Piyanka Jain, author of "Behind Every Good Decision: How Anyone Can Use Business Analytics to Turn Data Into Profitable Insight" and CEO of Arying.
“I was interviewing for software development jobs and I see this job for Senior Analyst on Craigslist for Adobe. I read the job description and like, Oh my god! Like my whole body, my whole being lighted up. This is me! This is like dream come true! I would love to do this!”
The BADIR Framework marries data science with decision science. It is an acronym for these five steps
- Business question
- Analysis plan
- Data collection
- Insights and
- Recommendation
It’s an equivalent of Lean Six Sigma Dynamic for data science. For those who are familiar with CRISP DM, BADIR adds stuff on the front and the back, as well as marries the entire data decision science aspect, the business constraints, and stakeholder interests.
Data Maturity, how accessible is data? Do the right people have the right level of access to the right data? Usability is another one. Single source of truth. That’s like data maturity. Governance, Infrastructure, and all of that is Data Maturity.
Data Literacy: are people at the right level of data literacy? Do they have a common vernacular? Are they coordinated? And when you’re talking about right level of Data Literacy, it’s not one-size-fits all. There’s a different level of data literacy needed. And so that’s the second part that we assess.
Data Driven Leadership; it all starts from leadership. Leadership that doesn’t emulate data driven decision making and doesn’t hold the team accountable will spend money wherever they like. Then you will not have a data driven organization.
The fourth D is Decision Making Process. That’s what pulls it all together. Do you have a structured process of making decisions?
That’s the first thing: business questions. They should be able to understand; What is it that project that you’re doing? What is your goal for that project? What is it? What is it that the business needs? And it’s typically not what they say they need. It’s when you ask very refining question. You have to refine that from an early ask to what they really need.
The second thing is you need to create a very decent plan. What steps are needed to get you to that goal, whatever that goal is? And that should have hypotheses. Your business folks, your data folks, and other folks around the stakeholders, and you yourself! You have very good hypothesis around that problem.
What do you want the business to do differently because of your analysis? One of the things that people fail at – especially data scientists, sometimes analysts – they put in numbers and they expect that somebody will create the dot, dot, dot to some action. Why make people work?? They will probably not work, and your analysis will go to waste.