Having worked in online marketing and web analytics for nearly a decade, I’ve heard it all when it comes to myths passed around small and large companies alike. Here is a top 10 list of my favorite web analytics myths and practical advice on how to dispel them.
There are several reasons why free software is never the best solution. Some of my favorite retorts to “why do we use Omniture rather than Google Analytics” often involve witty comebacks like “because I have to pay the bills” or “because my boss said so.” If that doesn’t work (and it never does), the primary reasons to go with enterprise analytics are:
- Service Level Agreements: What happens if your software fails? If you pay for analytics, you have a neck to choke; if not, you have to wait it out and pray nothing is lost.
- Data ownership: Free doesn’t mean consequence-free. Someone is paying the bill. Free software is often offered “at no cost to you.” Enterprise solutions enable you to take your data with you, should you so desire.
- Privacy: Enterprise solutions offer security and privacy through non-disclosure agreements protecting both sides of the contract.
- Customization: Hacking free solutions like Google Analytics is possible, but only to a certain degree. Enterprise solutions are built for customization with business objectives in mind.
2. Bounce Rate (or “Insert Metric Here”) is the Best Metric
Avinash Kaushik calls it the sexiest metric, but it’s not the best because there is no “best” metric. I know of several companies that employ teams of analysts whose sole responsibility it is to monitor a “God metric,” but rarely do these stand the test of time. It’s best to focus on a handful of metrics that actually drive profitable insights.
3. Everything Avinash Kaushik, Jim Sterne, or Eric Peterson Says is Gold
Don’t get me wrong, Avinash is brilliant, but none of the experts in analytics know your business well enough to provide a plug-and-play measurement strategy. On a high level, their best practices are indeed gold, but nothing beats digging into your data and creating an analytics playbook of your own.
4. Dashboards or Reports Should Have 4 Quadrants and Only a Handful of Data
Although it’s a lofty goal to aim for when producing any content (resumes, menus, etc.), it’s extremely difficult to integrate the data, insights and visuals on a single page that caters to everyone on a distribution list. A good strategy is to start bigger than necessary to showcase your capabilities, get the attention of several stakeholders in your organization, consult with unique business units, and fine tune custom reports for each audience.
5. Insights are More Important Than Data
Sometimes key data is all your executives need to make a decision. Should your company officially support IE6 for our next redesign? If only 2 percent of visits to your site for the last six months came from IE6 and incorporating development and testing for an application would cost several million dollars, the answer is easy!
6. Unique Visitors are Real People
Unique visitors is perhaps the single most abused metric in history. If you really think about it, the metric known as unique visitors is no more than: count of persistent cookies dropped in a browser. Unique visitors do not equal browsers, individual people, or computers.
7. Analytics Code Degrades Site Performance
8. Web Analytics is the Responsibility of Marketing/Research/Communications/Operations/IT/etc.
Web analytics is the responsibility of a data-driven organization. If your website influences your business in any way, it’s everyone’s responsibility within your organization to take a portion of the responsibility for coming up with actionable business insights that increases revenue, decreases cost, takes advantage of opportunity, or mitigates risk.
9. Metrics From Different Web Analytics Vendors, Web Logs, and Databases Should Match
Web analytics is inherently inaccurate and practitioners are rarely adequately versed in statistical theory, so to argue that any one data collection source should match another is futile. There are several factors that contribute to inaccuracies in web analytics data including:
- Cookie acceptance.
- Data corruption: receiving, executing, and transmitting.
- Server-side caching, scripting or configuration issues.
- Filters and processing rules: reverse DNS inaccuracies, data sampling, data encoding.
Look past the numbers and analyze trends, ensure your findings are statistically significant before coming to a conclusion, and always be transparent about web analytics limitations.
10. Insights From Web Analytics is Free
There are plenty of other gems out there, do you have a favorite myth that I missed?