Notes on a successful digital sales data mart development

"The development can be closed now. Thank you very much for the work done! Great job!”
We have received the above business feedback as an acceptance of one of our sales digital data mart developments completed in September. The development has got a green path, we can prepare for the live deployment , but how have we achieved it? What is the key to that the business world is satisfied with our work, appreciating our results in such feedback?
The key to success lies in the fact that in our project - adapting to the special demands and circumstances - we combine the waterfall model with the agile methods, so we take advantage of both worlds. We use the concept of the waterfall model, so that development is progressing linearly from the definition of a business requirement through planning and implementation to business testing. We keep processes built together, since in this field, after implementing the entire task and after completing the developer testing, we can provide the business customer with testable data. The purpose of the development is to create a data mart or marts that are used by our client for analysis, business development and various reporting purposes.
Because of the nature of the tasks, there is no room for 1 or 2-week-long sprints, so in the definition of requirement and at the planning stage, it is of the utmost importance to understand the business requirement as accurately as possible. In this needs assessment and in the planning stage, we use the benefits of agile methodologies.
In the definition of requirements and in the planning phase, we are in daily iterative relationship with the customer, as we keep personal status every week, and where business requirements are continuously refined and negotiated from their emergence.
With this working method, we cooperate continuously with the business area right from formulating the needs. More complex problems - which deplete the frameworks of the statuses and other communication channels - are discussed during operational meetings organized around an issue. As a result of this joint work and thinking, a definitive business specification is developed that fully takes into account business requirements and technical possibilities in this respect. While we mention technical possibilities, we do not mean technical constraints, but really opportunities as we have experience in the development process that it often involves extending the business needs, as long as we can help with business advice from the IT side.
There are also further parallel discussions from the business and developer side to companion IT fields. Ultimately, business needs are served directly by the data marts, so we can say that we are at the end of the development chain. In most cases, a dependency arises in the direction of the source systems that serve the data warehouses with data (middleware, front-end systems). Consequently, in the planning stage we must also focus on taking into account the capabilities of source systems and providing them with appropriate information. This is essential in order to provide updated and accurate data that are needed to serve your business requirements.
In the world of digital sales, circumstances change faster than average. A thorough preparatory work precedes a development with a 2-3-month lead-time, as mentioned earlier, however the time and energy invested here will bring multiple returns. When we reach the implementation phase, the task is precisely defined, well-planned and thoroughly documented. From this point, by the end of the agile reconciliation planning stage, we will actually move to the waterfall model. This does not mean, of course, that during development it would not be possible - or very hard - to modify, but based on our experience, this is rarely needed as a result of proper preparation. Here is the first point, at which the work and energy invested in reconciliation and preparatory process pays profit thanks to the smooth development phase.
By closing the implementation phase, we pass the development to business testing, which is the next stage of development where planning depths bring benefits to the team. The UAT (User Acceptance Test) typically consists of negotiating some issues, supporting business testing, and there is not much need for bug fixing or code modification. By reaching this stage, the development fully satisfies the business requirements, and thorough development testing before passing to UAT minimizes error occurrences. This phase usually ends in a short period of time (within a few days) and we can start preparing for development in live operation.
The third point, where good planning brings returns, is after the live deployment or maintenance stage of the task. Except for the operational tasks, bug fixings in this project almost never have to be sorted out.
Naturally, the digital sales data mart continues to evolve. New economic environment, new trends, new business needs, new challenges. To keep up with all this, we plan the next task in parallel with the business development, so that we anticipate for several months what kind of development is expected and what kind of results the customer can expect from the defined priorities. Thus there is enough time to negotiate the future tasks with business and technical companion systems, so that we can deliver them to the highest standards, expectations and in high quality to our customers.
Márk Kelemen
Photo: Pixabay
Office: 1139 Budapest, Lomb u. 15, 3rd floor
info@eisys.hu | +36 1 611 4117
EiSYS Számítástechnikai Kft. | HU13669555