Tips to Keep from Drowning in a Sea of Analytics
For the first Idea Lab (breakout session) for ASAE Day 2, I opted for the session on analytics with Marissa Goldsmith from Beaconfire. She broke the session down this wayâŚ.
Tips on how to think about analytics:
â˘You have to start by defining success, and avoid using analytics to do this. State your goals and then use analytics to see if you met them.
â˘Focus on trends less than the specific numbers. Trends help you add context to your numbers. Analytics are detailed, but there are limits.
â˘Be careful about comparing data from different sources. WebTrends and GA may get different numbers for the same stat and thatâs okayâŚthey measure stats differently.
â˘Only look at actionable analyticsâŚstats should tell you if your doing a good job or if you need to do something better.
Tips on how to present data:
â˘Be wary of overarching organization-wide dashboard reports. They are dull because they just present numbers, not stories. And as a rule, if itâs dull to create, it will be dull to read. They are often difficult to understand. And not everyone cares about everything. Consider a series of smaller dashboards instead.
â˘Never leave anything unexplained and never just say something is bad. Dig further into the analytics to find out why, then propose a solution.
â˘Add fun graphics when you can.
â˘always double check your numbers before you present them, and be aware that some issues are implementation issues.
â˘You do have to plan, itâs not just cutting and pasting code.
â˘You can use WebTrends and GA simultaneously, just donât mix and match numbers. Supplemental programs worth considering include: CrazyEgg, DoubleClick, Flurry, AdThis/ShareThis.
â˘Donât rely on your analytics tool alone. Take stock of your tools: CMS, eCRM, Blog, Community Tool, Online Membership Tools. Present reports based on a variety of sources.
â˘GA will not track PDFs, external links, or other specialized tracking (eg., between CMS and CRM pages). So you need to write scripts (or âborrowâ someone elseâs) to compensate for these. (consider using Brian Cliftonâs external tracking script for GA.
â˘use GA filters: eg. To filter out your officeâs traffic so you donât skew your numbers. But be sure to create one profile that is filter free (a ârawâ profile), just in case something goes wrong with your filter
â˘test your analytics. Use GAIgnoreMyHits to isolate your particular visit. Test each situation and in some cases waitâŚand test again to see if the results are the same.
Recommended Analytics Reading:
â˘NTEN Article on Dashboards
â˘Measuring Success blog (by Brian Clifton)