Re:Think: Isn’t it time to move beyond ‘one size fits all” model of estimation?
One of the pillars of every IT department is the person-month Model of estimation. The complete processes and tools have developed with the time-based models. The person-month model has been extended to with additional layers to present savings to business or develop business case. The complete model is based on improving process efficiency.
Tools that profess project management or program management inherently devolve to person-month (time sheets based) models. At the estimation phase itself, we confine the managers with defined templates that are perfectly suited for person-month estimation, with defined checklists, drop-down boxes, and built-in formulae. Though these templates have evolved out of the need to ensure that no critical factors are missed/omitted during estimation, over time they have become so rigid that completely confines the thinking process of the manager. Many companies also have proximity keys that deactivate a login as soon as the user moves away from computer.
These models have an inherent assumption – that all programmers will have the same speed or productivity. Just as an example, let us say there are three programmers, the first one is smart and experienced and can do the work in 4 hours. The second one does it in 8 hours and the third one does it in 12 hours. Assuming they all follow the complete principles of software engineering and write efficient code. Using the above model, the first one must drag his time to show 8 hours, and the last person must put in extra hours to complete his tasks. This has led to unnecessary wastage of resources skills and invariably in many projects’ effort getting buffered.
To remove this aberration, about 20 years ago, it was a common discussion in all project meetings about using alternate models of estimation and management, like function point, feature point et al, and the work allocation happens based on the skills and speed of the programmers allocated to the project.
But today, we have commoditized programming such that we expect a person with certain years of experience to be at a certain level of expertise, instead of taking his actual programming skills into consideration. The discussion on alternate models have seized to exist for over a decade now. We have also written iron-clad processes and procedures to meet this “one size fits all” model. Even under risk modelling, alternate models are scorned upon by management. The complete edifice from estimation to project management and delivery are built to ensure that every manager, irrespective of his intuition and gut feeling, is forced to accept and implement only this model.
This rigid thinking is inhibiting new development projects and encouraging out of the box ideas. These rigid systems are ensuring unquestioning corporate obedience. To encourage non-linear, irrational thinking many of these rigid systems need to be replaced with flexible systems that encourage innovative thinking.
As we get time to rethink, isn’t it time for us to start preparing for the future that completely aligns us with the business goals, including our estimation model and the way we capture performance?