Posted: 07 May 2021 8 min. read

Scaling data science to deliver business value

Practical ways to address common challenges

Data science, as a modern-day practice, was born decades ago by the coupling of the mature discipline of statistics with the emerging field of computer science.

Due to the exponential growth of data volumes and increasing computing capabilities, over the years, businesses have invested significantly in exploring ways to turn their data into insights that drive value through better-informed decisions. It is against this background that data science has evolved to also include Artificial Intelligence and Machine Learning capabilities and emerged as its own profession—one that is critical for deriving value from the vast stores of data available today.

Let’s dive into some of the nuances of data science—its current state, some practical applications, and even some challenges and practical ways to overcome those.

So, what is Data Science?

The classical definition is simply using scientific processes, methods, and algorithms to extract insights from data. Another common view is using algorithms to not only look back into what happened but also to predict what may happen in the future. Or even better, going a step further in being more prescriptive and suggest choices that will lead to the most favorable outcomes.

Data Science is a field that leverages existing data sources to extract meaningful information and actionable insights. A data scientist does this through business domain expertise, effective communication, and results interpretation, utilizing all possible and relevant statistical techniques, programming languages, and software solutions.

Some real-life examples of Data Science application

Today, most things are dependent on data. It does not matter if you sell paint, run a marketing agency, or have a chain of restaurants; data is essential. You use it all the time—from web search to hiring new employees or creating financial reports. It is all about data. 

Here are just a few examples of how data science leverages data today: 

  • Predicting win rates for opportunities to help leadership better understand their sales pipeline.
  • Providing client engagement teams with cross-selling recommendations that they can offer their clients.
  • Helping resource managers better staff engagements, and 
  • One of my favorites, especially in challenging times like these, helping executive leadership better understand the revenue forecasts using extremely granular data and external macro-economic indicators to do a much better predictive forecast of the revenue.

Challenges that hinder Data Science’s progress and scale in an enterprise

A key part of my job is helping Deloitte’s data scientists do their jobs more efficiently and effectively. Here are some of the most common challenges we have addressed when it comes to data science:

  • Access to data
  • Processing capabilities and capacity
  • Lack of data science talent

The first two challenges can be addressed by focusing specifically on the unique needs of data scientists and developing enterprise data capabilities that can accelerate secure access to quality data. For example, a “Data Science Lab” platform can be configured, leveraging all the latest and greatest cloud-native services that includes extremely elastic storage and compute, and is made available to all data scientists in the enterprise. The ‘lab’ platform can be made more valuable with seamless integration to curated, enterprise data, where data scientists can quickly and easily request data access for their studies. This type of Data Science Lab platform allows all enterprise’s data scientists to quickly and efficiently deliver the insights that their business users need. The common platform and tools will also foster significant collaboration and sharing between data scientists/teams that are often siloed.

Concerning the third challenge, lack of qualified and skilled resources, there are multiple options to address this. One of the most basic options is to simply hire more expert data scientists or the second option is to upskill existing team members. I tried both. However, my observation is both options can be costly and time-consuming. Another option is to add more junior data scientists to the team. Again, I put this into practice. And while universities are producing more and more junior data scientists and talent is available, this approach is likely to have challenges with team scalability, as there is a limit to the number of junior team members in a team.

Another option we’re now exploring is to enable a new persona—the ‘citizen data scientist’.

How does the Citizen Data Scientist differ from an Expert Data Scientist?

A citizen data scientist can be defined simply as a business user or analyst who clearly understands a business problem, the relevant business data, and has access to tools and technology that allow the individual to get to the required insights, with minimal or little ‘coding’ and without deep experience in statistical methods.

The critical component to this delivery model is both the availability and capability of tools that can abstract much of the complexity in traditional coding and algorithm development. These tools must work for the ‘clickers’ of the business world and not ‘coders’ of the technology world.

Now, with the potential for a citizen data scientist persona identified, does this mean we would no longer need expert data scientists? Absolutely not, and I cannot stress this enough! The premise of this staffing model to address the capacity challenge is centered on the expectation that citizen data scientists and expert data scientists will frequently collaborate to get results. As the citizen data scientists receive guidance from experts and even take on some of the work previously done by expert data scientists, the expert data scientist’s unique skills will be better leveraged.

The use of citizen data scientists is an emerging model for addressing the capacity challenge, but we’ve seen strong results in the early stages and the no-code/low-code technology is rapidly improving. As our teams gain experience and learn lessons, I do expect to see the model being used more frequently.


Looking at the transformational changes happening in Data Science over the years, the pace of change is only going to accelerate, and the value add will accelerate as well. 

If you are not reaping the benefits yet, it is not too late—now is the perfect time for you to jump on board and start the journey! 

Key contact

John Irvin

John Irvin

Chief Data Officer | IT Services

John is the chief data officer (CDO) for Information Technology Services (ITS), collaborating with Deloitte’s business leaders to understand enterprise-wide needs and shape Deloitte’s data strategy objectives for safeguarding trust, enabling value creation, maximizing cross-business capabilities, and driving new efficiencies. As ITS CDO, John is accountable for ensuring alignment and adoption of the Deloitte data strategy, standards, and governance, as well as identification and implementation of leading-edge technology solutions that support the data strategy.