Ambitious data intelligence projects contain many unknowns and can cause surprises. With a strong methodical approach through the project, you are able to avoid the most important pitfalls, such as lack of well-functioning solutions and budget overspend.
When conducting an ambitious data intelligence project, the business idea is the focal point. In essence, the assignment is to develop the data model and algorithm, which can bring the answers that releases the potential of the business idea. Therefore, the business idea must be analyzed from the viewpoint of a business data scientist. This gives a first indication of whether there is sufficient data to support the business idea. Afterwards, there will be one or multiple proof of concepts and pilots to clarify possible solutions, and this has to have participation from experienced domain experts in data intelligence projects.
Close cooperation and effective synergies in the project team are central elements to succeed in ambitious data intelligence projects. Often, this is a significant challenge as many different highly specialized competences participate in the project, and they may not have a full understanding of each other's contribution required to fully succeed. The trick is to get the project members to speak the same language. It means having them complementing each other with the contribution from their field and reach a common understanding of the business idea, the possibilities and limitations related to it, as well as the challenges to be solved.
For diffierent reasons, not all origianal business ideas will be implemented. A project may develop in another direction than originally thought. Stopping the project might also become the right decision. The foundation for these decisions often relies in IT-limitations, lack of sufficient data, too costly, or simple that more profitable ideas have occured along the way.
If the project takes a different turn than first planned, it shouldn't come as a big surprise, nor for management. This is, however, easier said than done. With a strong and methodical approach to data intelligence projects, it's possible to avoid surprises when implementing an ambitious data intelligence project.
Use Rooftop's Predictive Analytics Model as methodological tool. With the focus on the core of a data intelligence project, the model provides a step-by-step approach, which allows iterative development and agile working methodology, when developing models and algorithms. It also ensures the necessary quality, allowing the company to rely on the results. When the project team follows Rooftop's Predictive Analytics Model, they are automatically tracked in the optimal way in which predictive analytics are to be developed.
Prioritize conceptualizing the business idea from a business data scientist perspective. When the essens of the business idea is analyzed, the next step is, as a business data scientist to investigate the domain to determine, if there are recent rearch that can support the thesis of the business idea, as well as investigate, if the available data is sufficient foundation for develpoing the predictions. Our business data scientist also identifies one or more pilot projects which should be completed. If it turns out, is not expected to obtain the desired results, suggestions on how to adapt the business idea will be given. Insight to what type of data is required will be shared. However, the outcome can also be, that the initiative stops, because it is too costly to establish the required data.
Take advantage of one or more pilot projects. The main purpose of a pilot project is to get an indication of whether the data can deliver the information that the company aims for. Start with a well-defined and limited area. Depending on the complexity of the assignment, it may be relevant with several pilot projects to uncover the algorithms or the way to structure the data before it can be included in the analysis. Pilot projects can, first and foremost, verify if the data contains the needed level of information. It's also an effective way to work agile as you can adjust the project purpose, scope, and schedule in line with the knowledge gained during the pilot project. It creates transparency and a foundation for the management to assess whether the business case can be fulfilled even though the assignment may have been adjusted since the beginning of the project.
Set the project team with care. It makes a difference, when it comes to how well the project generates a good results and how quickly it happens. Ensure a strong representation from the business and a project manager, who can facilitate the necessary cooperation between the different highly specialized competences. The business should preferably be represented by the domain(s) from where data is applied and the solution is to be implemented. A business data scientist shouldn't be mistaken for an IT-competence and must participate from the beginning of the project.
Designate the change agent - the employee with the largest share in the solution to come. It will strengthen the project team's understanding of the assignment and the related opportunities, which arise along the way. Also, employees tend to trust the new solution and thereby accept the new ways of working, when a change agent is engaged. It's also instrumental in increasing the business readiness so that the lead time, from when the solution is completed until the solution is applied within the business, becomes shorter.
Choose a project manager with insight and data knowledge to bridge project participants as they are far from speaking the same language. The team members are highly specialized in their respective disciplines and have limited knowledge of each other's disciplines and how those are relevant in solving the assignment. The sooner common ground is obtained for the assignment and an understanding of who has knowledge of what, the faster the project team start becoming productive and generates a solution of high quality. The synergies count, and this is central to the project's plan, budget and when the company can expect the coming product or service to generate future revenue.
ABOUT ROOFTOP ANALYTICS
Rooftop Analytics has deep experience in managing data projects that include solution development based on data. We follow a strong methodical approach in data clarification, data model development, and programming of algorithms. We are able to think through the project before start, and therefore we can identify the real cost and time horizon of the implementation.
Rooftop Analytics offers services necessary for a successful execution of ambitious data intelligence initiatives.
When multiple technologies come into play at the same time, it creates powerful and unprecedented synergies.