How Smaller Organisations Can Build Data Analytics Capabilities
Roadmapping for effectiveness instead of efficiency.
Background
How should small organisations get started on their data analytics / data science capability building and punch above their weight? Typically, an organisation with a few hundred employees will begin to experience friction in information flows, leading to sub-optimal decisioning. The value that data analytics (including data science and AI) can bring to address such organisation challenges has already been proven. So the question isn’t abiut what I should do, but rather how I should go about doing it. Is there a specific framework that a “small-ish” organisation can employ for its data analytics roadmap? And so I dedicate my 48th weekly article to unpacking this topic.
(I write a weekly series of articles where I call out bad thinking and bad practices in data analytics / data science which you can find here.)
Large versus Small Organisations
Thousands versus hundreds of employees; that’s a simple way to categorise large versus small organisations for the purpose of this article. For large organisations, maintaining consistency is paramount. This is just a natural outcome of size. Hence, the discussion around the pursuit of data analytics capabilities and maturity typically revolves around the automation of intelligence. This is essentially an efficiency play. How do I reduce and automate my business reporting? How do I develop and deploy my predictive models at speed?
When a small organisation wants to invest in either creating or upgrading its data analytics capabilities, consultants often hold up the large organisation implementation template as the go-to approach. This often overwhelms the small organisation and puts it off in wanting to seriously pursue this data analytics initiative. While the cost of technology and platform cost have arguably dropped, coupled with the liberal use of open-source tools, the large organisation implementation template requires deeper technical skills and experiences that may not be readily available in a small organisation.
Flawed Decisions
To punch above its weight class, small organisations should be laser-focused on driving effectiveness through its data analytics initiative. This means looking for flawed decision-making within the organisation. These flaws can be due to flawed or incomplete data, incorrect interpretation of the data, or simply bad reasoning. Looking for flawed decision-making is somewhat different from looking for sub-optimal decisioning, which is what large organisations tend to focus on. For the latter, the sheer size suggests sufficiently good decisions would have been forged through time and test, but sub-optimality arise due to trade-offs in consistency versus flexibility. The same crowd-based convergence of decisioning would be less likely in small organisations; they are not at the limits of their market and business opportunities.
The data analytics team in a small organisation should be actively challenging operating assumptions around (a) revenue, (b) risk and © customer experience; in that order. They should be conducting diagnostic analytics to look for and validate supporting evidence (in both data and logical thinking) for the current set of operating assumptions. They should be asking good questions.
- Revenue: Why is our target market so defined? Why have we set our pricing at this level?
- Risk: Are we overpaying for segments of customers who would not give us sufficient value? What are the early signs of customer disengagement that we should be looking at?
- Customer Experience: Why do we think our customers will want this product feature (utility)? Why do you think customers would give us feedback through these channels?
Why this order of Revenue → Risk → Customer Experience? The data analytics team in a small organisation needs to create or enable positive cashflow to fund its own growth in capabilities, be it in incremental headcount hire or investments into better data analytics tools and platforms. Playing the efficiency game is easier but won’t get you enough dollar-impact given the smaller size of the company.
Roadmapping
Given the stated focus above, the data analytics team in a small organisation should build its capability along the following path: Diagnostic Analytics → Predictive Analytics → Descriptive Analytics. Simply put, they should be solving the WHY that leads to solving for the HOW before improving on the WHAT. This may seem counter-intuitive. Many organisations embark on their data analytics roadmap by starting with Descriptive Analytics, i.e. improving business intelligence reporting and visualisation. Organisations treat reporting as low-value work (I disagree with that view!) and thus the whole approach becomes an IT engineering requirements exercise. Pursuing Descriptive Analytics also leads to flawed decisions to create offshore centres of excellence (COE as they are called, although I don’t know what they are excellent in) to move “low value” work to “low paid” locations. Small organisations cannot afford to get caught up in this mess.
By pursuing Diagnostic Analytics as the starting point of capabilities building, the data analytics team must get into the field, must be should-to-shoulder or embedded in their stakeholder functions. The aim is not to standardise data analytics tools and workflow at this stage, but to generate findings that have the potential to become actionable insights (i.e. improve the quality of decision-making). The Diagnostic Analytics work will surface up gaps in the small organisation’s current data strategy (collection, treatment, and organisation of potentially useful data), which is an important gap to address if the organisation is to achieve data analytics maturity.
During the next Predictive Analytics phase of this roadmap, one of the key areas of focus would be on standardisation of tools and workflow. Automation isn’t necessarily the aim because you want to retain a fair amount of agility and adaptability at this point in the maturity cycle. There’s a lot of learning about operationalisation and execution that needs to be achieved during this phase. Don’t be wedded to your process. Use open-source tools (or solutions built on open-source foundations) whenever you can; the days of proprietary tools are dead.
During the final phase of Descriptive Analytics, the data analytics team should focus on enablement of better intelligence rather than automation. Everyone is a junior data analysts in a small organisation. In a small organisation, everyone is already much more comfortable with their own (albeit limited) data. Everyone just needs to get better at data sensemaking and data storytelling; this linked to reducing the flaws in the operating assumptions due to incorrect data interpretation and logical reasoning.
Conclusion
Lifting and shifting the template for data analytics capability building for a large organisation onto a small organisation is generally going to be inappropriate. The opportunity space for the latter is fundamentally different. Their roadmap to be analytically mature must start with a clear focus on effectiveness rather than efficiency — create value instead of optimising value. Building critical thinking competency around the business operating model is the most important agenda to pursue for the data analytics team at the beginning. And ending with enabling the stakeholders across the (small) organisation to get better and more comfortable with their own data usage.
