Web analytics play a crucial role in building a high performance web business. Whether it’s basic site statistics like unique visitors, page views or sessions, analytical data like click paths or funnel abandonment or campaign performance metrics like cost per lead/sale, conversion or value per visitor, knowing your numbers help define your next steps for growth and improvement.
However, besides the greatest hurdle of web businesses actually taking the time to become aware of their numbers, there is also the concern with the numbers being accurate.
I read an excerpt from the July/August 2007 issue of Revenue Magazine in the Tools section that stated,
A recent comScore study suggests that most sites that count unique visitors based on first-party cookies may be over counting their visitors by as much as 150 percent.” The statement continues, “The reports says that 31 percent of U.S. Web users delete their first-party cookies during a given month that means sites are then recounting return visitors as new by a factor of 2.5.”
The “cookie-deletion” issue has been a topic of discussion for quite some time. But the importance I want to pull out from the quote is how an inaccurate unique visitors count can greatly throw off other metrics including conversion rates. And maybe for the better if you thought you generated 100 sales from 10,000 visitors when in reality it was only 7,500 visitors! On the other end, overstating visitor counts may cause you to re-think your SEO relationship if 150% less visitors are actually arriving on your website.
One thing seems to be certain – it’s better to focus on trends than absolutes. If you are using inaccurate numbers due to cookie issues but those same numbers increase over time, then the trend shows positive growth (in the case of unique visitors.) As web analytic providers research new ways around inaccuracies you focus should be on watching the trends in context of your goals. Downward trends for bounce rates, abandons and expenses; upwards trends for unique visitors, conversion rates, sales.
Comments
