Our years of collective experience in delivering both large scale and Proof of Concepts (PoCs) in the BI and analytics space have highlighted the value of adopting “tiny steps” towards addressing a larger business problem.
Across the industry different methodologies have evolved for implementations of IT projects and more specifically BI. Gone are the days where BI projects would span across multiple years with no tangible business deliverable for months and even years. Business is demanding more for their money and IT spend, and want to see value in most cases in as little as three weeks now. Technology itself has facilitated this business change, for example Kimball methodology of delivering function specific data marts is much more preferable for most businesses than a large scale Enterprise DWH encompassing the entire organisation. Agile BI is also more widely used than a standard waterfall method for delivering BI solutions.
Implementations usually adopt a corporate or project mandated methodology. One such common methodology is – “Waterfall” which is based on a philosophy that all detailed aspects of the business are well understood and documented for all phases of implementation (requirement, architecture, design, development, testing and deployment). Our experience of projects have shown waterfall works well in most instances but face challenges especially on instances where
- Requirements are unclear. This may be due to varied reasons and few are highlighted in this article (New technologies or greenfield nature of the project make implementation difficult). The very nature of BI projects mean business users start to understand more about what is being delivered only after few iterations of the solution.
- Requirements are subject to change frequently or change during the development process.
- Stakeholders are aware of the timelines for each phase of the project; however the fear of “missing the gate” is inherent in the methodology. Late requirements are usually not reprioritised and do not a find place in the current plan. Based on the relevance and business urgency of the requirement, these are accommodated in subsequent releases.
- Business users not involved during build phases.
The closest we could find to the suggested “tiny steps” approach on project execution is a methodology in Agile called “Spikes”. Most agile teams working on a “spike” (duration of which is usually time boxed within a sprint) had valuable findings to report back to the team on their “fact-finding” mission. The outcome of a spike usually empowers decision makers (product owners and team) to make a call on the way forward. Further we have realised that Agile is well suited for transparency with tangible results but projects adopting Agile requires management commitment to the core principles of the methodology.
More often than not, most projects are faced with technical challenges during various phases such as Analysis, Architecture, Design, Development and Implementation. This is particularly true in the following scenarios
- Greenfield projects or where organisations are adopting new technologies;
- Unclear business requirements;
- Organisations deploying resources not adequately skilled with the technology.
Other challenges that we have commonly seen is the limited availability of technical documentation required for implementation.
The answers to most of the above seem directly related to factors such as cost, budget and most importantly organisational and business drive.
One feasible option to address the above challenge is adopting a PoC or a prototype to assess viability. We will delve further into this topic in our next blog.
At DataQuarks we assist organisations attempting to embark on BI/DWH, Big Data or Customer Analytics through a set of pragmatic PoCs, pre-packaged as “Starter Kits”. Each starter kit is well structured and hand-holds in-house resources during implementation phase. The starter kit is structured to addresses essentials required towards fulfilment of a larger strategic initiative within the organisation. Each starter kit is a time-boxed and cost effective offering to assist and assess feasibility of a concept within the enterprise.
For more details, contact DataQuarks – https://www.dataquarks.com/contact-us/