Digital transformation strategy
A business perspective for analytics
by Stan Maklan
Management has truly grasped the significance of the data-rich era in which we now compete. Whilst there is little new in the deployment of statistical methods to business problems, i.e., analytics, the context has changed.
From a marketer’s perspective, the combination of enormous quantities of consumer behavior data, the technology to assemble all that data, and artificial intelligence generate step change improvements in business operations as well as opportunities for disruptive new business models. The potential of big data will only increase with progress in blockchain and local printing. Companies are making big investments in data and analytics, and this generates increased demand for managers and skilled analytical talent.
Business context, customer and market know-how are necessary complements to modeling and machine learning.
This growth in analytics creates several business challenges. Firstly, where will the skilled data scientists come from? How will an increasingly numerate analytics team work with managers? How and where do analytic capabilities fit into the organization? Which investments in big data shall firms make? Where is the evidence for return on those investments? What impact does this have upon the organization and relative power between functional groups?
A prominent stream of research suggests it is not the technology or algorithms that make businesses prosper. Rather it is the ability to benefit from their use that is rare and valuable. All well-resourced companies can buy the best software, hire excellent data scientists and build excellent data pools; not all will generate context-specific insight and turn that into winning strategies. The “answer,” the new algorithm is only one piece of the puzzle, albeit a vital one.
Most “how-to” prescriptions for digital transformation offer a standardized change-management recipe: Start with the CEO (poor CEO is very busy with all these initiatives), appoint a digital transformation steering committee, create separate and protected spaces for new digital models to take root, and maintain a radical vision.
But this can guide almost any management initiative and assumes a managerial, top-down driven program of change. Our initial research into digital transformation suggests that the analytics revolution has some unique characteristics:
- Unlike most technology-led transformations, those building analytical capabilities do not seem to start with large capital expenditure. Small teams, sometimes even one data scientist, can analyze data with open source software, write algorithms and implement them on commercial websites for large firms in a matter of weeks. Contrast this with the two- to three-year, hundred-million-dollar (plus) implementation of supply chain and customer relationship management software (ERP and CRM). This is not “big IT”; any organization can extract value from its data. The challenge is to sustain “bottom-up” change over time.
- The rapid experimentation inherent in analytical thinking makes digital transformation more of a learning endeavor than a managerial, investment-driven change program. There is a greater role for reflective managers, inquisitive analysts and innovation throughout the program. Do we have the structures in place to support this? Moreover, are marketing and other business managers confident enough in their analytical abilities to participate fully?
- Business context, customer and market know-how are necessary complements to modeling and machine learning. The projects that produce the largest business benefit quickly are those where business managers work closely with analysts supporting rapid cycles of learning and experimentation. Functional roles and power are “in play” in this new environment. Marketing typically has the customer and market knowledge but lacks deep analytical skills and data technology engineering. Data scientists have the analytical and programming skills, but often lack privileged access to data and the market context. IT is rarely designed to support interactive, learning analytics, nor does it have the business context. There needs to be a realignment of roles, power and control in the data-rich environment we are entering.
I do not see one unique solution applicable to all companies (despite what is often implied by generic change management rule books). Each organization faces unique circumstances and objectives that favor one approach over another, or a unique mix of approaches. The changes are so profound that it may prove impossible to create the “if this, then that” instruction manual; a wiring diagram or “playbook” that offers an ideal play for each set of circumstances.
There are just too many contingencies to consider: Does strategy emphasize growth or profit maximization? Local or global reach? Business versus consumer customers? Rapidly changing versus stable markets? Being a broadly positioned market leader or specialized niche player? Participating in volatile high technology markets or stable, well defined ones? Disruptive strategies versus evolutionary ones?
Cranfield is working through these issues in the context of market- or customer-focused analytics – big data transformation. We are focusing on the roles, power and scope of three focal functional groups: marketing, IT and business intelligence. Will digital transformation be more or less successful if led by any of these three groups? What are the ideal configurations of structure, tasks and fit with strategy that promote the most rapid and successful change? Where do organizations start? How fast does the transformation happen?
We suspect that the answer we eventually generate will involve a number of themes evident from our initial work:
- Upskilling business people so that they can participate in the analytics revolution. JMP democratizes access to data and advanced analytics given its intuitive interface. I use it for teaching marketing students who are generally not trained in statistics or programming – and neither am I.
- Access to data will need to be democratized and no longer the exclusive competence of the IT department. The data that is interesting for the front office lies outside the firm’s operating systems and databases. Open access to large data sets is moving ahead. Soon, intuitive tools will allow business functions such as marketing to access data directly.
- Business functions, particularly marketing and sales, will need to develop their own data and technology strategies or at least be full partners at the top table when firms make such decisions. Analysts working closely with the business, or better still, embedded in business functions, can play a pivotal role enabling marketing to understand the data that is available, needed, how difficult it is to access, etc.
- Business analysts and marketers will work more closely and learn from each other. Core concepts of agile development (scrums, minimal viable product, etc.) will permeate throughout marketing, growing demand for analytic capabilities. Much of this demand must be met by an upskilled business establishment; we cannot train data scientists fast enough. This offers opportunities for analysts to assume greater leadership roles.
Software is advancing to address the skills gap and democratizing big data for business managers. Analysts will have the opportunity to enhance their own roles in coaching, supporting and contributing to an evolution of business.