Fog Creek Software
Discussion Board




"Incentive Pay Considered Harmful" Revisited

I re-read Joel's insightful article on incentive pay (http://www.joelonsoftware.com/articles/fog0000000070.html), as well as Frank Hayes' article on how to avoid employee burn-out(http://www.computerworld.com/careertopics/careers/labor/story/0,10801,79619,00.html) and I was wondering what people's thoughts are on this issue in post-bubble days.

Take a company that has a well-established performance review program and pays a very reasonable salary and is a great place to work.

What if that company decided to pay a once-off incentive to a dev team to meet a specific objective, such as "cut the number of open issues to less than 10", or "keep a system running without failure for 3 straight weeks"? This incentive would be totally seperate from salary and performance review.

Does anyone have any experiences to share or any other references about incentive pay?

Astarte
Monday, March 31, 2003

I have experienced this before, everything you said was there; good pay, good working environment and good team of people. As a one-time deal I was offered extra pay for not scheduling my vaccation time until after we finished the project we were working on.

In my book, this is not harmful or bad, because you know its a one time deal, and you can "see the end of it" in lack of a better wording.

Whats considered harmful in my book is ongoing demanding incentive programs; they usually end up with people backstabbing eachother instead of helping eachother as a team.

Patrik
Monday, March 31, 2003

Two experiences come to mind where this has worked and failed.  One case where it failed, you mentioned.  We have a bug/enhancement list of about 300 items.  Someone thought that if we incented people we could cut the list in half.  As a technical leader, I was in on some of the meetings and here is what happened...
  - Incentives for every bug closed.  The GUI presentation people loved this idea.  The A/R people hated it.  While GUI had more requests outstanding, the A/R peoples were far more time consuming to fix.
  - Incentives for every hour on a bug beyond 45 worked. Younger managers did this one in.  I would even say "inexperienced" managers.  Their question was "how would we know they are not taking a long time fixing a bug so they get more money?"  Trying to explain the depth of mistrust the very question posed, as well as the serious issue "it" provided, was more than most of us could swallow.

One area that did work was System Support.  It is very difficult to get experienced people to stay on System Support.  The hours stink, the benefits are zero, and in general as a senior person, you are needed during the day regardless of the hours you spent at night.  The fix:each member on System Support was given the week following their on-call rotation off.    This worked great.  We got experienced people back onto the rotation, support calls were resolved in less than half the time and permanent fixes were implemented at an astounding rate.  Then the complaints from people not in Systems Support started.  They did not want to be oncall but they did want the extra four weeks a year of vacation or more specifically if we were giving them to others they wanted them too.  (Time to differentiate between the leaders and the managers -- deal with it.)

Now for the obvious, but often overlooked.  Do not attempt to incent people with the same thing.  Joel was on target here.  Bob make want the cash, but Jane might want days off instead.  Offering Bob days off is NOT an incentive.

If you attempt to incent people to do their jobs you will lose.  If I am responsible for bug fixes and you incent me when over 100 are in the queue, nothing is really important until we hit 100.  We suddenly start looking at 100 being a pass/fail grade for Mike.  If Mike cannot keep the number below 100, maybe the solution is to bring in some help (assuming Mike is not the problem).  If you are going to bring in *gasp* temp help you should consider incentives for someone from another area within the organization. 

Mike Gamerland
Monday, March 31, 2003

I have seen this sort of thing work well in the past:

We won a large contract, and each member of the team working on it was evaulated, up front, regarding their responsibilities, scope and challenges by the heads of the project.

Then specific milestones with associated testable/quantifiable goals (not stuff like researched feasibility of...) were set.  As each milestone was met, the team would get a report card ranking how the goals were met - on time?  met or exceeded performance requirements?

At the end of the project, the bonus pool (total amount determined by the value of the contract and the report card scoring) was divided among the team according to their "importance" ranking.  The numbers were very significant.  My bonus cut was in the tens of thousands of dollars over my normal salary.  The project duration was 18 months.

Nat Ersoz
Monday, March 31, 2003

In my experience, there is no such thing as "one time" or "temporary" or "nothing to do with salary".  Either you agree with and implement a policy (such as incentives) or you don't.  Otherwise, that "one time" will just turn into the "first time".

Sounds like you don't understand/agree with the policy of not using incentive pay.  Just remember that people will adapt to their incentives, at the expense of the larger company goal, and ask yourself if that is what you want.

Scot
Monday, March 31, 2003

What I took out of Joel's article is that generally, most of HR culture's ingrained notions of employee motivation are patronizing and toxic to technology professionals who happen to be on payroll. 

My opinions and rambles:

Minor incentives and perks come off like Bill Lumbergh's "Hawaiin Shirt" day idea in "Office Space", and larger incentives create a Pavlovian training mentality.

Almost all technology jobs in IT today are fraudulently presented as career situations, so WTF cares if there's an incentive payment for a crunch? It's just another variation of treating employees like contractors. It's kind of saying that "you, the employee, probably know the score, and we really need this done now, so take this bribe and work as much as we need in this crunch."
Or, less cynically - the kind of company that would care enough about equitable treatment of employees to pay a "one time" incentive is probably a decent place to work and paradoxically, probably doesn't even need to fool with incentive pay.

The opposite end of the axis from this is the sweatshop that figures that all salaried employee's time is potential unrealized working hours - places like this need  and deserve a swift bankruptcy and repossession of the haughty owners' half a dozen luxury vehicles in order to restore equity and balance in the universe. But damned if a place with this business model in place will ever pay fair incentive pay. Typically, it amounts to a bonus of a dollar an hour or less - which impresses only the stupidest, and insults anyone who considers their time worth something.


In the semi socialistic closed system paradise of the well run geek company, nobody would resent pitching in extra; extra effort would be appreciated and respected, not expected; staff would look forward to long, stable careers, protected from outside business cycles; and just as management could expect employees  to chip in each according to their gifts and abilities - employees could expect to reap the rewards of trust and could expect to take off compensatory personal time in exchange for contributions made in crunch mode.


But I'm a contractor. The "closed system paradise" doesn't seem to exist which is why I am at where I am. And I don't fit in where there's a culture of  long term balanced symbiosis between workers and management. Which is generally, never - but most companies play pretend that they're "just like the old HP".

Bored Bystander
Monday, March 31, 2003

Nat - doesn't this encourage padding your estimates?

Philo

Philo
Monday, March 31, 2003

It is easy to focus on the technology professional (which most of us are), but these ideas, I believe, transcend that and apply to any intelligent worker.

A bonus should be a gift by the corporation not tied to any activity not directly related to the employee.  That is, "If we keep the number of bugs below 100 you all get dinner out." is meaningless because those bugs are not all equally valued by the client, the programmer, and the company.  Bug #1 might be so difficult and important that the other 99 might as well wait until its resolved - instead the incentive keeps programmers focusing on those other 99 less meaningful bugs while leaving the big important one exposed and undealt with.

Giving employees off or extra pay for dealing with difficult situations - such as the Systems Support personnel referenced above - is a reasonable measure taken to ensure employee satisfaction.  However a better measure would be to put this in as part of the employee contract - any employee who is required to work X job will be given an additional 4 weeks.  Employees may complain about inequality but they can't argue that it isn't published and they weren't aware.

Companies pay a price to keep good employees around - most seem to relegate themselves to paying it out in secret bonuses rather than giving everyone who deserves it a raise and raising the comfort level of all.

Lou
Monday, March 31, 2003

Rd: Estimates.

Well, that was an interesting thing.  There were actually 2 programs which had incentives like this.  And there were no opportunities for engineers to provide any estimates in terms of the milestone schedules.  They were purely driven by negotiation with the customers.  So, of course, the biz team tried to negotiate the best terms possible, but in the end it was utterly ridiculous.  Bottom up scheduling was ad-hoc, and only as particular subsystems showed risk.  We worked like dogs, and came in about 2 monts late (out of a 12 month schedule). 

On the second, it was much more reasonable (18 months as I recall) - I think the biz team negotiated much better terms.  Also, they were pushed into penalties for missing the final delivery date.  They didn't take it lying down, however, and insisted on penalties should the customer not take delivery.  Very shrewd.  We met their schedule, they failed to take delivery.  It was a billion dollar contract.  Penalties were $50 million.  There are some residual articles out there regarding it:

http://telephonyonline.com/ar/telecom_cold_shoulder_california/

"Of the 3 million digital wireless cable set-top boxes ordered from Thomson Consumer Electronics in 1995, only 250,000 have actually been delivered, with Tele-TV paying a substantial penalty for canceling the rest of the order, a source close to Tele-TV said."

You'd think that I'm violating someone's NDA, except I never signed one.  But I never signed one. The NDA we were given when Thomson bought the place was so absurd that anyone who knew better tossed it (almost everyone I knew at the time took a look and round filed it).  And TeleTV is long defunct (although I still have 2 shirts with their logo).

Nat Ersoz
Monday, March 31, 2003

That's interesting. One of my targets, against which my bonus will be based, for the coming year is how far under/over estimate my stuff comes in. Guess who gets to do the estimates? "Yes, its defintitely going to take a week for me to change that string in the string table" (because I will goof off for two days, then fix it in 5 minutes and still come in ahead of estimate). This is a fucked up world.


Tuesday, April 01, 2003

Thanks, Bored. That post has given me some cheer this morning as I realise I am not alone in a cold uncaring universe.

I have contracted off and on (mainly off, though) for a number of years, the last stint of which started 2 years ago following redundancy. I tend to contract for a bit, then get lured back into perm work for a few years with empty promises and lies. They all seem plausible at the time though.

Almost everywhere I have worked has turned out to be at the lower end of your axis, owned or run by people with the business ethics of weasels ('axis of weasel'?).

I don't have a problem with a company making vast sums of money on the back of my labour; that's what companies do. They exist to generate as much profit as possible for their shareholders. Independent contractors are no different, or should be. Indeed a common criticism of us is that we run our businesses like the enthusiasts we are, not as business people.

Having said that of course, I realise its not completely true. I would like some of the profit to filter its way down to me. What grates is working every hour God sends for smug bastards who secretly laugh at being able to con non-business savvy people (ie anyone who actually cares about doing a good job) into working stupidly long hours for no tangible benefit.

It’s the lies, greed and general blackheartedness that are so unbearable.

Justin
Tuesday, April 01, 2003

What I am concerned about is the old saw "you get what you pay for". I am dead scared that this approach will discourage the pleasant, cooperative environment we currently have, and encourage an adversarial (to management and colleagues), competitive environment. I think inexperienced developers will burn themselves out desperately trying to jump through the hoop we're holding up. The whole idea seems too Pavlovian to me, but I thought maybe I am just being naive.

I feel like this would not motivate me. But maybe I am not realistic about what motivates other people?

Astarte
Tuesday, April 01, 2003


"It’s the lies, greed and general blackheartedness that are so unbearable. "

I have to agree with you Justin.  The overall waste and overhead costs in most companies/projects is extremely distasteful.

In a lot of ways, I see incentive pay as just another disguise for MBO... it works just as well in causing dysfunctional behavior.

Joe AA
Tuesday, April 01, 2003

Astarte:

I'd suggested you read "Punished By Rewards: The problem with A's, Praise, Gold Stars and Incentive Programs" by Alfie Kohn.

Failing that, check out www.alfieKohn.org or my fourth article on "leading the way":

http://www.cadetstuff.org/features/200112/200112_huesser_styles_4.htm

regards,

Matt H.
Tuesday, April 01, 2003

The problem with "one-off" incentive schemes is that they are only "one-offs" if they don't work.

If they do, then the temptation to say, 'Well, this worked great last time, so let's try it again' is just too great.

Incentive schemes cause "local optimization".  If the particular goal you have is more important than anything else in the company fair enough.

Also you should be aware that incentive schemes, or simply stressing certain aspects, can have the most peculiar effects.

Our local steel factory decided it was going to have a crackdown on safety. It always posts outside its gates the number of man-days without accident and decided it was going to try and see if it could attain five million. It did, but what happened on the way wasn't quite what you could expect.

If you had an accident at work and went to the doctor or nurse, then they would ask you if it happened at work. If it did, then there would be a full investigation (part of the safety drive) and you could pretty well guarantee that neither your supervisor nor higher management were going to be the ones responsible, even though they probably were by putting the emplhasis on production instead of safety. So, to avoid possible disciplinary action, you announced the accident happened outside work, and that, yes all those bruises and concusssion happened as you staggered out of the bar in Bahrain drunk last night and you had only come in to work semi-comatose for love of the company. You were then given a forthnight's sick leave, no questions asked.

This continued for months; people continued being injured, no report was made - which meant that the blackspots were not identifieid - and the banner outside the factory gates crept closer to the five million.

It all fell apart when they had a couple of deaths within days of each other. Both spectacular. In the first one a couple of tons of iron ore crashed down on a father-of-four's head when the frayed cable went loose.

The banners were taken away, and they started to report accidents as they should have done. The management wriggled out of the debacle by claiming that what was needed was extra training, and I got called on to give suggestions, since I had helped devise the last safety course. The fact that they knew full well all of the safety rules but didn't apply them because "in the factory it's quite different from college" was ignored.

Stephen Jones
Tuesday, April 01, 2003

Matt,

Alfred Kohn is a real piece of work. Radical communist doesn't even begin to describe him. In my mind, nothing he says should be taken seriously, but to each his own. You should check out his radically anti-american screeds. Some of them he links to from his website.

Tony Chang
Tuesday, April 01, 2003

I checked out his website but couldn't find any such links.  Perhaps you could indicate how we might find them.

Kyralessa
Tuesday, April 01, 2003

http://www.rethinkingschools.org/special_reports/sept11/16_02/kohn162.shtml

He argues that we need to teach our children that the motivation of the 911 terrorists was understandable, since the US has created and funded armies of terrorists, killed 100,000 iraqi civilians in 1991 alone, and some of the 911 terrorists suffered -personally- at the hands of US imperialist agression. Oh and also he argues we need to teach our children that 911 happened because the US supports "brutal" Jews. Just the  standard deranged antiamerican screed.

In his other articles he argues that it is basically abusive to praise children.

B: "Daddy, look at the picture I drew."
D: "That's a great picture, Bobby!" <-- this is the wrong thing to say according to Alfie.

Anyone other than Matt want to argue his position about praising people for a job well done? Or paying people for competence? Kohn says that is bad too. To each according to his needs after all.

What's it all about with Kohn I just don't know...

Tony Chang
Tuesday, April 01, 2003

----"Oh and also he argues we need to teach our children that 911 happened because the US supports "brutal" Jews"-----

Of course it didn't. The hijackers all decided to commit suicide because they had an irrational aversion to Ronald MacDonald and democracy

Stephen Jones
Wednesday, April 02, 2003

Care to clarify here Stephen? Your McDonald's comment sounds like you intend to be sarcastic. Does that mean that you also agree with Kohn's suggestion, or is your sarcasm nonsequitorial?

Tony Chang
Wednesday, April 02, 2003

Everybody who is not in denial or with a hidden agenda, knows that the US's unfailing support for Israel under all circiumstances is the prime reason fot its being hated throughout the Arab world.

No doubt Bin Laden himself would be anti-American even if the US turned aganst Israeli expansionism, but he wouldn't find it anything like as easy to get supporters if US foreign policy were not so obviously subservient to Israel.

Should we teach the truth to schoolchildren? a dangerous departure no doubt. It might be an idea ot start teaching it to adults though.

Stephen Jones
Wednesday, April 02, 2003

So what are you saying Stephen? Are you saying that the Arab nations are all racist antisemitic bigots or what exactly?

Tony Chang
Wednesday, April 02, 2003

Not in the least. The Arabs are quite right to be against American and Israeli foreign policy. To brand any criticism of Israel as anti-semitic is a cheap trick used by those such as Wolfewitz who want to stifle any kind of reasoned debate.

To dislike a Jewish state demolishing your houses, evicting you from your land and killing your children isn't anti-semitic, it's a normal human reaction.

Stephen Jones
Thursday, April 03, 2003

Have to say I found the Kohn article linked to made a lot of sense, although a lot of his other stuff didn't.

David Clayworth
Friday, April 04, 2003

*  Recent Topics

*  Fog Creek Home