Skip to main content

Agile or Waterfall: What's Best for Your Software Project?

When HubSpot’s development team started working it was clear to them that they are going to work agile. It was equally clear to Raytheon’s air traffic control development team that they should use traditional methods like waterfall (and CMMI). And if Raytheon would have decided on agile I would never fly again. However, these projects are in two ends of the spectrum for many in the middle the answer isn’t obvious.

I recently went on a business trip to San-Jose California to help establish MIT’s product realization lab. I met with managers in Cisco software and E-bay and learned that this is an important issue. In fact, E-bay uses both methodologies depending on the project.

To come at the right conclusions development teams should consider organizational, product and industry characteristics. I developed this questionnaire with 8 questions. By filling it out and summing the points, one can arrive at the most suitable method for a given project.

Questionnaire

Rank each question in a scale of 1 to 4

Industry related questions (uncertainty, time to market)

1. How uncertain are costumer needs in your project?

1 being high uncertainty and 4 low uncertainties

2. How important is time to market?

1 being critical and 4 not critical

Product related questions (reliability, dependencies, maturity, and type)

3. How important is reliability (or safety) of the product?

1 being not critical and 4 critical

4. How would you describe the impact your project has on other projects/products?

1 being a standalone project and 4 being part of a cross-functional product

5. How would you describe the maturity of the product?

1 being new product (no legacy systems) and 4 being maintenance of mature product

6. How would you describe software type of work?

1 being not algorithmic / analytic intensive and 4 being algorithmic / analytic intensive

Organization related questions (disparity, culture)

7. How close is the development team?

1 being co-located and 4 located in multiple time-zones

8. How would you scale the organization’s culture?

1 being developer centric and 4 developers as back office culture

Score of 8-16 recommendation: Agile

Score of 24-32 recommendation: Waterfall, Spiral or another traditional method

Score of 17-23 recommendation: ?!?

Some tips on how to answer the questions. Consider a diverse set of projects and rate yourself in comparison to them. Examples: car brake control software needs more reliability and safety than help-desk software. Time to market is more critical for a startup with cash for 6 months than to a fortune 500 company.

Finally, a warning: this questionnaire was not validated by empiric research (yet).

Labels:

Comments

Post a Comment

Popular posts from this blog

SDM from up on the bridge

I am taking the first spring semester as a long distance student in MIT's System Design and Managment (SDM) program. The long distance option, which I am grateful for, poses some challenges. A month into the semester, I am still trying to improve the interaction with the classroom, faculty and fellow students and maintain a balance between school and work. Setting up a work station I am using my home office (/living room /dining room/ guest sofa) from which I connect to the ‘bridge’. I find it useful to work with two screens. One displays the class, the mike controls and my video (I want to see how I appear to people looking at me from the classroom). The other displays the presentation slides and notebook (OneNote). The webcam is positioned above the main screen and when I look at the other screen, I am not facing the webcam. Class participation The video and sound quality are good. Material is always available before hand on the web. So keeping up with the studies is eas...
3 comments
Read more

Frugal Innovation in Healthcare

I have been working on EVOCAM a low cost endoscopy system with my partners at Evolving Technologies. Using frugal innovation techniques we developed a light, portable endoscopy prototype for a fraction of the price of existing solutions. Recently Medicine for Humanity physicians used our system to operate in Uganda. To allow training to continue from afar, after volunteer physicians returned home, we are planning to add a tele-medical feature to our device. I joined the team after they already had a working prototype and focused on market research, creating a concise massage and preparing a parametric engineering evaluation that exemplifies how EVOTECH is positioned in the market place. Basically everything it takes to turn a good idea into a good business plan. I am currently working rendering another EVOCAM out of MIT’s hobby machine shop. Together we took EVOTECH into MIT’s $100K business plan semi-finals (still competing) and to a bunch of other competitions in MIT and out...
6 comments
Read more

10 Recommended Readings on Business and Policy

Piles of business and policy books are published every year. While almost any book has something worthwhile in it, only a small fraction are good. And for every profound idea there is a load of simplifications, unbiased assertions, after the fact reasoning and hype. When considering what to read I know what to avoid. For starters, I’d avoid anything with a title similar to this blog title – anything with a number in it: “365 days of success”, “8 new roles of money” (from the Kiyosaki who brought us rich dad poor dad) or “101 ways to turn your business around”. I also avoid biographies and autobiographies. They tend to present a simplified view where the fortunes of companies and nations are over attributed to one person. This view is misleading and even dangerous when people start believing in saviors and messiahs . Autobiographies have all the flaws of biographies and they are also biased and are too often just pure narcissism. Finally, I won’t read a book about a company written f...
3 comments
Read more