The 4-steps that we have adopted and honed over several implementations to successfully rollout data analytics.
The 4Ds
The 4Ds are Define, Design, Develop, and Deploy.
Define
Define is an essential step to capture the requirements and deliverables. The project lead carries out small sessions with executive sponsors and key stakeholders to scope the requirements, understand the current challenges and define the goals and indicators. The indicators are required measures that will be monitored to improve performance and aid in decision making. The discussions must happen with the key stakeholders.
Design
This step is driven by the project lead who designs the wireframes and mockups of reports and visualisations. I have often used pencil, erasers, and sticky-notes to present my designs. Several tools are available to create wireframes and mockups. You will end up with multiple iterations and versions of your Design, but it is time well spent.
If you intend to create mockups using a Business Intelligence (BI) tool, be aware that it can become a time and resource-intensive activity. You will also lose time to create data sets for the BI tools.
Another essential part of the Design step is to identify the data sources, define the data mart/warehouse schema, and the data models. The technical team gets involved in this activity. The talks are usually with the IT or the database team of the customer.
Two Principal Steps
Define, and Design ensures that the requirement is documented and well understood by all. It is also the basis to create use-cases, test models, and user acceptance.
Develop
In consultation with the project lead, the technical team takes over. They develop the data mart, configure the ETL (Extract, Transform, and Load) process and designs the user landscape. It is crucial to create a well-designed data model. It assures that the performance of data mart/warehouse is optimum and flexible enough to address the dynamic nature of a growing business. As per the scope in the Design step, the reports and visualisations are also developed. The key to the adoption and usage of stories and viewing is that it should be well laid out, clearly marked, and un-cluttered. It is essential to test and validate data and visualisations internally and with stakeholders throughout the development step.
Deploy
The solution is deployed out to the production environment, on completion of all the tests. Usually, training to end-user and administrator are carried out. If it is an enterprise-wide rollout, it needs to be planned well in advance along with the stakeholders. The key stakeholders should be a part of the rollout process to ensure the adoption of analytics.
Break it down
If the adoption of analytics is found to be slow in an enterprise, it is suggested to pick smaller battles and achieve quick wins. Pick an area of business, LOB (line of business), or a department where the necessity for Analytics is the most, and its leader is prepared to make decisions based on facts. Implement it well for that area of business and secure initial success. This initial success will rub-off with the others and will provide the necessary impetus for the rapid adoption of Analytics.
Your thoughts and experiences
Please share your experiences in implementing analytics. What has worked for you? What have you done differently to ensure the adoption of analytics in your enterprise?