Alex Holman of Mezzo Labs tells us how to spring-clean your web analytics account and keep it looking fresh for your stakeholders.
I thought I’d share my top tips for cleaning up a web analytics account, whether it’s a yearly audit or inheriting someone else’s broken dreams.
Just to give you a bit of background, I studied microeconomics at university and the one non-drink-related lesson I learnt was “too much choice is bad”. As supermarkets Lidl and Aldi have taught Tesco, the consumer doesn’t necessarily need infinite choices if your one product is great value.
The truth is, when you give most people a choice of greater than six, it becomes just too hard to make a decision. That decision time costs you money, the anxiety about the choice, is a further cost.
The same is true of data. You need to get that balancing act right. Give people just enough data to do their jobs, in a quick, efficient way (no matter how inefficient the tool for the job is). Too much data, and they quickly drown.
Here are my top tips for keeping your web analytics house clean…
1. You’re the Admin, you’re the boss
The more people you grant admin rights to, the more likely you’ll turn your house into a Greek tragedy. Think arguments, manipulation and duplication. Limit report creation rights.
2. Standardise your metrics
Depending on your organisation this could be all or just one. This way you’re never jumping in and out of reports looking for the one with the right metric. If you can agree with everyone the only metric you need is Visits and only ever use it in all reports and for analysis – congratulations, you are awesome, please tell me how you did it!
3. Name your reports consistently
Agree the naming structure for all reports. Agree that if anyone hasn’t followed the structure, you’re allowed to rename it without them throwing a tantrum. How I label…
Structure: Project – Filters – Dimension1/Dimension2/Dimension3
Example: Account Logins – Men’s Fashion – Account number/Account action
No one can see the filters you add, so please put them in the name or you will forget.
4. Mimic GA
If your platform is Webtrends or SiteCat, but your end-users only know Google Analytics, mimic the GA structure in your menus and report categories. It takes a bit of effort but you’re removing the most common objection to your platform: “I can never find my reports!” You may not be able to make marketers love your customisable non-GA tool but you can make them hate it less.
5. Avoid duplication
So one person wants a report on device split by browsers and the other browsers split by device. Don’t write two reports. Rock/paper/scissors. The loser has to take an extract and manipulate it in Excel. Or set them up with an API extract which does this for them.
6. Keep a log
Keep a central record of everything created, where they are to be found, when you created them and for who. When that person leaves the company, ask yourself is this report useful to anyone? Or was it made because that individual twisted your arm to do it? If so bin it. This includes external users and test reports.
7. Ask for help
If you ask them, Webtrends, Adobe and other platform providers can give you awesome extra data to help you clean the crap. For example, how many people have viewed each report? Which users have viewed each report? What users have not logged-in in the past 6 months? Sorry, if you’re on standard Google Analytics, no such luck!