.
In Australia, the initiative to adopt analytics-based websites is really only now gathering momentum. The time is ripe, there are great opportunities to really drive the web channel while competition is still lagging.
With the growth of ecommerce generally, the potential for innovation with analytics is really only beginning to be tested. Web analytics is already proving to be one of the most useful yardsticks available to a business, but time is also beginning to show that it is one of the most fundamental.
The rapid rise of web analytics has had its side effects. Like any industry that quickly gains notoriety and popularity, the scramble is usually to just get on board and get it going before those concerned with the industry’s situation can really begin to consider its long-term role or best practice.
One growing issue is a lack of benchmarks according to Caslon’s Bruce Arnold. “There is a lack of standards across the industry in how this info is collected, in how it’s handled. There are still basic disagreements about fundamental definitions and a lack of a meaningful code of practice. For example, say you are talking about print advertising, there is independent monitoring you can trust that that particular paper has a circulation of so many thousand. You can trust that figure because it’s independently audited. With web advertising and web analytics generally, there is no such auditing,” he says.
The problem becomes clearer still if you compare the same month through two different web analytics systems. You are likely to find that there are quite significant differences in the results simply because of the differing ways that the systems collect and track the data.
Every system will have different guidelines and approach the data collection from different viewpoints. Where one might view a visitor who remains inactive for an hour as an entirely new session when they become active again, other vendors might see this as the visitor continuing the session.
The great difficulty is that vendors view the way they calculate metrics as a source of competitive advantage, and so are uninclined to make their methods public. Amplify’s Tom Petryshen says this is something that really needs to change. “It doesn’t help the industry and it doesn’t help the companies I deal with. I think it’s a way of locking people in,” he says.
In the US there are efforts to establish a professional organisation to address some of these issues. Most of the members of the Web Analytics Association (WAA) for example, are actually analytics vendors, but how effective such a body can be remains to be seen.
There is also a rapidly growing global consciousness about data privacy. A lot of these issues are starting to appear and because the web analytics environment is fairly new and dynamic, there is very little established in terms of industry code of practice. Just what could happen to data is not certain, and there has been a number of privacy-related legal settlements in the US surrounding analytics and the use of cookies. Cookies received plenty of attention at the time of the legal action (2002) and some perception problems still reside for some users who see cookies and analytics as stalking or a form of spyware (and anti-spyware vendors seem to somehow encourage this).
Last year, a Forrester report titled Web Analytics Cookies: Perception Versus Reality, recommended that businesses adopt a variety of best practices designed to allay many common users’ fears and “perceptions”. These included suggestions that businesses switch from third party to first-party cookies (cookies from third-party analytics vendors attract far more anti-spyware attention), being open and transparent (rather than vague about what you do with data collected), offering value for users to retain cookies (rather than deleting after each visit—enabling tracking over repeating visits), and even offering a separate company analytics policy in addition to a company privacy policy. As a last resort customers could also be offered an “opt-out”.
In the long run, these issues will emerge and be tested. A key stage in the maturation of the industry, the sooner best practice and industry standards can be established the better.
About Technology & Business Magazine
Technology & Business MagazineTechnology & Business is Australia's premier enterprise technology title, providing useful and high-quality news, reviews, analysis, interviews and opinion on issues key to Australian business and IT leaders. Technology & Business is the only title to provide independent lab testing and analysis of enterprise products, looking not only at performance but also interoperability, future proofing, return on investment and service as testing criteria.
Visit Technology & Business Magazine Online (http://technologyandbusiness.com.au/)
Subscribe to Technology & Business Magazine (http://magshop.com.au/Technology_and_Business_Magazine.htm?CID=284)
(12 months subscription for $24.95)