Web Analytics and Measuring Success

How do you know if you've been successful with search engines? You can check your "rank" at search engines for particular keywords, analyze log files to see the actual terms people used to reach your web site or make the ultimate jump and "close the loop" by measuring sales conversions and return-on-investment (ROI). This panel explores ways to measure success and what statistics you should really care about. Allan Dick of Vintage Tub & Bath is moderating along with Matthew Bailey of Site Logic Marketing, Lionel Largaespada of Fathom Online and Laura Thieme of Bizresearch presenting.

Matt starts things off. People typically pull up their analytics, look at their dashboards and freak out. Those that are on top are the ones who define their goals from the start. How can you measure without a goal to measure it by? Seems simple but many miss that important point. Web sites without analytics are not worth running. You are losing money... you just don't know where. Pages visits, top ten pages, unique visitors, etc. are from the caveman days. They don't really define what is going on.

SEO - if they can't find you, you aren't there. Usability - if they can't find it, it snot there. Analytics - they tell you what just happened. Analytics are not necessarily about the numbers but the process. What hinders people from getting to point A to point B? This starts with segmentation. Numbers don't show you the whole picture. Without segmentation, you will not know the whys. It tells you the factors that lead to the actual numbers which will then help you to adjust your strategy.

If you use a single conversion rate for example, a e-commerce site, you cannot tell which products are doing well and which are not. You need to have conversion rates for each product and even each section of page of the site.

Three C's of analytics -

- Context
- Comparison
- Contrast

Find what works. Look at key performance indicators, then segment (where did they come from). This will tell you what they are looking for. One thing Matt points out is that product pages or specific interior pages typically convert better than having someone enter at the home page. As far as engagement, visitors from blogs and articles are engaged longer than for example social news sites such as Digg. Reason for this is that a blog for example if like a word of mouth referral. Social news really draws people who are just looking for the latest news and trying to keep up on things. So the context of the link is important. In other words, go after links that will not only give god ranking juice but which exist in a context that will help visitor engagement.

Next up is Lionel but not before Alan mentions that his company got stuck in the "caveman" phase that Matt was talking about. He discovered a lot of metrics they were looking at did not allow them to focus on what really needed to be done. When they re-defined their metrics, they were able to make better judgment calls.

Lionel polls the audience as to how many are actually using an analytics package and secondly how many can make heads or tails out of the information. The response for the second question was less than the first. Lionel's definition of web analytics is the objective tracking, collection, measurement, reporting and analysis of quantative internet data to optimize sites and marketing initiatives.

82 percent of people reading analytics think it is confusing, mostly because there is too much information to collect. As far as where people are coming from when they land on our sites, there are so many inlets. Trying to measure the effectiveness of each inlet is a daunting task. Therefore we need to consider technology that goes beyond typical analytics packages. Good "marketing analytics gathers data across all channels - not just online. This can include mobile, print, call center sales, sale people, brick and mortar, etc

A question to ask s what data do we really need to make quality decisions. A second question is how do we even get this information. Third party tracking is key to integrating your marketing programs for analysis. A goal with analytics should be to connect disparate conversions into a single analysis.

Getting started entails setting up campaign objectives, what data is available and what is not, don't limit efforts based on data, plan to work regularly with data, and re-evaluate data options. He repeats what Matt said in that all data needs to line up with the actual goals the company has set out for themselves.

Towards the end of his presentation, Lionel shows an example of how they used cross channels to analyze a client campaign and make marketing decisions based on the data they received.

Finally Laura takes the stage. She asks how many are blogging and actually tracking that activity. The landscape is changing in what you need to track. Are you also analyzing campaign activity at the keyword level. The initial information marketers see when logging into PPC management consoles is their dashboard - CTR, CPC, impressions, ROI, etc. However are you using tools available to look at advanced KPIs? Companies need to ask themselves the question of what is their acceptable customer acquisition cost. They also need to make sure that the data they are looking at is accurate. Is everything you use to track installed and working correctly?

Track everything -

- visibility
- spider activity
- traffic
- sales
- latency
- KPIs
- ROAS
- competitors
- social media

She then highlights some of the tools that are avaiable - Omniture, CoreMetrics, Web Trends are a few that were listed.

Favorite reports -

- NetTracker - Robot/Spider report
- Google Analytics - Network properties/Network location
- Google Analytics - Goal conversions

Don't discount the traffic that blogs can bring but do look at it, analyze it and make marketing decisions based on it.

Analyze bounce rates by keyword in order to lower them. Of course there are going to be obvious keywords that will bounce users but you want to use analytics to identify bounce rates for keywords that should create more stickiness.

With web analytics, being analytical is not enough but should be combined with creativity. You have to spend time with clients to discuss what web analytics reveals to you.


David Wallace - CEO and Founder SearchRank