Can You Lower The Quality To Save On Costs?
How much would you pay for a bad heart-transplant? How much would you pay for a ticket onboard on a faulty airplane? Is 80% of the regular price okay? Is 60% okay? What would you pay for a television that is only slightly likely to blow up and kill you? I say this, because some people seem to get confused about the Traditional 'Quality Triangle'. The three corners of the quality triangle are: - Cost
- Time
- Complexity (or 'features')
You can only control two of these at once. If you specify the time it will take and the features needed, then you cannot control the cost. Some people seem to think that the third corner is 'Quality' (instead of complexity/features). This implies that if you are happy with lower quality you can save on time or cost. Not true. When quality is lower, the triangle snaps. Costs suddenly blow out, deadlines whoosh past, complex workarounds are needed. Your customers leave and you are out of work. Don't f*ck with quality. It will f*ck you.
'Farmer Jeb' on Thu, 26 May 2005 22:33:51 GMT, sez: Amen, Brother. Amen.
'Wildranger' on Mon, 13 Jun 2005 03:18:05 GMT, sez: Man, I TOTALLY agree with you on this one.
Ive been using the old phrase, "you can have my services Good, Fast, or Cheap, pick two", for years. In the software world, yea, you cannot have all three, period. All these outsourcers seem to be trying to make the case overseas with cheaper IT people, but its not panning out to be cheaper OR higher quality. So the theory still holds true, looks like.
But quality is wrapped into the equation, even though the "triangle snaps" as you say. Most project planners in IT still could care less and have few means to measure quality or complexity in IT, so dont care and keep getting burned project after project in the US.
The problem is business people, managers, and CIO's still dont know how to keep from undermining the work IT people do in the Western world (ie offshoring now replacing us), and no one Ive ever met in all the corporate boardrooms Ive been a part of, knows how to measure quality or ROI in something as simple as two web sites and two different web shops and the code they use in making those two products successful. They simply dont care to take the time to really evaluate IT beyond pretty little graphic pages and UML charts. In other words, IT is absolutely about sacrifies of quality over cheaper and faster work (ie offshoring Indians). But people are still choosing cheaper IT over high quality IT and keep getting burned and still cant figure that one out.....when will they learn? Only after more money is spent I guess. Offshoring is not turning out to be their salvation as they once though (read recent Deloitte and Diamond research)
In addition, complexity in software can be good and bad....more complexity means more confusion for the poor smucks who have to customize someone elses code to meet a business requirement thats changed. Thats bad. But complexity could mean more code to address contingency issues, like what happens when someone presses the wrong button in a program. If the code is too simple with few "catch" statements, it could return nastiness. More complex error checking, it fixes itself! So complexity is a good thing in that case. Simplified complexity....thats a whole other animal no one has figured out yet...not even Microsoft. - MS
'secretGeek' on Tue, 14 Jun 2005 10:15:25 GMT, sez: Hey 'Wildranger'
You don't seem to 'totally agree' with me on this one at all.
The phrase: "you can have my services Good, Fast, or Cheap, pick two" is //exactly// the cliche I am attacking.
I think you're really keen to rant about your own ideas without actually understanding mine at all.
cheers
lb
'DaveW' on Wed, 22 Jun 2005 16:48:51 GMT, sez: Your examples are loaded to prove your point: all of them involve life-critical safety issues. In domains where this is prevelant (medical equipment, fly-by-wire control SW, etc.), people are generally willing to pay more and take longer to get higher quality.
But what if the cost of failure is lower? Would you be willing to use a phone system that occasionally doesn't complete your call or gives you a crappy connection? My answer: sure, we all do that. If the connection doesn't go through or is crappy, you just hang up and redial. Now if that happened all the time, it might make the system unuseable. But at the level it happens in the US today, it's generally tolerable.
How much more per month would you be willing to pay to get a guarantee of 100% quality connections? $100/month? $1000/month? My answer: not much more than what I'm paying now - the current level of service is good enough for my needs. Your mileage may vary, but I bet there is some cost level at which you would refuse such "premium" service, even if it was available to you. And I would further assert that most software quality issues are closer to my phone system example than your exploding TVs.
'Sandro Magalhães Costa' on Mon, 27 Jun 2005 09:42:18 GMT, sez: Por que então a China está se tornando uma potencia mundial vendendo produtos de baixa qualidade?
'sg' on Mon, 27 Jun 2005 20:00:29 GMT, sez: Yeh - there definitely are some types of quality that you can lower to save on costs.
i'm still thinking about this.... hope i don't go mad like robert persig in 'zen and the art of motor cycle maintenance'...
'boardtc' on Wed, 29 Jun 2005 10:24:39 GMT, sez: i did some googling....do you have any references (urls, book) for this traditional quality trialgle? Ta.
'Moz' on Fri, 01 Jul 2005 05:00:19 GMT, sez: Quality is a risk measure, not part of the complete/fast/cheap triangle. Dropping the quality does gain you time, sort of, in the sense that you might get the project finished sooner. Of course, it might take 10x as long as it should, and you have no way to know until afterwards. Hence the risk...
'sg' on Fri, 01 Jul 2005 10:10:27 GMT, sez: Thanks Moz...
Yeh -- that's right. You *can* drop the quality at times and end up in front. But it's a risk. Sometimes -- it's a huge risk.
Gamblers love low quality solutions. //Consistently// succesful people avoid low quality solutions.
cheers
lb
'boardtc' on Mon, 04 Jul 2005 14:54:44 GMT, sez: Sure, i agree. Stop the feature creep or the quality goes down. They don't get anything they haven't paid for. Was just interested to know where you were pulling this Traditional 'Quality Triangle' from, is it your theory?
:-)
'sg' on Mon, 04 Jul 2005 20:04:56 GMT, sez: Hi boardtc --
sorry i didn't respond to your question the first time.
it's definitely not my theory, it's an oldie and a goodie. If you google for 'project management triangle' you should find some good resources.
'http://' on Tue, 05 Jul 2005 11:17:49 GMT, sez: Right on. Thanks. I found http://www.projectsmart.co.uk/print/introduction_to_project_management_print.html
which says "More recently, this has given way to a project management diamond, with time, cost, scope and quality the four vertices and customer expectations as a central theme. "
|