The research budgets are growing: In 2007 61.5 billion Euros were spent for research purposes in Germany alone – five years later, expenses have already risen to 79.4 billion Euros. Two thirds of that amount are invested in the business sector and one third is used for financing fundamental research conducted by universities in research projects.
Before joining TEAMWILLE I participated in a fundamental research project. Our goal as an interdisciplinary team was to expand complex and numerical approaches. Already at an early stage, experience as well as intuition were necessary to steer our research into the right direction and to invest our research budget feasibly. Moreover, our research team had to face new and unknown situations time and again as requirements for the project’s outcome could not be specified exactly at the beginning. Those only became clear little by little as the project progressed.
What are the characteristics of research projects?
The combination of these tasks mirrors the common description of research projects in general :
Research projects are enterprises limited in time during which either single researchers or teams of scientists try to obtain scientific results within an interdisciplinary context.
Research projects usually feature the following characteristics:
a) High thematic or technical complexity
b) Unclear requirements
c) Flexibility of goals
Which project management approach is best suited for research projects?
To answer this question we first have to decide between classical (e.g. Waterfall Model) and agile project management (e.g. Scrum).
A decision-making-matrix helps us to compare both approaches based on certain criteria (table) . These criteria are evaluated on the basis of their validity for research projects. In the case the answer fits the criteria for research projects, the box in question is marked red.
|Classical PM||Agile PM|
|Clearly defined project goals||Yes||No|
|Clearly defined project scope||Yes||No|
|Frequent changes in to be expected in the project||No||Yes|
|Importance of documentation||Yes||No|
|Self-organized project team||No||Yes|
|Clearly defined costs||Yes||No|
|Clearly defined project duration||Yes||No|
Table: decision-making-matrix for classical and agile project Management
In the last row of the decision matrix we compare the arithmetic sum of the number of rows marked in red both in classic and agile case. Since “5>2” it is clear that research projects show more characteristics for agile project management (5 characteristics) compared to the classic project management (2 characteristics). This is why we subsequently only focus on the agile method based on Scrum. Its base is further described in more detail in the following blog articles: „Scrum ist nicht die Lösung, die Lösung lauert überall“ and „Von Storypoints, scheinbar wirren Zahlen und dem Mut, die Dinge auch mal anders zu machen“
Is there a particular project management method for the chosen Approach?
Yes. A suitable agile method is based on the SCORE (SCrum fOr REsearch) method [4 – 6] and considers the following core tasks:
a) Counting starts with 0:
At the beginning of a research project it is recommended to invest a couple of weeks into an initial Sprint, also called Sprint “0”. During this Sprint, the following groundwork should be carried out:
- Capturing the research status of Science and Technology. What has been published in this research area so far?
- Defining the initial scope: Even if the whole image of the research project and an idea of the forthcoming results is still missing it should be stipulated, which steps must be initially implemented. To what basic tasks do these steps refer? These tasks are then finally prioritised and collected in an initial to-do list, i.e. in a plan of the forthcoming task in Sprint “0”.
- Selecting software and tools: What tools are needed within the course of the project? What tools are available? What tools need licences that must/should be bought?
b) Prioritising is essential:
Making priorities is a very important planning activity, both during the initial phase and within the course of the Sprint. Prioritising the different tasks should ideally take place together with the research project team during stipulated meetings. These meetings help the team to create a common and deep understanding of the project content and to identify grey areas in due time.
The cost-benefit ratio should be in the centre in order to prioritise new tasks and functions, so that tasks with the greatest benefits and the lowest costs are developed first. Since cost and benefits cannot be clearly calculated in a research project, these criteria are replaced by the ratio of scientific interest to complexity. So, for instance complex and less relevant tasks and functions can be prioritised at the bottom of the to-do list.
c) Ready, steady, Sprint:
After having created an initial to-do list, the tasks described on that list should be divided into short Sprints (e.g. biweekly). In the beginning of each Sprint, the tasks on the top of the list are chosen and fragmented into sub-tasks. These sub-tasks should be prioritised, too. At the end of each Sprint, an incremental work result can be delivered.
Mainly, research projects differ from other project types, since not all requirements are known in the initial phase. To assure project success, agile methods are the best choice, such as Score (adaption of Scrum for research projects). After having captured the status of technology in a Sprint “0”, the Scrum approach applies in full, including a prioritisation of the tasks from the to-do list and an iterative development within Sprints.