Today, fewer than 50% of corporate strategies call out data and analytics as key components in delivering business outcomes, according to Gartner1. With 65% of enterprises planning to increase analytics spend in 20202, it is important to understand why the best of intentions and investments may not translate to business outcomes.
A greenfield analytics journey begins with descriptive analytics to make sense of the data in place and understand current and historical performance better. Sounds simple enough, but setting the right stage in this phase is key to harnessing the right buy-in, sponsorship, and ultimately, driving adoption.
Problems that stymie a successful start arise from multiple aspects:
- Buy-in: The technology is ready, but business is not. A common fallout of a central team focusing largely (and often, solely) on the technology decisions like choice of tools and licensing models early on. While the answers to these questions are imperative, focusing on just these will mean a mad scramble for business ‘owners’ when the clock starts ticking.
- Requirements: The value of the initial use cases is not well-defined. When the business objective is not clear, use cases get identified and prioritized primarily on a first-come-first-serve basis.
- Data: Reaching for more than what the data is capable of. Disparate, inaccurate and irrelevant data will be a ground reality in the beginning and these need to be dealt with before addressing more complex use cases. Leapfrogging without a ‘single source of truth’ in place is tempting, but not sustainable.
- Tools: Succumbing to ‘spreadsheet hangover’. Replicating Excel reports in the new Business Intelligence (BI) landscape not only deters from leveraging the power of a visualization tool but also impacts performance. When a tool’s objective, functionality and limitations are not considered in the design phase, the outcome is unlikely to be well-received.
- Methodology: When the approach hinders more than it helps. Project methodology decisions are driven by the availability and constraints related to time, scope, budget and resources. Agile methodology is the popular choice for greenfield projects as it allows for iteration of requirements. But lack of clarity in what constitutes a successful minimum viable product (MVP) can derail projects with multiple unproductive iterations.
These issues ultimately lead to a lack of adoption once you go live, because users don’t see enough differential value to switch over from the status quo. What can be done to avoid these roadblocks?
- Align business from the get-go. Cross-functional champions are as important to success as cross-functional teams are.
- Balance business outcomes with quick wins – shape use cases with the relevant business functions that address existing pain points. If your organization is grappling with increasing customer churn, for example, a good starting point would be understanding attrition patterns and pockets. This could give business insight into the channels and markets where retention programs would be most impactful. As your analytics journey evolves, your roadmap could include building models to predict churn and eventually proactively engage with high-risk customers to extend customer lifetime value.
- Conduct a pragmatic feasibility assessment from a data perspective. It is also important to outline a data quality and governance approach at the beginning to ensure current and future data are more ‘usable for analytics’ and to build data trust.
- To build ‘fit for business’ use cases with the chosen landscape, the business needs to understand the tools and what it can do for them. Arrange for training early on and not just at the time of going live. But this alone is not enough. It is also equally important for the technology teams to understand the business a little better. Sharing requirements over emails will not work. Schedule workshops to brainstorm use cases and outline imparting basic business knowledge to technical teams as part of the agenda.
- Define priorities in requirement definition - which are the most important KPIs, which ones are good to have, which currently exist and cannot be done without, which cannot be considered at this time. Essentially, a MoSCoW of sorts so the team spends time debating requirements proportional to the value.
- Take the time to analyze post-deployment feedback from users and feed the learnings into the next phase.
A successful initial phase with demonstrable benefits will help steer your capabilities towards the ultimate objective of data-driven decision-making and set up the analytics program as an able supporter of a larger business transformation.