How important is data culture to your overall data strategy? On our DataFramed podcast, Taras Gorishnyy identified the key pillars for data science, AI, and data strategy within an organization:
- Executive support
- Vision for analytics
- Robust data foundations
- Establishing the impact of analytics early on
- A distribution of data skills and a data culture
”Everyone at any level, whether it’s C-level or entry level, should be looking and diving into data the same way you were expected to start using email 20 years ago.” —Tanya Cashorali, Founder at TCB Analytics
Build data fluency across your organization
A data-fluent organization is one in which everybody knows how to dive into the data that they need to do their job. For example, our VP of Marketing might not need to write SQL code or any code at all, but to do her job well, she does need to know how to access the analytics dashboards and interact with them. She also needs to know how to ask the right questions of data scientists and how to use the results that they give her. The way we think about this at DataCamp is summarized in the figure below:
A strong data foundation requires starting with a data engineer, and then hiring data scientists and data analysts. Then, we can bring business analysts into the fold, along with other departments. This process helps everyone become data literate, leading to the company becoming 100% data fluent. Basic understanding of data tools and resources across a company greatly improves the quality of interaction among colleagues, allows teams to make better requests, and empowers everyone to make decisions autonomously.
In order to gauge the state of data fluency across industries, we conducted a survey in which we asked over 300 organizations to what extent they have taken actions to become data fluent:
On the left, we have the mature companies with mature data fluency competencies, and on the right, immature competencies. Of the mature companies, only 39% had implemented process redesign and culture change with respect to data and only 7% of the immature companies had. There’s a huge amount of improvement to be made here.
Involve your employees
It’s important to prepare your employees for the future of work when crafting your data strategy. This requires ongoing conversations with all of your employees about what type of tasks you think will be automated and what won’t, how your employee base can coexist with the tasks that are automated, and what that human-machine interface actually looks like. For some roles, we will see job automation, but the bigger challenge across the board will be task automation, since certain parts of someone’s job may be automated away. It’s far better to figure out how to upskill and re-skill these employees to transition them to do more high-level, meaningful work.
Prioritize upskilling employees
Another question we asked in our survey of 300 organizations was, “What type of business challenges prevent companies from building or improving data fluency?” The majority of organizations flagged their difficulty in hiring top talent, as so much data talent chooses to work in tech these days. This means you won’t be able to hire as many unicorns as you’d like, so you should prioritize upskilling and reskilling everyone in your organization.
To build data skills, many of the organizations we surveyed have considered the following options:
- Establishing data universities or center of excellence
- In-person training
- Online learning platforms
- Internal campaigns to promote and incentivize learning
- Sourcing solutions from vendors
In polls we’ve conducted on DataCamp webinars, over 50% of respondents have said that less than 50% of their data work was actually used to inform decision making. This leakage is due to culture!
Call to action
List three to five outputs of data work in your organization and audit them with respect to how much your organization actually uses them to inform decision making. Too often, organizations hire a lot of great data analysts and data scientists to build dashboards and machine learning models, only to see them not be utilized in the intended way due to cultural challenges. All this good work can be wasted, resulting in a poor return on investment.
Find out more about best practices to building your data culture in The Definitive Guide to Machine Learning for Business Leaders.