I believe I've read several times that cognitive psychologists have convincingly demonstrated that people would rather have AnAcceptableWayOfFailing
than a risky way of succeeding.
The experimental evidence so far the fact is that ED (now XP) has been rejected as too risky
for way too long when it's obvious to any sane person that doing things in smaller steps means much less real risk. Something wrong in our psyche/heuristics is a reasonable conjecture. --RichardDrake
I explain TheAlmightyThud
This topic shows up early in AgileSoftwareDevelopment
, and it adds a good deal to the discussion of why it is so difficult to convince people to adopt new development approaches. Clearly the psychological and statistical angles described on this page have something to do with it, but I think there is something deeper at work in managers' blind adherence to an acceptable (or possibly "accepted"?) way of failing, and that has to do with personal accountability. If I do everything that my corporate development methodology tells me to do and my project fails, I can always point to other factors that seem to cause the failure (business requirements changed, critical team turnover, vendor code that didn't deliver). But if I stand up and push hard for a new, better way of approaching development then I am more likely to be held personally responsible for the failure. So with AnAcceptableWayOfFailing
, I can take credit for success and fend off blame much more easily than if I adopt a novel approach.
There's a similar "irrationality" observable with economics. The theory is that $20 is $20, whether it's earned or spent or saved. Yet we seek out $20 discounts on small purchases (software games, say), but wouldn't bother seeking out a $20 discount on that spiffy new $2,000 monitor. We'd mow our own lawn to save paying $10 to some kid, yet we wouldn't mow our neighbor's lawn for $10. -- EricScheid
We'd drive through the rain to attend a show we've already bought tickets for, but wouldn't go buy those tickets if it was raining.
That's a rational economic decision, provided the value of attending the show in the rain is greater than zero, but less than the cost of the tickets.
Earned money doesn't spend as fast as found money
I think it has something to do with relative values.
Well, different people have different habits. I might pay $10 to someone to mow my lawn, if I had a lawn. When I was living in NewYorkCity I routinely paid strangers ~$5 to do my laundry. I wouldn't have paid $30, though.
, see any of innumerable articles at http://thedailywtf.com/
, starting with http://thedailywtf.com/Articles/What_Could_Possibly_Be_Worse_Than_Failure_0x3f_.aspx
''According to the Standish Group’s CHAOS Report on Project Success Rates, "84% of projects fail or are significantly challenged." This directly contradicts common sense, which dictates that most projects are successful to some degree and run into a few bumps along the road.
For better or for worse – actually, for worse – once the code makes it to production and the initial bugs are worked out, it’s time to celebrate! Everyone’s hard work paid off and, despite a few delays and weekend crunches, the project finally went live. And before there’s even time for a post-mortem review, it’s time for the next project. Memories of the bitter arguments over poor design and hacks fade with time, and the project goes down in most people’s mind as a success. That’s right; to those that initially developed it, The Customer-Friendly System was a success.
That last point bears repeating. The consultants behind The Customer-Friendly System – a system that programmatically utilizes Visio documents to map its workflow – actually considered it to be successful. This means that they will take the techniques they learned from this project, and apply them over and over again. After all, it was through lessons from their previous "successes" that the system was even conceived in the first place.'' - AlexPapadimoulis?
The Customer-Friendly System
I've never understood why it's so often assumed that "acting rationally" is the same thing as "maximising your expectation" (in the probabilistic sense of "average"). For sufficiently-often-repeated decisions, it's certainly true, but (1) lots of decisions aren't repeated or even repeatable, and (2) some of the ones that are involve very low-probability events that make a big contribution to the expectation, which is exactly the circumstance in which you need a lot
of repetition before maximising expectation is obviously right.
If you have any sort of insurance, you are doing something with negative expectation because you prefer to minimise risk. Seems fair enough to me. This may not be an entirely fair example because your utility function may be highly non-linear, so you may be minimising expected utility by having insurance. But then, perhaps the same thing applies to situations in which people prefer AnAcceptableWayOfFailing
to a risky way of succeeding. -- GarethMcCaughan
UK political analogy
"All political careers end in failure" -- Enoch Powell
I believe that wide acknowledgment of the same rule for software development projects would help us all a great deal. The time comes even in the best XP or ED project that the cost/benefit equation says: don't do further development. The fear of this very visible "inevitable failure" is what is holding us back in many situations.
Add to that the bald economics of WaterfallBudgeting
: get the commitment for something big, pocket a large part of the loot and have plenty of time to work on your excuses. "You changed your mind" almost always comes usefully to mind at this point. This is why we need the SoftwareManagementManifesto
to receive widespread backing. -- RichardDrake
I think I could use a little more clarification on the point you're making here. Are you saying that risk-aversion is a good way to avoid being fired? As in "Nobody ever got fired for choosing Microsoft" ?
- The computer is acting up again. No, actually it's your Microsoft software that's acting up; possibly because you broke it by tweaking some registry key.
- Your Linux thingy is broken again. Well that might be true. Are you going to allow time to fix it, or just curse it and install something else?
Acceptable ways of failing are:
- fail because what you did was obviously stupid but was what you were told to do.
- fail because what you did was obviously stupid but was what everyone else also did.
Unacceptable ways of succeeding are left as an exercise for the reader.
Do we need to define the term failure?
See also: WhatIsSuccess