This is the first post of a collection of them which are created to talk about how to create a data analytics department. The main reason to write them is to materialize my thoughts and experiences around working in data analytics for corporate environment. Those thoughts are opinions being created in a very specific context of experiences that were composed by working in a department with highly focus on financial data, but having some degree of responsibilities around product, marketing, customer support and sales analytics.

In this collection of articles I will try to place the ideas about technologies to be considered, guidelines on how to balance tradeoffs and better strategic paths to take in order to grow quickly to a steady stable state and deliver fast and good along the way.

The post does not have vocation to become a bible for building data analytics department, but to give some guidance that can be helpful. If you landed here because it is your first experience building teams, I hope that it can help you gather a wide landscape view of your problems to come. If you have already experience but you want to explore new opinions about the topic, I hope you can find here an innovative point of view about these problems.

Which are the high-level challenges you will have to face?

The first weeks themselves will be a whole challenge. A lot of new people and new information are coming after you and filling your time.

You will have to carry out your first steps wisely and be quick identifying black holes of time and energy that you will have to avoid, and also shortcut paths that can be good to walk through.

Understanding the business model of the company, learning about key processes around IT or HR, meeting the contact people that are key touching points and hearing about what it is expected from you are the first things to come.

But no matter how chaotic or overwhelming these first steps can be, you will have keep your mental anchor and always have in mind the high-level points that you will have to address.

• Manage context: you will be sitting in a company surrounded by co-workers. You will have to work with them. For that probably you will have to define and establish dependencies with them. Also a framework of collaboration and interaction between your team and them in order to be as much productive as possible.
• Business understanding: by understanding the business model it can help to trace the best way to support the specialized domain business units with data more efficiently. Also it can help you to define an ideal way to represent the company data in one unified data model. That can be useful if the end-goal of the data team can be unified analytics and provide a single source of truth.
• Tech strategy: decide which is the best tech stack to ease your problems. A good decision about technology can solve a lot of problems and avoid to reinvent the wheel. That can allow the team to focus and spent their energy creating value.
• Data needs: we need to understand which are the needs in terms of data and mainly which are the most pressing needs to be solved first.
• Organization management: how to the exposed problems translate into roles to hire and organization structure.

Manage context

You are in a company and you have to understand that company is paying to obtain an outcome. The outcome sometimes is difficult to measure and for the others will be only measured in a very subjective way.

Understanding your labor context is always the key to success, specially in data. And among all your surroundings, the most important context is the people.

You will need to work together with many other people and understanding them culturally and personally is a must. Many capable people demonstrate themselves failing in some people surroundings. They are un-capable to flexible adapt to the labor context and shine there.

The first point you have to understand is that your future will be tied to them and so for your future successes. Building trust and on-boarding them in your challenges is as important as understanding their problems. I saw so many times technical service providers to fail in here just because they are driven by their own ego. They are so attached to their own ego that they can not establish a constructive relationship with other co-workers that probably are also driven by ego. At the end all of them, and the company, lose.

Once you build a relationship, the efforts must not end up there. Having a trust sometimes does not mean having their respect. Having a respect means establishing equal and balanced relationships. You need to make the outcome of the relationship be credited equally. This point can seem that is driven by ego, or by a need of recognition that of course it should not be the main driver of a data lead (as individual). But remember, you represent a team and the people on this team needs to have their work recognized and respected outside the boundaries of the team. If you want to make the team members happy and motivated to grow, don’t forget on building respect for your team and team work. And you have to keep it in mind while you draw your plannings.

Business understanding

Another of the key context environment points is the business. You need to understand what is making your company earn money, how are the processes done and how this processes translate into data.

By understanding this

by understanding the business model it can help to trace the best way to support the specialized domain business units with data more efficiently. Also it can help you to define an ideal way to represent the company data in one unified data model. That can be useful if the end-goal of the data team can be unified analytics and provide a single source of truth.

What is your purpose of existence?

Along solving all these questions exposed previously, we have this transcendental question. A key question to be ask to yourself. Why your position exists? How was it decided to spent the budget they commit in your role? What is it expected from you and your team?

Without not being able to answer these questions you will lost in your new journey. It is not a just simple silent exercise of observing, guessing and inferring. It is highly recommended that you ask direct questions to your colleges in order to understand the history of the decision. How the decision was backed and taken?

By failing in doing it can put you in a painful trip in direction of the exit door. Now, in the first weeks, it is the moment to re-assess expectations and also search for an alignment with the rest of the organization. Probably they have unrealistic ideas of what is expected from you to deliver. Or probably you realize which are the main problems and pain points that you can alleviate.

The first weeks in the company are critical and will condition the rest of your performance. You need to understand your role, the pain points of the stakeholders, build trust and carry out some quick delivers. But all that must be done without losing the long term view of what it is needed to be build.

Understanding the company needs

The first round is about hearing but not listening. Data is one of the key start components of a company to operate. The company survived without you or your data team. So for sure there are a lot of data processes established all across the company. By understanding how there are done and being able which is the maturity level of the company towards data, you can save a lot of miss-understandings and open you a door for quick delivers.

In this step, collecting information like:

• Which is the data that they use to carry out with their data dependent processes?
• How they collect the data?
• Which tools they use in order to transform and process this data?
• What are the data that they missed and they would like to have available?

Going across different people in the company with their different data needs and their own implemented solution for their unique problems will get you the opportunity to find unnecessary redundancies and inefficiencies that can be easily solved by your data team.

All these high-level information will help you do define tech requirements and to set direction of your data team and identify the low hanging fruits to collect.

Understanding the company direction

The next round of understanding

On-boarding them as stakeholders

Your impact will not be conditioned to who gets the credit. You come from nowhere and the company will the opportunity to compare the state before your arrival and after your arrival. Don’t bother yourself in credit fights and try to maximize output delivered around data.

Making the stakeholder believe that they are on top of the whole design of your data infrastructure and data ecosystem will not be a

###

what is being done and what is the reason for which is done in that way

Talk about

Understand context as first task:

• You need to understand which kind of challenges

Game of balances

Self-sustainability of the data team. Avoid lock-in.

Key points:

• Infrastructure
• Team
• Data strategy (self-service, distributed, centralized. Dynamic from one to another)
• Data outputs

Measuring an analytics team