Are you really working at a

What do we mean when we say “data-driven”?

Many organizations aim to become more data-driven, but it is difficult to know where to start. While having Excel spreadsheets and integrating business intelligence tools is a good start, it might not be enough to permeate through the levels of a company.

What do we mean when we say “data-driven”?


Data-driven leadership

How is your executive team currently making strategic decisions? Are there tangible metrics that they’re tracking to evaluate their performance and identify new opportunities? In order for a company to become data driven, it’s critical that top leadership champions the cause and prioritizes concrete results over hunches. While expertise certainly is valuable, many studies show that rudimentary predictive analyses can have similar, if not higher, accuracy. Last year, CaseCrunch challenged lawyers to predict the outcomes of court cases and their AI program beat their prediction accuracy by 20%. This isn’t to say that lawyers will be automated (context is still king, after all), but leveraging predictive algorithms can supplement strategy and improve accountability across an organization.

It’s no longer enough for top leadership to delegate the data responsibilities – executives have to understand how data science can be applied to their industry and how they can leverage it to gain a competitive edge. They have to set the direction of data collection and allocate resources to data storage and infrastructure. Without a good knowledge of the data landscape, it is difficult for the top team to build out a cohesive and current data strategy.


Key questions your executive team should be asking:

  1. Which decision makers should be involved in planning a data strategy?
  2. Does your manager or leadership understand data science?
  3. Does your manager or leadership support data science?

Data infrastructure and tools

Do you have the right data infrastructure and tools to be able to collect and analyze your data? This is one of the biggest lifts for any organization, especially if they’re new to data collection. Through our Data Science survey of professionals, we found that a big portion of frustration came from the time consuming task of data cleaning and preparation. Many data scientists and analysts spend time deleting duplicates, editing column names to merge data sets, replacing N/As, and other tasks that can be prevented by good data practices. Our follow-up piece about the Plight of the Frustrated Data Scientist delves into 4 key best practices that can help substantially reduce the amount of time your data people spend cleaning data. Whether you use Hadoop (a distributed database system) or SQL Server, the overarching rule is to build out standardized data dictionaries and guidelines so that your data sources will be formatted similarly.


In terms of data tools, the question you should be most mindful of is “How will this tool help us achieve our objectives?” We’ve seen many companies invest up to millions of dollars in business intelligence tools that end up collecting dust or only being used by a handful of people. While a lot of data tools can be very useful in building data dashboards or quickly putting an analysis together, a tool does not fix underlying data issues nor does it check for underlying assumptions and context. The best implementation of a data tool comes with the right data to input and the right training to get your staff up to speed on how to use it well. Make sure that you plan for both of those facets when you’re thinking about purchasing a data platform.


Key questions your team should be asking:

  1. Do we have a written data dictionary or data guidelines that we are using to ensure that our data is standardized and clean?
  2. Is our infrastructure built for scale and is it secure?
  3. What data tools will help us meet our objectives, and do we have staff who will be able to leverage them?


Data workforce capacity

Even if you build a flawless data strategy and have your infrastructure set up, it won’t run without a team of data professionals to execute it. This is typically one of the more challenging aspects for any company. Because data science is still a nebulous field, and it can be hard to determine which employees / candidates have the skills you need, it is imperative to understand your data science goals and data infrastructure in order to guide your hiring and training practices.


Key questions your team should be asking:

  1. Do we have the infrastructure and the need to hire data scientists?
  2. How many current staff feel comfortable working with data or are interested in working with data?
  3. What training opportunities are we providing for current employees to become data driven?


Related Blog Posts