Extreme Implementation: Makeover Analytics Edition

At a breakout session during Adobe’s EMEA Summit last month, Adam Greco from Analytics Demystified did something really cool. Prior to the event, he got in touch with users of Adobe Analytics, around ten in total, who all give him access to their Adobe Analytics account so that he could review it and later present improvements during the summit in front of everyone.

If you are not familiar with Adam, he’s considered an industry heavyweight in analytics. He was voted the best breakout session speaker at the last EMEA Summit and has, among other thing co-founded Omniture, which was later acquired by Adobe. He’s also written hundreds of blog posts about the tool as well as a handbook on the same subject.

Anyway, at the summit, described how he had been given brief access to some of the users analytics tool before the event and that he would now offer them some feedback and wise words.

Adam selected went on to both praise and heavily criticize how these people were implementing analytics. Thankfully, he did it in a fun and informative manner.

Here are a few points on analytics implementation that Adam spoke about at the summit. All of these should help you bolster your overall analytics implementation strategy.

• Does your organization have a Solution Design Document? If not, create one! A Solution Design Document contains your most critical business questions. This is the foundation of why you are collecting data in the first place. Mapping your business requirements to which variables they correspond to, and which questions they help to answer, creates a solid foundation for your implementation. This has become even more important now due to GDPR as you need to know and explicitly tell your customer’s what you’re doing with their data.

• Is your analytics implementation easy to use for analytics users on all levels? Not all of your organisation’s analytics users have time to read blog posts about analytics. Make sure it is as easy as possible for everyone to implement your processes. Here are some top tips from Adam:

• Don’t track variables using both a s.prop and s.eVar unless it’s necessary. Having multiple variables with the same name makes it very hard for entry-level users to know which report to open. Make sure you only use s.props when needed for hit level reporting and pathing.

• Make sure all your reports are collecting data and that the data is correct. Inaccurate data or no data at all will lower the user’s trust in the data and the tool. Make sanity checks that your data is being collected as expected and make sure to remove old reports as your implementation and website/app improves.

• Expect junk data. This can be both from bots, your organisation’s performance testing tools or test orders that mistakenly was placed in your production environment. All these examples can be avoided using a virtual report suite that curates your data by excluding this type of traffic.

• Use the variable descriptions. This will make it easier for users interpreting what the data is showing. On the same note, make sure you also tag your calculated metrics and segments. It might not seem important for you, but sooner or later you will have created loads of them and you won’t be able to remember what they are all intended for. It also greatly helps others to re-use what you have created when tagged appropriately.

• Use a group login for all your official and verified reports, dashboards, segments and metrics. This makes it easy for all users in the organisation to use verified data in the way they were intended to be used.

You probably put a lot of effort into planning and configuring your analytics implementation. Make sure that these hours weren’t wasted by not maintaining the quality of your implementation or by making your implementation too difficult to use for people in your organisation. After all, if you’re like me, you want as many people as possible to make the most informed decisions as possible, using your analytics tool.