Getting Started with Analytics in your Organization: Think Big, Start Small

This article is the second in a series of four, where we mention some of the most important considerations before taking the big leap towards analytics. Please read the first article if you have not, before proceeding.

Introduction

To be able to succeed in today’s digital era, business managers have to be hyper-aware—of an agile business scenario, their organization, competition strategy, ever-changing customer demands, and rapidly evolving market disruptions and opportunities. Analytics is the only way to capture these insights while keeping up with the speed of digital business.

With the widespread adoption of emerging technologies like artificial intelligence (AI), machine learning (ML), and intelligent automation, the trend of analytics adoption will only go up in the foreseeable future. Despite a proven promise of analytics, traditional businesses (think of all 20th-century industries producing tangible goods) might just be scratching the surface.

While some companies have developed pockets of localized analytics processes, very few have ingrained analytics or believe it to be a differentiating capability within their organization. More surprisingly, companies continue to struggle with fundamental issues related to analytics adoption, including organizational setup, workforce challenges, and change management.

Many companies are rushing to climb on board without rigorously evaluating business needs or competencies, often leading to below-par RoI from an analytics implementation. Here, we discuss some key pointers for business leaders to keep in mind while planning their analytics journey.

Getting Familiar with the Analytics Landscape   

Sponsorship and Strategy: A Leap of Faith

Creating a long-term vision for analytics in the organization and backing that vision with the right commitment and sponsorship is crucial. Sustained investment in tech, resources, and training is needed to ensure enthusiasm and buy-in for analytics across the organization. A lot of puzzle pieces across business processes need to be put in place, and different stakeholders need to be in-sync with each other before zeroing in on an analytics strategy. Hence, you should not expect immediate results.

A long-term data-driven strategy needs to be chalked out, keeping in mind the evolving business environment and organizational goals. While technical overhaul and the process of setting up data and IT systems to support analytics might seem more overwhelming, what is often a slower and more intensive process is the organizational shift in decision-making and the adoption in day-to-day operations.

Alignment with Business Goals: The Marie Kondo Way

Most organizations still use analytics in a decentralized form, scattered across regions and business functions. It might already be in use on an ad-hoc basis, in limited pockets, and without a proper strategy or roadmap. An enthusiastic group of analysts in APAC might be using predictive models for sales forecasting, while teams in the US might still be relying on traditional methods – taking inputs from sales reps.

While such a distributed approach was the only way out in the last decade, given tech constraints, the advent of cloud solutions has made analytics concentrated, scalable, and deployable across geographies. Analytics as a transformation exercise can now be managed more efficiently and effectively with a centralized team and vision. Ownership from managers, who are in the best position to pick the right KPIs, processes, and business decisions that can benefit the most from analytics, is key here. Such managers are often well placed to do a cost-benefit analysis for each analytics initiative.

Putting an Analytics Roadmap in Place

Prioritizing Activities: 

Most of the eye-catching and flashy use-cases are helpful only when the analytics practice has reached advanced stages of maturity. Often, analytics teams end up with more project proposals than they can take on. Much of the initial success depends on the team’s ability to punch through hyped technology (think of Elon Musk’s Neuralink or humanoid Sophia).

However, instead of trying out ‘moon-shots’, the focus should lie on tried-and-tested value-delivering solutions. The decision to develop analytics capabilities should be based on moving towards a value-driven approach rather than reaching towards the claim of being an ‘AI-driven organization’. The following figure showcases one such prioritization framework.

Target Setting: Expectations v/s Reality

Each business is unique, and hence, expectations from analytics should be clearly chalked out at an organization and project level. While it is important to set benchmarks by internally assessing the current state of capabilities and comparing them with leading industry practices, one should account for subtle differences in business models and scales to set realistic expectations of possible RoI from analytics. External consultants with specialized experience might also be able to help make RoI estimations and plan investments accordingly.

It is important to start small, with a couple of pilot projects, stay focused, and deliver quickly on these initiatives. This approach not only helps provide visibility to analytics within the organization but also proves the potential value of deploying analytics, helping pitch to and get acceptance from stakeholders for more challenging use-cases. It might not hurt to employ successful stories from analytics leaders by customizing such solutions for one’s business and provisioning for differences in scale, business model, processes, and industry.

Team Setup and Training

Organizational Structure: 

Organizations need to define the placement, structure, and cost and resource allocation to business units for their analytics teams. While there are multiple possible team structures, including business-unit led, centralized, or outsourced, a ‘Center of Excellence’ (CoE) approach works well for most global organizations. A CoE acts as a cost center, and analytics projects are assigned for different functions or business units. Further, there should be primary assessments carried out for multiple teams or functions to check if they are ready for collaboration with analytics teams. The team structure should be defined after considering the technical and resource dependencies.

For example, in a CoE model, data silos from different IT systems need to be moved to a central data repository. In such a case, businesses should also invest in allocating talent accordingly, using internal movement, hiring, or outsourcing. Also, define the Ownership and coordination guidelines for cross-functional initiatives clearly. It might need functional and technical expertise, and hence, most organizations either have external consultants or hire industry-veterans as Chief Information Officers (CIOs) or Chief Data Officers (CDOs).

Workforce Challenges: May The Force Be With You

Qualified analytics talent is scarce, and talent sourcing and retention to deliver on analytics ambitions can be quite challenging. While analytics is an umbrella term, covering anything from simple excel pivots to advanced GAN models, hiring an analytics team with a fusion of technical know-how and domain exposure can be tough. Further, talent mix requirements keep on changing with the analytics curve of the organization.

For instance, while setting up analytics, more data engineers and database experts are needed to set up databases and ETL (extract-transform-load) processes, but once these are in place, developers with skills in Business Intelligence (BI) and Machine Learning (ML) are required to build applications and tools on such databases. To drive analytics adoption, hiring business managers with an analytical mindset is as important as hiring technical resources. It is not only the talent attraction process that is time and cost-intensive, but retention is an equally challenging process. In the case of small teams, attrition of 1-2 specialized members can potentially derail multiple projects. The following table summarizes some common challenges in managing analytics talent and some solutions.

Closing Comments

Most of the challenges discussed here can be addressed if businesses take out the time to develop a comprehensive, enterprise-wide analytics strategy, not just on paper but supplemented with an operating model designed to tap the potential of analytics. Often, there is a tendency to overlook simple yet effective solutions. However, simple solutions are often necessary stepping-stones to developing more advanced analytics capabilities.

Much of analytics success lies in an organization’s ability to identify business problems, formulate those as analytics problems, and choose from available solutions across digitization, analytics, automation, and reporting. A clear roadmap to go from ‘issue to outcome’ is critical, as it allows organizational analytics to focus where it should – on directly enabling problem-solving, decision-making, and implementation, hence delivering improvement in business performance.

About the Author

Amit Kumar

Amit, a Data Science and Artificial Intelligence professional is currently working as a Director at Nexdigm (SKP), a global business advisory organization serving clients from 50+ countries. He holds over 15 years of experience across industries and has worked from both perspectives, as an internal functional expert (in Vodafone, Aviva Insurance, GE) and as a consultant. A passionate advocate of data science, Amit constantly endeavors to create optimal, actionable solutions, that help derive measurable business value from data.

Connect with Amit – LinkedIn.