analytic community
Blogs home Featured Image

CIO’s globally ranked analytics and business intelligence as one of the most critical technologies to achieve the organisation’s business goals, with data and analytics skills topping the list as the most sought-after talent. As we embrace digital transformation, it’s clear that the need to upskill and resource data science teams has become far more pronounced.

Building a successful and lasting data community can often be one of the most significant hurdles to overcome, which takes time and effort. Often establishing and maintaining a thriving collaborative approach to analytics, with the right ecosystem for your community, can be a challenge.

Naturally there’s a growing need to ensure analytic teams have access to the best tools and latest methodologies to perform their analysis and find business wisdom. Alongside identifying the analytics skills already in place, a great place to start is also to identify the best tool for the job.

Nurture and aligning members of the community

Pulling together existing disparate data science resources into a single, connected community of practice, creates a secure foundation to grow analytic talent. Having such a community means the business will have a better understanding of the skill sets that exist within the organisation already, as well as best practice examples for approaching different scenarios and a better awareness of the tools and solutions that can be used.

Defining the right tools for the community

R and Python are still the two most popular and adopted programming languages. Both tools are open source, free to use and cover pretty much everything data science-related.

R was developed specifically for statistical analysis, so naturally is the popular language choice for statisticians. R has a large user community and an actively developed large library of packages which enables effective analytics. However, R can require a steeper learning curve and people who do not have prior programming experience may find it difficult to learn.

Python on the other hand, is considered the easier of the two most popular languages to learn. Its domination in machine learning is well-known. With an increasing community base, Python is commonly taught in Computer Science lessons in Schools and therefore the rated language of choice in academia. However, Python can be considered to have its limitations especially around speed and memory, so best practice use should be applied when considering Python.

It’s not a debate as such on which language to use, but more a conversation around empowering a team to become multilingual and multiskilled, so they can use the best language for the application.

Up-skilling of analytic talent

For an organisations analytics function to thrive, it’s critical to continually attract, develop, and retain key data skills & capabilities. Understanding the mix of skills within a data science team, as well as identifying gaps to unify skills & knowledge, is vital to drive analytic value. Establishing the support of a dedicated Learning and Development partner, who provides live, instructor led, data science training programmes, designed to equip and enthuse a data team with the latest approaches, can help address this challenge & unlock business gold.

Enabling training at all levels of data awareness will be critical, and this should even include training on how to use information, to guide decision-making.

Building a successful community provides a solid basis for working out where the talent pool needs to be extended, unifies talent across the business and enables quick wins towards embedding the right culture and building the required capability.

After 20 years of experience, we are a trusted data science L&D partner to leading brands worldwide. We train thousands of data science and analytical teams every year from a range of industries and backgrounds.

 

data science team
Blogs home Featured Image

There continues to be a lot of hype around data science (and it’s evolving world of connected buzzwords such as AI, ML etc).  This is driving unprecedented investment in the field, driving the race for data & digital talent into overdrive. But there is scarce data available on the return on investment in data science.  This can make it difficult to make a good business case for investment, and can hamper good discussion around how to size your data team. 

To resolve this, in November we released a report, Delivering Value Through Data Science, conducted by Data IQ, the only membership business focused on the needs of data and analytics professionals from large and midsized businesses. That’s an important start because if you want expert outcomes, we believe you should ask the experts… 

Before we get into the results, I want to sound a note of caution (well, I’m a statistician after all – no figures without caveats from me!).  Whilst the data we got back is interesting, this is a relatively small sample size and more research is needed in this area.  And of course, surveying companies associated with Data IQ will also introduce bias around the maturity of the organisation and the leadership qualities around the data teams. 

That being said, here’s an eye-popping but very debatable figure: The successful application of data science could lead to a 17.9% uptick in revenue for organisations with a “mature” data science capability. 

Ambitious but not impossible, perhaps? Previous successful generations of major IT change such as ERP or e-commerce could justifiably be said to have significantly inflated revenues at many companies (think of Tesco in online retail, for example). But we need to acknowledge that it’s very hard to pin success down precisely to a specific action; growth comes in many ways. How, for example, do you measure quality of execution, committed leadership, motivated employees? These challenges do not, as yet, have ready answers within the data science community. 

So while impressive, a figure of 17.9% feels unachievable for an organisation starting on their data journey.  Too high to be believed, perhaps (but clearly a figure many of the leading data-enabled organisations will stand behind).But, working in the space and seeing data science in action across organisations, the average figure of 6.7 per cent revenue growth (across all organisations surveyed) seems plausible. So, let’s agree in principle that careful investment in an ability to interrogate data and (crucially) act on these results and good things will happen for your business. 

We have definitely seen plenty of anecdotal evidence through our 19 years delivering data science projects, that back up at least this 6.7% figure: 

  • We’ve seen successful numerous data-led customer retention project reduce churn by at least 8% in our customers;  
  • Our higher ROI single data science project (a next best action engine in finance) was directly responsible for an 8 figure uptick in revenue;  
  • Deploying analytic solutions in areas where subjectivity has previously reined will routinely yield incremental gains of over 30% in target metrics. 

In our experience, the scale of the return is typically driver as much by change and culture as it is by algorithms and code.  Blocking the road to change are the familiar foes of inertia, cost justification, change management and skills. Ultimately, leaders must lead and drive their organisations forward to become data-enabled and data-driven because, as ever, business alignment is key.  

The DataIQ survey found a 4x increase in value where strategy is aligned with investments. Again, this matches well with our experience – doing data for the sake of doing data is very unlikely to yield results, and walking around your business with an AI-shaped-hammer looking for something to hit is not the way forward.  Data and analytics needs to be focused on, and deliver on, your business ambitions. 

Those that don’t lead from the front and pursue long-term data science programs will see rivals make better, more auditable decisions faster and will become cannon fodder for internet giants edging into their markets and for startups with no legacy. 

Today, pundits queue up to tell us becoming a data business is essential, and of course they’re right. We can’t all be Google but we can learn about data, harness it and apply it. Goldman Sachs hired more software engineers than Facebook, Gousto applies data to “little touches” that delight customers, Black and Decker reengineered production and dispensed with what one executive called “lazy, rusty asset syndrome”. These are all companies applying data to understand customer behaviour, improve service and drive competitive differentiation. This is not all about Tesla, Netflix and Amazon: data science needs to be pervasive and ubiquitous across all sorts of organisations. Start small to convince with data science quick wins by all means, but leaders also need to be persistent and not expect overnight gains. As with other game-changing capabilities such as AI, they need to be in it for the long term and learn as they evolve. 

“Data that is loved tends to survive,” said Kurt Bollacker and companies need to learn to love their data. The opportunities (and the risks) are too high to act it any other way. “Culture eats strategy for breakfast,” said Peter Drucker, so build a culture of continuous business improvement supported by data (and data science) now that will lead to data maturity and better, more consistent results over time. 

Start now. Download the research paper, compare and contrast against your own experiences and start thinking about your next steps. You may not see a 17 per cent revenue hike anytime soon, but you’ll be on the right path.