The insurance industry is riding a wave of cloud-based data warehouse implementations and migrations from on-premises environments. As insurers plan their next moves, they find that they must make several critical decisions from a vast number of options. The choices they make now will have a lasting impact; moving quickly (while avoiding missteps) is critical.
Key decisions include which cloud platform, data warehouse, data lake, data movement, data governance, and business intelligence vendors insurers should partner with. Existing relationships, team expertise, ease of use, performance, capabilities, switching costs, and solution costs are prime considerations. The vendor community in each area is vast and diverse, and having an approach to decisioning will accelerate the project and generate confidence in the new direction.
Complete knowledge of all future needs is impossible at this stage of the effort, but developing a list of known primary use cases for the data environment will be the starting point for evaluations. Developing proof-of-concept projects based on time-boxed use cases and having pre-determined success criteria is a best practice. The time box will keep costs down and expedite results. Requiring that success criteria are defined upfront will accelerate implementation and ensure that the proof of concept is well-designed.
POCs can be scored on the following dimensions:
- Ease-of-use: The extent to which the team found the solution intuitive and team productivity.
- Performance: Measures execution times and resource consumption for complex computing challenges, and varying volumes.
- Capabilities: Define upfront the desired product features and compare them against the solution’s capabilities. Consider leveraging the MoSCoW prioritization approach, i.e., must have, should have, could have, won’t have.
- Costs: these proofs on concept should provide information on resource consumption sufficient to extrapolate it to production use cases, providing input for planning processes and enabling bulk purchasing commitments.
Reaching for the “best” decision may not be the most effective objective. An evaluation that focuses on “best” may drag on. Further, “best” may not be feasible, as significant amounts of investment and innovation characterize the cloud information management space. These are long-term choices; avoiding regrets is more important than reaching the “best” decision.
Add new comment