Estimation of software development projects
Posted November 8th, 2009 in Outsourcing IT by Wojtek GurgulRecently we have had (not the first time) a pleasure to host professor Magne Jorgensen from Norway. It’s been the third or fourth time when he did some of his studies using PGS Software as an example of Poland-based software development company.
For more information of professor’s work please look here. Apart from his studies, we got a great seminar about typical problems with estimating software development projects. I won’t jump into details today, but I wanted to say something different. This is a fantastic and immediately ready to use piece of science which any software company should use as well as their clients.
Each of companies like PGS Software struggles with two things: preparing right (and realistic!!) estimation of the project. The second problem is how to convince the client that this is a real one. In the real world the typical scenario looks as follows:
The client asks for quotation for particular project several potential vendors. They all give quotes. Based on Simula’s studies all (almost all?) of those estimations are too low!!! The client chooses either the cheapest vendor, or one of the cheapest based on other criteria. What has to happen is that in the course of the project (or even before it starts), the vendor has to realise that the budget is too low. It can choose to cut corners on quality hoping that the project will be accepted, or reduce its margin and provide more work for the same money, or the combination of both. There is no other way. Does the client benefit from the lower estimation? Partially yes, but mostly no. The client will get the product of lower quality than expected and will need another budget in the future to fix it. Also, it risks reputation in front of its clients, because of the poorer service. So who wins? Nobody. Vendor did not make enough money, the client got poor product. If the client is not familiar with technology well enough to check estimations, which is usually a case, it has a difficult dilemma: which vendor to choose if the one with the lowest estimation is the cheapest, but probably is also the biggest liar (or the least professional) from all the others. I don’t have the answer here other than the price cannot be the only driver. I’ll get some of professor’s advises together with PGS Software experience in the next posts.
