Data Strategy for Google Analytics: Tips From Yehoshua Coren

D.D.:You're helping brands scale their business and marketing with reliable campaign tracking. What are your key tactics for building a solid analytics foundation?

Y.C.: First and foremost, I strive to understand what business questions need to be answered so that I can make sure that we’re collecting data to answer those questions.

Next, I do my best to hammer home the importance of analytics being a part of the process of any “campaign” / “project” / “dev cycle” etc. It’s sometimes difficult to change a common aspect of company culture where people build new pages/components/content and only after launch do they ask for analytics. I want analytics to be a “line item” for any project or campaign. I’ve seen this happen across a number of large organizations that we’ve worked with and it is very fulfilling to see analytics no longer treated as an afterthought.

Additionally, I work with development teams to have a technical framework through which they can adhere to having a data layer and HTML5 data-* attributes within markup on the page. In this manner, Dev Teams will be able to build analytics into their infrastructure in a consistent manner.

Lastly, having a dashboard or report planned after launch will help product teams and other business owners have visibility into their efforts. They’ll be able to see the “fruits” of having analytics be a part of their day-to-day business requirements.

D.D.: What are the most challenging analytics problems that companies are facing today?

Y.C.: There are a few. Sometimes it is a lack of care and investment in what is required to maintain and use a data collection pipeline. Other times it is a naive belief that analytics is either “plug-and-play” or that you can just buy some fancy, expensive tool that has a Google Analytics connector, and then your analytics is turnkey.

Even with good processes in place at an organization, it is critical to recognize that you need people to make it happen. That’s the other big challenge; a lack of highly qualified analytics professionals to fulfill the data demands that exist in the business world.

D.D.: You’ve worked and are working extensively with Google Analytics. How can organizations leverage GA more successfully? Where should they focus their efforts?

Y.C.: To leverage GA successfully, you need to have careful strategic planning of your data collection. Using custom metrics in addition to traditional event tracking makes a big impact on how well GA functions as a tool. Honestly, though, I think that companies should focus on how to use GA to “go beyond” GA. Databases and cloud computing resources are a great place to get your data ‘out of GA’ so you can do more with it.

I think that is one of Google’s biggest moves by re-platforming to the App+Web platform. They want people to use BigQuery. I think it is a smart move by Google, even though the current iteration of App+Web is still very “beta” and the data collection limits of the platform are pretty ridiculously rigid and limited.

D.D.: How can analysts turn data into insights? Are there any low-hanging fruits worth looking into?

Y.C.: I’m not fully convinced I agree with the premise of the question. That data should be turned into insights. I.e. We have all of this data!! Tell me what to do with it. Rather, I think that the primary role of data is to answer business questions. Tell me what your questions are and the business problems we are trying to solve for, and then use data to help answer questions.

That said, some of the low-hanging fruit will always be measuring the users’ browsing experience. Doing a behavioral analysis based upon Device Category and Browser generally tends to lend insights into how a particular experience (for example, Mobile / Desktop) translates into whether or not a user was able to accomplish what they were trying to do on a website.

D.D.: Some research, as well as plenty of anecdotal evidence, suggests that data quality is often poor in Google Analytics configurations. That is a sad truth that many companies are facing. Why is it important to have accurate marketing data coming into Google Analytics?

Y.C.: Having accurate (or even more pointedly - precise) data coming into Google Analytics will lead to greater usage and adoption of the data. If stakeholders don’t trust the data, they won’t use it. Trusting data and having the adoption of using the data within an organization is critical.

D.D.: If you were to predict the biggest future analytics trend, what would that be?

Y.C.: More and more tagging is going to be “first party”, in other words creating/maintaining your own analytics pipeline using internally owned cloud infrastructure. As browsers try to lock down privacy, analytics is going to be pushed server-side. Machine learning and AI applications are going to be layered upon analytics data to try to automate the discovery of insights.

D.D.: What are the 3 analytics metrics you find most valuable & why?

1. Task completion rate.
   This is where the money is left on the table.
2. Attempts per task
   I.e. something like XX events per user. Another measure of online friction. If you see that users are clicking on an add to cart button three times more than they are adding to cart, you can identify something wrong with the add to cart flow.
3. Cost per acquisition
   If you’re spending money, you should know whether or not that spend was effective.

D.D.: What are the tough truths within digital analytics that we don't talk a lot about?

Y.C.: Knowing how to code is mission-critical (Javascript and SQL in particular). It’s not sufficient to just log into GA or build a Data Studio report.

Knowing HOW the data flows from the website into analytics (technically) is needed.

Psychology. We must always remember that analytics is not about HITS, it is about people using a website. Sometimes analysts will over-trust their data. It’s also important to trust your brain, as well as to be able to extrapolate what that data is saying about human behavior.

D.D.: If you were to predict the biggest future analytics trend, what would that be?

Y.C.: More and more tagging is going to be “first party”, in other words creating/maintaining your own analytics pipeline using internally owned cloud infrastructure. As browsers try to lock down privacy, analytics is going to be pushed server-side. Machine learning and AI applications are going to be layered upon analytics data to try to automate the discovery of insights.