ellipsis flag icon-blogicon-check icon-comments icon-email icon-error icon-facebook icon-follow-comment icon-googleicon-hamburger icon-imedia-blog icon-imediaicon-instagramicon-left-arrow icon-linked-in icon-linked icon-linkedin icon-multi-page-view icon-person icon-print icon-right-arrow icon-save icon-searchicon-share-arrow icon-single-page-view icon-tag icon-twitter icon-unfollow icon-upload icon-valid icon-video-play icon-views icon-website icon-youtubelogo-imedia-white logo-imedia logo-mediaWhite review-star thumbs_down thumbs_up

What you need to know about web analytics

What you need to know about web analytics Brandt Dainow

Around the middle of each year, a survey is conducted into the state of web analytics by training and publication group Econsultancy and web analytics consulting firm Lynchpin. Together they've been running the survey each year for the last four years, tracking the changes in the web analytics industry from year to year. Their survey offers an opportunity to watch the development of the field of web analytics in general.

Multiple products
When looking at the technological aspects of web analytics, the survey found that most companies use more than one web analytics package and that Google Analytics is becoming ubiquitous. Eighty-seven percent of those surveyed use Google Analytics, even though half of them also have an enterprise solution such as Omniture or Webtrends.

Stephane Hamel is director of strategic services for web analytics consultancy Cardinal Path and a senior member of the International Web Analytics Association. In his view, people use Google Analytics even when they have other products because Google Analytics is the easiest to customize and offers the best range of ancillary skills and services.

"Even if they have a commercial product, people still go to for products which are easy to use and where resources are available," he says. "Google Analytics (GA) is easier than other products for instant customization and segmentation, even if it is less accurate. In addition, there is now an ecosystem of GA skills and products."

In my own experience, the easy availability of the Google Analytics API is a significant factor in the development of an ecosystem of products and services based on Google Analytics. As a result, Google Analytics can be integrated into a wider range of other systems than any other web analytics product.

Reconciliation and data integration
Data integration is the major problem facing the web analytics industry at the moment. As management and other branches of the organization become more web analytics aware, they are starting to ask questions of web analytics relating to their own spheres of activity. This trend is enhanced as web technology permeates a wider range of business processes and thus gives rise to a wider range of measurable activities. The most obvious example at the moment is measurement of social media and mobile. While we all know about the importance of mobile, few companies pay any attention to it. While mobile now forms at least 10 percent of traffic or more (peaking at 25 percent in India), only 20 percent of companies are measuring mobile traffic. Social media cannot be reliably measured by most web analytics systems at all. As a result, the social media measurement field is populated by an emergent ecosystem of startups. This inevitably leads to difficulties integrating data into existing analytics systems. Furthermore, we can anticipate a period of consolidation as the market matures. This places some degree of risk on the choice of product used to measure social media and other emerging areas such as media streaming.

This risk highlights another key issue confronting web analysts. Changing analytics products or adding new ones is a major challenge with current technology. There is a complete absence of technology standards equivalent to what we find in other web fields. There is no web analytics equivalent of HTML. Unlike other branches of web technology, there is no XML dialect for web analytics, which could serve to standardize data formats. As a result, shifting from one web analytics package to another while retaining the data is almost impossible.

For the same reason, exchanging data between web analytics packages or integrating data from multiple sources is almost impossible. Reconciliation (the integration of data from differing analytics sources) is currently the major challenge facing web analysts in their day-to-day jobs. The survey found that most data integration is done by hand. My own experience is that I spend much more time working in Excel, cross-referencing data from multiple sources, than I spend looking at dashboards in individual packages. The survey shows that many of us are in this position.

What to measure?
There is also the challenge of working out what to measure. As we do more and more online, a wider range of activities need measurement. Web analysts today are confronted with a dazzling range of metrics. They are unsure what metrics matter, and they are unsure how to report these to the decision-makers in a manner that will make them take notice. This problem has not really improved in the last 10 years.  Jim Sterne has been running the eMetrics web analytics conference around the world for more than a decade. I remember that the principal problem confronting web analysts discussed at one such conference many years ago was how to identify the key metrics that management needed to consider. While members of management are more interested than they used to be, Sterne says identifying key metrics hasn't improved. "We are getting more senior managers at the conferences now," he told me. "But they're frustrated because they want to know what they should measure."

In other words, management knows it should have web analytics, but doesn't know how to make use of it.

The general consensus amongst those I discussed this with was that this was primarily due to the way we approach web analytics. We're overly focused on what we can measure and tend to ignore what it is we actually need to measure.

Hamel put it like this: "We don't need to measure everything. Why measure things you can't change? I am only interested in stuff which helps the bottom line. Give me a business challenge first and data second." He continued, "Being flooded with data is a sign of having of not having clearly defined objectives. People don't define smart objectives. It's not easy to do, but once you have them you can limit your scope to what matters."

This is echoed by Andrew Hood, MD, of Lynchpin Analytics, who helped design the survey. "At the moment businesses start from what they are measuring, but they should start from the top down by defining key business objectives first, then working out how to measure them," he said.

I had this experience recently with a client -- it was being presented with a 91-page Webtrends report each month by its IT department. By understanding what its current online business objectives were, I was able to show the client only needed to read one page of the report; the rest was irrelevant.

The cause of this problem was that the managers who used the data had no involvement in designing the report, while those who designed the report had given no thought to what it would be used for. In Hood's experience, management involvement in identifying the metrics that match business objectives is the key and has numerous benefits. "Senior managers are more likely to trust metrics if they've had a role in developing them, because then they understand their accuracy and limitations as well as their meaning for the business as a whole," he says.

Bryan Eisenberg, author of "Call to Action" and "Always Be Testing," is one of the founders of web analytics, and arguably the most successful consultant in the field. He thinks part of the problem is that web analytics software is sold as a solution, rather than a tool. "People don't make money having web analytics tools, but from making business decisions based on web analytics tools," he says. "Who cares what metrics it provides -- if web analytics isn't generating a to-do list, it's a waste of time."

In my view, this is really the key to the problem. Too many people see creating their website as the end of the construction process, and that once built it will remain fairly static forever. In reality the only way to get best performance from a website is to constantly improve it. No one starts at their best, you improve as you learn (if you learn). If you don't plan to change the site, you can't really use any metrics -- there's nothing to focus on.

As Eisenberg put it, "For web analytics to be worthwhile, it needs to be business-changing. I make testing an integral part of web analytics -- testing and experimentation. Whatever change you make is an experiment in a continual process of optimization. A handful of changes each month is not enough to be business changing. Amazon is typically running around 200 tests at any one time. Most companies don't do this because it's hard work and managers would rather throw money at things instead."

Organizational issues
Eisenberg's description of Amazon's 200 concurrent tests raises the issue of organization. Not all websites are technically capable of running 200 tests at the same time, but that's just a software issue. With sufficient budget and will power, any site could do this. However, how many organizations are structured in a manner that would enable them to implement and coordinate such a high level of activity?

Many of the web analysts I spoke with now distinguish between data-centric companies and the rest. Data-centric companies base their e-commerce decisions on web analytics data. Eisenberg cites HP, Dell, Amazon, and Expedia as data-centric companies. I would also include Yahoo and Lovefilm.com, simply on the basis that web analytics is a board-level role in these organizations.

Organizations that put web analytics at the center of their business processes are doing so more and more. The survey shows that organizations with annual web analytics budgets over $100,000 are increasing their spend, while those with budgets below $100,000 are holding expenditure steady or decreasing it. 

Data-centric companies are also scooping up all the good staff while the size of web analytics teams is growing. This makes it harder for less data-centric companies to maintain good web analytics teams, so there is a move to outsourcing web analytics to specialist agencies. Thus organizations that are not data-centric are actually losing their own web analytics capabilities.

Andrew Hood thinks this is caused by the widening role of web analytics. "These days you need a multi-role team, from JavaScript programming to Google Analytics skills and from database analysis to presenting to the board on business strategy," he says. "This represents an opportunity for the consulting market for those businesses which cannot afford such teams."

The place of web analytics within the organization is also changing. As web analytics becomes of value to departments outside marketing, it needs to find a place within the organization that suits this wider role. It is moving into the department most experienced in connecting business objectives to numbers -- finance. Both Jim Sterne and Stephane Hamel told me they think a key reason for this is that web metrics has become part of shareholder communications.

Andrew Hood thinks finance is the logical place for web analytics. "Web analytics should have a direct line to the director responsible for online performance," he said. "Finance makes sense, though not if the finance director is incapable of understanding online. I would say that as a general rule web analytics should be part of finance unless there is a good reason not to place there. Web analytics represents a crossover between technology and marketing and should therefore be independent of both. Finance makes sense because it is used to relating numbers to business performance."

Hamel thinks we should see web analytics as a branch of business intelligence, stating that the topics at traditional business intelligence conference are exactly the same as those at web analytics conferences. As universities start to include web analytics in their degree programs, we see that it is introduced to business management courses, not marketing or IT.

Hood agrees. "Whether there'll be a web analytics industry in five years, I don't know," he says. "I think it could become the part of business intelligence that just happens to have an online focus -- I don't know if we'll be able to call it a separate industry then."

Hood thinks this is also a product of growing management awareness of web analytics. "There is far more maturity around boardrooms about web metrics," he says. "This puts pressure on web analytics departments to do more analysis, rather than simply reporting stuff because we can measure it."

While awareness of web analytics is growing, what is notable is the lack of progress within the industry itself. Linus Gregoriadis is research director at Econsultancy and supervised the survey.

"It's encouraging we've got more companies that want to engage with web analytics," Gregoriadis told me. "What's discouraging is that it's becoming harder to draw actionable metrics from all the data. The issue of getting actionable data has not improved since we started the survey four years ago."

I, like most of those who've been in the web analytics industry since the 1990s, am surprised at the lack of development in web analytics software. Most of the major vendors 10 years ago are still the major vendors, and their software hasn't improved very much. That's not to say that it hasn't improved at all, but if you contrast the state of other branches of web technology in 1999 and today, you see radical change. By contrast, improvements in web analytics systems are relatively minor.

Hood thinks this is the result of where the R&D investment has focused. In his view, web analytics vendors have focused on making systems appeal to people who don't really work with web analytics on a day-to-day basis. There has been too much emphasis on pretty pictures, ease of deployment, and finding ways to justify the investment. These are all signs of products that must appeal to people who won't use the products, don't see the point, and don't really understand the task that the software is designed for. If you have to sell to such people, what else can you do but focus on stuff that doesn't require any skill to understand, such as fancy graphics? Furthermore, it is my own opinion that web analytics vendors are hampered by their lack of interest in common standards and protocols. Vendors have never bothered to form an industry body to agree detailed technical standards by which they can all comply. What few standards have been produced have rarely, if ever, been acknowledged by vendors, and I know of no case where a vendor has changed its software to comply with standards. Imagine if web browsers were never upgraded to use new web capabilities -- if things that were impossible in 1999 on a website were still impossible. No streaming video, no VOIP telephony, no blogging, no social media. That's what it's like with web analytics technology today.

By running web analytics conferences around the world since the beginning, Jim Sterne has a better understanding of how the industry has evolved than most.

"From my point of view the industry has always developed at a snail's pace," he says. "People are totally occupied dealing with fires and emergencies, especially in the current business climate, and opportunities get put aside till later. Web analytics is positioned as a nice, but not necessary, tool. Web analytics people would be better pointing out the website is burning money and web metrics is the fire extinguisher to put it out."

The full survey is available here.

Brandt Dainow is an independent web analyst. He works with data-driven companies to develop, monetize, and tune their web intelligence.

On Twitter? Follow iMedia Connection at @iMediaTweet.

Brandt is an independent web analyst, researcher and academic.  As a web analyst, he specialises in building bespoke (or customised) web analytic reporting systems.  This can range from building a customised report format to creating an...

View full biography


to leave comments.

Commenter: k l

2011, September 13

"Exchanging data between web analytics packages or integrating data from multiple sources is almost impossible". I totally agree. That's why over at Bime HQ, we decided to change this. Today we launched the 'QueryBlender', a new feature in Bime (a SaaS analytics tool that you can use on top of your existing data storage investments like Google Analytics, Google spreadsheets and Excel) which allows you to mix data on the fly. That means you don't need to move your data around with ETL, and you don't need to federate everything in a data warehouse. You simply query your different systems in the pivot table and join everything on the fly. I think this is going to be a huge advantage for anyone who analyzes data, as you said, the major challenge facing web analysts in their day-to-day jobs is the integration of data from differing analytics sources. Well now Bime can save you that time and effort you are spending working in Excel, cross-referencing data from multiple sources. To get a better idea of how it works, I'd invite you to watch a short screencast of the QB in action: http://vimeo.com/28939854.

Commenter: Spencer Broome

2011, September 09

Good stuff.