Supporting Analysis — Team Structures
In the last post of this blog series on being successful with digital analytics, I shared my thoughts on how to approach conducting analysis. Having the right mindset can go a long way towards driving the overall perception of your team. In this post, I’d like to discuss the way that your team supports the organization in terms of digital analytics.
At most organizations, the digital analytics team is there to support other groups. As such, the analytics team has internal customers that it hopes to assist. Like any group serving customers, there are different approaches to providing service. These approaches will vary based upon factors such as business model, size of company, location of employees and the sophistication level of the team and its internal customers. The support model utilized must match the organization — a model that works for one organization won’t necessarily work for another.
In my career, I have seen many different approaches to supporting digital analytics. At some large organizations, I have seen a centralized model. In this model, there is a core digital analytics team that performs most of the analysis and internal customers who request assistance as needed, often called an analytics center of excellence (COE). Many times, these requests utilize a ticketing system submitted by internal customers and fulfilled by the core analytics team. The advantage of this model is that the core team is intimately familiar with the implementation and past analyses performed. They are able to answer incoming analytics requests more swiftly than novice internal customers and can avoid reinventing the wheel if they know that a request has been asked and answered before. The downside of this model is that internal customers often have to get in line behind other requests and are not always empowered to pull the data they need themselves.
In a decentralized model, there is a smaller core analytics team and most internal customers are trained in the analytics tools and expected to self-serve when it comes to digital analytics data. This approach puts the data directly in the hands of the internal customers so they can pull the data they need whenever they need it. However, training non-analytics team members on how to use your implementation and your analytics tool can be quite difficult. This is especially true for internal customers who only need data occasionally and end up being casual users of the digital analytics tool. I have seen many situations where business users have pulled the wrong data and made misinformed decisions as a result. This can be exacerbated if you have many office locations and people of varying technology skill levels. This model is often extolled as executives proclaim they are “democratizing data,” but I see many situations in which the organization is fooling itself about how data is really being used.
The other approach I have seen is the hub and spoke, which is essentially a hybrid of the preceding two models. This was the model I chose to use when I managed digital analytics at Salesforce. In this model, you have a small core team (hub) that focuses on some of the most critical business questions and educating a small group of people throughout the business (spokes). Often times the spokes are key people within strategic business units who have an affinity for digital analytics data and are more teachable than the masses. The core team educates these spokes (I called them Ambassadors at Salesforce) and relies on them to be the main educator of their team and their first point of contact. If the spoke runs into advanced questions or issues, they would reach out to the hub for assistance. That way, the hub has a smaller number of people to support and can delegate some of the responsibility to the spokes.
Which model you choose really depends upon your organization and its culture. It may take you a few times to figure out which model is best for you, but it is worth taking the time to consider the options.
Your homework for this post is to:
- Determine what your current support model is today and how well it is working.
- Consider meeting with your internal customers and bosses and discussing the different models and the pros and cons for your organization.
- Discussing whether you’d like to try adjusting the current support model and, if so, what resources that would require.
In the next post, I am going to discuss training your internal customers on digital analytics for those who decide to go with a decentralized or hybrid model.