October 8, 2013 | Written by: Paul Nangle
Categorized: Customer Analytics | Marketing
Share this post:
There is a lot of discussion in the industry about the value of a good digital analytics analyst. Data in and of itself doesn’t really have any value. It is only when you use data to drive action that it becomes valuable. There are a lot of “analysts” who are really just report builders or analytics data enablers. While this is helpful it still doesn’t really unlock the value in the data. A good analyst is one who helps understand what actions the data is telling us to take. But after working with many of our largest customers with IBM Enterprise Marketing Management, I have learned that how analytics programs are structured can be the foundation for creating good analysts and as such is crucial to a company’s Digital Analytics Success. Below I will discuss at a high level the different ways enterprise analytics programs can be structured and the advantages and problems each model presents.
Staffing Organizational Structures
At the highest level there are two basic models for organizing your Digital Analytics staff: a Centralized Model and a Decentralized Model.
A Centralized Analytics program is one where there is a team formed specifically of digital analytics practitioners. In this model other teams who need analytics insight work with this team of digital analytics experts. This creates a Hub and Spoke model placing the analytics team at the center of the digital analytics world. The other verticals go to them with their questions and to get reporting and analysis.
A decentralized model is where each of the business verticals has some one or a few people that do digital analytics reporting and analysis usually along with their other responsibilities. In other words each vertical is responsible for their own digital analytics reporting and analysis.
Decentralized – the Default Structure
A Decentralized Model is what most companies default to because as the Digital Analytics tool is implemented management expects each team to automatically begin using it as part of their function. As the data it reports on is applicable to their vertical there is an assumption that it will be adopted by the team in conjunction with their other duties. It is or isn’t adopted by different team members based on their desire or affinity for doing analysis. So each team adopts it to varying degrees and for those who take it on they may end up doing so in what is otherwise an analytics vacuum in their particular business vertical.
Problems With The Decentralized Model
The problems with this model are as you would expect as you think about how this would continue to unfold. With analytics being a second or third job for team members, in this analytics vacuum, they have little time or resources to gain any real expertise. But the siloed nature of the decentralized model also causes problems for management. There ends up being few standards and little consistency across the different verticals in reporting or analysis. And any reporting will, intentionally or not, tend to be subjective and self serving. In the end the biggest problem with this model isn’t the model itself but the reason it is a default. The assumption that it will automatically be adopted and the little to no executive sponsorship that accompanies this assumption means there will be little to no adoption for many of the verticals.
Centralized – The Companies Who Go For It
Banking On Analytics And Funding A Centralized Structure
Companies whose top management are committed to being “data driven” will sometimes decide on developing an analytics department around the implementation of their Digital Analytics program. This center of analytics practitioners is the basis of the centralized model. Since this team is staffed and funded it does what decentralized models often fail at. It develops expertise, creates standards and encourages consistency. It also drives adoption. It has to. That is its mandate. Obviously an analytics team has to adopt analytics.
Executive level mandated adoption is why almost all centralized digital analytics models are more successful. It isn’t necessarily the model itself but the sponsorship, the buy in that executives made in the value of analytics, that make this approach appear more successful. The success comes from the fact that it is driven, expected and funded from the very top down.
Why The Success Might Be An Illusion
The centralized model itself has some inherent problems. In this model each member of the different verticals that need actionable data have to communicate what they need to the analytics team. This adds an extra step between people who need the actionable data and the ones who can make data actionable. That step, that required communication, slows down the process and opens the door for misunderstanding. This separation also stifles innovation. Once again the apparent success of this model doesn’t come from the models inherent superiority but from the executive commitment that allows this model to be implemented.
A Third Path
Over the years I have learned a couple lessons from the successes and failures I have witnessed with both models. The first lesson is that of executive sponsorship. Notice I specifically haven’t been saying executive Ownership. Executives do not need to become Digital Analytics experts to create and sustain a data driven structure. But they do need to require its organization and mandate its continuation. The second lesson is that both models have their strengths and weaknesses and that we can capitalize on that.
What About A Hybrid Structure?
Can we use the best of both models? Yes we can! It takes work and executive commitment and sponsorship but with the right organization you can create a hybrid model. The trick is to have members of the individual verticals that are responsible for analytics in their group also align cross vertically with the members in the other verticals who work with analytics.
There needs to be a fully functional cross vertical team of digital analytics practitioners. They need to answer up through their executive sponsor (who once again doesn’t need to be an analytics expert) and have the expectation to meet regularly as a unified digital analytics team.
The Advantages Of A Hybrid Model
There are a few advantages to the Hybrid Model. One is that having the folks in each team do their own analysis is far more efficient. Less gets lost in the translation between an analyst and other team members when the team member does their own analysis. And this efficiency can allow for more innovative analysis. Another advantage is that since they need to meet and discuss their analysis across the verticals their reporting will be more objective and less self serving. Finally as a team they can better prioritize things like new tagging requirements they need the IT team to implement.
The Difficulty Organizing Across Verticals
The executive sponsor can facilitate and mandate the organization of the team. They can help organize meetings and at a high level define some rolls. But unless they are an analytics expert they will need a team lead who can be the focal point. The team lead can be from one of the verticals but will need to be objective. They will need to have a deep knowledge of the analytics tool and will need to be experienced in making analytics data actionable. They will also need enough free time to take on the responsibility of keeping the team going. This may not be a full time position depending on the size of the cross vertical team and the frequency with which it needs to meet.
The Problem Maintaining A Cross Vertical Organization
The problem with maintaining the team is that there needs to be a lead who has as much at stake for keeping the team running and progressing as they have at stake within their vertical. And when I say maintaining the team I don’t just mean keeping the meetings going each week, that can be required by the executive sponsor, I mean continuing to drive deeper use of the data and growing team members expertise.
Organizing Around The Customer
Some of the more successful Hybrid model clients organize the Hybrid team around the thing that they all have in common, their customers. The teams are often call things like “Customer Insight Teams”. So instead of just looking at the data from the perspective of their vertical: Marketing, Merchandising, Content, Usability, etc., they also look at the data as building a holistic view of the customer. Digital analytics data is really customer behavioral data and understanding the customer better is usually the point of implementing digital analytics in the first place. The leader of the team can also use this customer focus to drive deeper dives into data and challenge the team member to deepen their analytics expertise.
Team Lead Options
Many companies don’t have an objective member of one of the verticals who has a deep enough knowledge of the analytics tool, enough experienced in making analytics data actionable and enough free time to lead the team. While hiring someone of this caliber is an option we know that a great many of the “analysts” available are rarely more than people adept at building reports. Though report builder level analysts are usually more affordable they don’t have the skill set needed to lead a cross vertical team. A highly skilled and experienced analytics practitioner available on a part time basis may be the best option for this position.
Now it is time to step back and think about how and why your analytics practitioners are organized the way they are, whether you are getting all you want from your analytics tool and how you’d like to move forward. Feel free to contact me if you would like to discuss this further.