One needs to use the top-down approach as well as the bottom-up approach to build a solid foundation. It won’t help a business if the leadership needs months or a year to get the high-level picture and create strategic alignment. It won’t help the strategic intent if the operations drown in work without a compass.

Processes and data are interrelated. I wanted to say that it’s useful to start with the processes, though frankly there is a mix between mission, vision, goals, objectives, functions, processes, and data, and the interplay between these elements need to be considered.

For the leadership and the knowledge worker it must be clear to what questions they need an answer to support an organizations’ operations, tactics, and strategy. Then one looks at the infrastructure and tries to identify the gaps. Moreover, alignment is needed at various levels, and this takes time.

Also jumping blindly into data cleaning and other activities without setting priorities and knowing the usages of data can easily lead to waste of effort. Conversely, operations and data science will (always) need row data. Having the data already into the data lake can shorten the time from idea to results.

Data capabilities can be decoupled from the data based on the declared goals and objectives. I’ve been working for more then 23 years in the field, and I can’t tell that the input for the Data Analytics program coming from the top can be upon case important and must be addressed adequately, though in what concerns the infrastructure a lot can be achieved in parallel once the high-level requirements are defined and communicated across.

Aiming to reduce uncertainty sends the wrong message. You’ll always have uncertainty no matter what you do. It’s more important to understand the business, which are internal/external enables and factors, what opportunities and risks are there, respectively of having the agility to change when the environmental or circumstance change.