Defining the Five Aspects of Web Analytics March 12, 2008
Posted by debbiepascoe in web analytics, web standards.Tags: compliance, data integration, data storage, implementation, web analytics
add a comment
The analysis of site traffic has become quite complicated. Even Eric Peterson has come right out and said “web analytics is hard“. On that we agree.
A really interesting conversation, begun by Ian Thomas from Microsoft, predicts that in five years, there will be no Web Analytics vendors, but Web Analytics will be everywhere.
Regardless of who’s driving the activities five years from now, the same five basic things will have to occur:
- Initial implementation
- Ongoing maintenance
- Analysis and interpretation
- Data storage
- Data integration
Within this framework are the data quality, site quality, and compliance issues and the need for a scientific approach to ensuring the data quality, both discussed in previous posts.
Aspect One: Initial implementation
Even if the initial implementation is flawless, the challenge then is to keep it that way. Implementation:
- Requires a commitment on the part of the site owner – page tagging is a contact sport
- Requires guidance by the vendor or by third-party consultants that know the selected vendor’s product
Aspect Two: Ongoing maintenance
Site owners struggle mightily with this. Sites are becoming larger, more complex, more volatile, with more user contributors. Errors creep into the tag implementation, site quality issues infect the data, and compliance issues impact the way visitors interact with the site. Ongoing maintenance:
- Is the ultimate responsibility of the site owner
- Requires support by vendor or third-party consultants that know the vendor’s product
- Requires special product knowledge and expertise to diagnose and correct issues
- Requires continuous automated validation of implementation and diagnosis of issues
Aspect Three: Analysis and Interpretation
The analysis and interpretation of web analytics data is the source of much consternation, worry, hair-pulling, obsessing, and frustration. If you don’t believe it, drop in on the web analytics forum, and see the types of things practitioners struggle with every day. A lot is being written and pontificated now about the future of web analytics. No matter what happens, without analysis and interpretation, all the rest is simply an expensive, time-consuming and resource intensive data collection exercise. And without accurate data to analyze and interpret, practitioners are challenged even further. As the discipline matures, data validation processes are put in place, and companies’ internal organizations mature, those people that are skilled at interpretation will be an even hotter commodity than they are today. Interpretation:
- Requires an understanding of the business drivers of the site owners – analysis must align to goals
- Requires special knowledge and expertise, and has given rise to a new profession choice
- Currently part art and part science, it carries with it the challenge of moving internal stakeholders past “page views” as the measure of success
- Needs standards – standards development is still in early stages – Web Analytics Association has a committee devoted to it
- Is dependent on data quality. Defects in data quality negatively data impact interpretation
Aspect Four: Data storage
Somebody has to do it.
- Off-site, on vendor-supplied machines – it is part of their cost basis. Being in the web analytics business means being in the infrastructure management business, lots of machines, lots of bandwidth, highly secure environment, guaranteed uptimes, power backups, etc
OR - On-site on site owner-supplied or controlled machines
Aspect Five: Data integration
Web Analytics vendors are now acquiring other companies and building out partner networks that enable them to combine data from other sources. This is a natural next step, to get a combined view of the things that drive people to sites and the actions they take once they arrive. Sources of data for integration include:
- Web Analytics
- Live Onsite Chat Session
- Onsite surveys
- Ad banner-generated traffic
- Keyword buy-generated traffic
- Email campaign-generated traffic
- Call centers
- POS devices
Conclusion
The adoption of outsourced web analytics is not a decision to be made lightly. It involves a lot of moving parts in a shifting landscape. In some ways, web analytics as matured – “it’s in the august of its years” says Ian. As for data integration, it is still early days. That said, things move really fast in this space, and there is tremendous interest in and motivation to solve it. These five aspects will be as much a part of the picture in five years as they are today.
Web Analytics 2.0 – How to Get over Information Intoxication November 18, 2007
Posted by debbiepascoe in accessibility, multiplicity, privacy, quality, search engine optimization, usability, web analytics.Tags: accessibility, compliance, data quality, Omniture, privacy, site quality, Unica, web analytics, WebTrends
add a comment
Information intoxication – it’s that state we find ourselves in today where we have miles and miles of data about our site visitors. We’re swimming in an ocean of data, drinking it in as fast as we can. But the real question is:
Is the data any good?
Site owners are beginning to realize that the quality of their data is directly linked to the quality of their web analytics implementation, and is directly impacted by site quality factors such as broken links, search optimization problems and duplicated information, and compounded by compliance issues like accessibility and privacy-related concerns.
These three things – data quality, site quality and compliance – are the inextricably linked gears driving your online engine. Organizations are beginning to understand that changing vendors is not the answer. Whether you use Omniture/Visual Sciences, WebTrends, Unica, Coremetrics or any of the other vendors, the underlying issues remain the same. Only by addressing them will you be able to have confidence in your data.
Data Quality
Organizations that take the step from log file analysis to page tagging make a commitment, not only in dollars, but in human resources as well. While your vendor of choice will do everything they can to ensure a successful implementation, the responsibility ultimately falls on the site owner to know if the pages are tagged, and if the beacons are functioning. And your site is constantly changing. Untagged pages and malfunctioning beacons result in systemic errors in the data, meaning you are missing traffic information about those pages 100% of the time. Your assumptions about the resulting traffic patterns will be incorrect because your data is incorrect.
Site Quality
Even if your web analytics implementation is perfect, you can have site quality issues that show up – or don’t as the case may be – in your traffic data. A broken link is a path that can’t be taken. Pages missing from your internal search are destinations that can’t be arrived at. Inadequate search optimization for search engines like Google result in less traffic. Is there anybody that would say, “I’m happy with the amount of traffic I’m getting and don’t want any more, thanks”? Of course not – sounds ridiculous. But sites missing opportunities to optimize for natural search placement are settling for less.
Compliance
Site owners can’t know the impact of poor accessibility or the insecure collection of personally identifiable information (PII) by looking at traffic patterns. When these issues result in visitor departure, the result in the data is absence – “raise your hand if you’re not here”. The assumptions made by reviewing traffic patterns will be flawed, and the actions taken as a result will be misinformed.
Compliance issues impact the visitors use of the site, as well as putting the company at risk. Accessibility and privacy continue to attract attention as the web continues to mirror the bricks and mortar world, where you can shop, bank, research, catch up on the news, secure government services, arrange travel, communicate and socialize with others. Many governments have already implemented legislation, and the focus on these issues will continue to grow. People following the Target lawsuit will have already noted that the National Federation of the Blind is not looking to federal legislation; rather they are pursuing legal actions in state courts, where accessibility has been addressed.
Bottom Line
The good news is that this is a manageable environment. Data quality, site quality and compliance are three pieces of the same puzzle, and must be treated that way, with an automated solution that frees the web analysts to have confidence in the data and the assumptions, conclusions and recommendations they make as a result.
If you haven’t seen it, take the time to read Avinash Kaushik’s blog post titled “Multiplicity – Succeed Awesomely at Web Analytics 2.0″.
maxamineWeb Analytics Myth Number 1: There aren't enough web analysts October 19, 2007
Posted by debbiepascoe in quality, web analytics, web design, web standards.Tags: accuracy, compliance, eMetrics, myths, quality, web analytics
add a comment
I just returned from eMetrics DC, where I talked about the five myths – we call them “myth-understandings” that are prevalent in the market today. This message resonated and, in fact, it was easy to see the light bulbs going on as we talked through what is myth and what is reality. It is fresh on my mind, already. Then this morning I received my email update from the web analytics forum, and one conversation thread was screaming “myth number 1 is alive and well and living in the forum!” Since I haven’t written in the blog before about the myths, I’ll list them out, then I want to talk specifically about number 1.
Here’s the countdown:
5. Accuracy doesn’t matter – random errors can and do occur, and as we saw in my previous post (Web Analytics Shootout), different vendors report different numbers even in a quality-controlled environment. That is not the same as the systemic errors that appear in data due to untagged pages, malfunctioning javascript, and site structural problems that in turn skew the data.
4. My checker checks out – Conventional scanners get 40-60 percent of site. They can only audit what they can access, which results in an inadequate and arbitrary subset.
3. One, done and run – Quality and compliance management is continuous process, not a one-time exercise. Because sites are becoming larger, more complex and more volatile, the need to continuously monitor quality and compliance is even more crucial.
2. Optimizing the suboptimal – significant time, effort and money being spent to optimize online performance. However, common approaches are violating a basic axiom of optimization theory: you cannot optimize in a suboptimal environment. Concurrent enhancement of the environment to improve data quality and correct site problems is essential to achieve positive and sustainable outcomes. To put it plainly, if the floor in my house is unlevel and the doors won’t close, the roof leaks and air blows in around the windows, it might not be the right time to fuss over just the right carpet.
And the number one myth…..there aren’t enough web analysts.
This is still an emerging market, and as with all emerging markets, it is suffering chronic labor shortages – this is nothing new or even unexpected. The real problem is not just a lack, but rather a WASTE of analysts. Too much of their time is spent manually checking, finding, and fixing problems that shouldn’t be there to begin with.
The conversation thread that got me going this morning started with a question about a Google Analytics implementation in a php environment, and why traffic is not being recorded – the code seems to be there, is there something special about php, well maybe you need to write some special script, etc., etc, – aaarrrgghhhhh! The time termites are eating away at this person’s precious limited resource, along with all the others who are trying to help him solve it. In so many organizations, site quality is an afterthought, and the web analysts are not equipped with the proper tools to do their job. They spend time on time-wasting exercises like this, when they could be focusing on how to drive value from their data.
Bottom Line:
There could be 10 times the number of analysts that we have today – wouldn’t matter. Without addressing this fundamental, foundational issue, there still won’t be enough. Fortunately, as we get out and debunk these myths, we are seeing the lights come on, the cobwebs disappearing, and people increasingly becoming clear-eyed about what they can do going forward to get control of their web environments.
