Fog Creek Software
Discussion Board




Performance related pay and buying lawnmowers.

Joel refuses to add reports to Fogbugz that can be used to monitor individual performance.

While out shopping for a lawn mower, I think I saw some evidence that shows his right.
   
Here in the UK we have a chain of stores called Argos.  They're a catalogue warehouse where you choose your items from the catalogue, pay for them at the till and then collect them from the warehouse.  I'm sure that every country has something like it.

One of the big problems is the wait at the warehouse.  Originally you would have to wait around by the desk keeping an eye out for your goods, and then try to attract one of the server’s so they could hand over the goods.  It would be a stressful half hour and the biggest drawback of this style of shop.
   
Then the clever people at Argos invested in a new computerised system.  The warehouse collection section is full of comfy chairs and screens that show the progress of the various orders.  You sit in a comfy chair, browsing the catalogue for the next thing you want to buy.  Any time you can glance up and see how your order is doing.  Finally a synthetic order declares 'Order XXX is ready for collection.'  Then you wonder up to the desk and take your items.  It all worked beautifully and I was very impressed.
   
Last night I visited Argos and went through the usual routine.  I was very pleased when my order raced across the screen and my order number was called out within a couple of minutes.

Then I spend half an hour waiting around the desk keeping an eye out for my lawn mower.  Everything was back to exactly how it was before the system!
   
I can think of only one reason for this degradation - management are using the system to monitor the speed of order handling, and most likely setting targets.  I'll bet there are bonuses involved.
   
So now the staff are motivated to hit the buttons as soon as possible, certainly before the order has actually been picked.  Head office is probably loving the 250% improvement over the last few months.

The reality is that we customers are back to the stressful warehouse scrum and the money invested in the lovely warehouse system was wasted.

Ged Byrne
Tuesday, April 01, 2003

>> comfy chairs

Not in the ones I go into. 8-)

Mike Grace
Tuesday, April 01, 2003

Ged, I hope you don't mind, but I linked to your post from my blog, since it was dead on target for today:
http://www.saintchad.org/blog/

Philo

Philo
Tuesday, April 01, 2003

Philo,

So thats whats happened to the CAMEL project.  I was starting to miss it.

Ged Byrne
Tuesday, April 01, 2003

No you're mistaken, Ged. It has nothing to do with performance monitoring.

Apparently the job printer in the warehouse is broken just before and just after your order goes through, with the result that orders XXX-1 and XXX+1 are ready for collection, but yours is not.

Justin
Tuesday, April 01, 2003

Where's the RSS feed Philo? :)

Andy
Tuesday, April 01, 2003

I'll be commenting to the same effect in another thread.

Although I agree with you in general, I believe that another reason could simply be that the computer system doesn't gel with reality. Whatever triggers the computer to announce your lawnmower has been ordered it is not the same as the lawnmower actually being physically moved.

The guy who wrote the computer system doesn't work in the warehouse, and nobody thought to sit around for a few days to see how it worked in real life.

It's where you get the thermoclyne; all those swimming above the thermoclyne thnk the whole thing is hunky-dory, while all those swimming below know it's one big balls-up but their waters never mingle.

Stephen Jones
Tuesday, April 01, 2003

Ged implied that the system, when first implemented, worked perfectly. So it doesn't seem like it's the programmer's fault/lack of understanding.

Martha
Tuesday, April 01, 2003

---" the system, when first implemented, worked perfectly. So it doesn't seem like it's the programmer's fault/lack of understanding. "---

The words of a true programmer Martha, and the reason why forty years of software investment has yet to have any discernible effect on productivity.

The fact that something works the first time means nothing. The programme is not supposed to work the first time, it is supposed to work all the time.

If there's a physical hold-up after the program has been told all is dandy, then the program is badly designed, since it's supposed to mirror the physical process.

Your attitude is the well known response to a trouble shooting request: "Well it worked fine on my machine".

This might not be the case here, but it all too often is.

Stephen Jones
Tuesday, April 01, 2003

" The fact that something works the first time means nothing. The programme is not supposed to work the first time, it is supposed to work all the time. ""

Bingo! That was well said.

OTOH, I find that when things work well at the beginning and degrade once everyone is used to it, it usually means that they've fallen back on old procedures rather than relying on the shiny new ones, many of which have less to do with new software than refined processes.

Ron Porter
Tuesday, April 01, 2003

I re-read this thread, trying to see where you're coming from, but I fail to see it. The programmer was told to implement a system which did X. He did. People found a way to use the system to do Y instead. Y is bad. This doesn't imply that the programmer failed to understand the requirements of X; it means he failed to foresee Y. But preventing Y was not in his requirements, at least not as far as I can tell.

The point here is, I think, that going back and preventing Y at this point is an exercize in futility; the users will just find workaround Z instead. The problem is in the basic system, not the programming used to model it.

Martha
Tuesday, April 01, 2003

The problem Martha is that you fail to see it.

The point is that a program is supposed to solve a problem in the real world. If the programmer writes a program that fulfills the documented requirements he has done his job on paper, but if the program then does something different in the real world, it is time to go back to the drawing board, and revise it.

Looking for blame is a waste of time; it doesn't matter who is to blame, the point is that the program doesn't fulfill its purpose, and therefore it needs doing again or revised.

If you worked in a proper engineering job and designe a plane that fulfilled all specs but didn't fly when it was windy or cold, or raining or night, then you wouldn't pretend you had done your job because the documented requirements had been fulfilled.

Stephen Jones
Wednesday, April 02, 2003

Stephen,

Were not just talking about a computer system here.  It's a wider system used throughout the warehouse.  The computer element works fine.  The screens give a good graphical representation of where the order is.  The problem is where the system requires input from the users (order pickers within the warehouse)

Attempting to create such as system that is 'dishonesty' proof is a mistake.  I have personally worked with systems that treated the user as a idiot thief and I did not like it. 

Systems should be supportive of the skilled worker, rather than idiot proof.

It is as important to get the human element to the system right as well as the software element.  It is a typical programmer error to think that all problems can be solved through the code or that the program should rule as master rather than serve as slave.

If a system is reliant upon the honesty and co-operation of the users, then you should make sure that honesty is the best policy.  This is down to the management, not the developers.

Ged Byrne
Wednesday, April 02, 2003

Stephen,

To take your plan analogy - once a plane is built it needs  regular maintance.  If management decide to increase profits buying halving the maintenance staff, is the engineer to blame when the plane falls apart?

Ged Byrne
Wednesday, April 02, 2003

You and I are in agreement here Ged.

The problem, however, lies in this split between the management and the workers. A computer system doesn't work fine just because everything  shows on the screen correctly; it must interface with reality.

Now what happens in these cases is that MIS insists that it has done its job because the system does what they were told to get it to do, and general management insists that they are following the system MIS devised.

You may be right that the problem is that the computer system is the evidence management uses to track its workers, and these have learnt that if they enter data correctly they will be held to account, but it could also be that the system asks for data to be entered just before an action is done instead of just after. It could also be that this matter hasn't been laid down.

Either way MIS acting as guinea pigs and sitting in the lounges waiting would have bought out the faults in the system.

Stephen Jones
Wednesday, April 02, 2003

You may be right on this particular case Ged, but to follow your analogy about the plane, if the engineers designed a plane without taking into account the maintenance requirements then yes they would be partially to blame.

Barcelona airport has a fabulous design, but the cost of cleaning the glass is exorbitant, and the floor tiles are in fact wall tiles, and the replacement costs there are another drain on resources. Yes, the architect was partly to blame (the exonerating circumstances being the hurry to get the plan ready for the 1992 Olympics)

Stephen Jones
Wednesday, April 02, 2003

Stephen,

I'm glad we are in agreement.  You make a fair point.

As an external observer, I can't know for sure what the internal workings of the system are.  However, having lived under the tyranny of management statistics in a number of helpdesk call centres I'm pretty sure that I'm right.

Ged Byrne
Wednesday, April 02, 2003

"The guy who wrote the computer system doesn't work in the warehouse, and nobody thought to sit around for a few days to see how it worked in real life."

Unless you are confessing, you are guessing.

"It's where you get the thermoclyne; all those swimming above the thermoclyne thnk the whole thing is hunky-dory, while all those swimming below know it's one big balls-up but their waters never mingle."

You mean like the tree of monkeys? All the monkeys at the top see a bunch of monkey grinning up at them, while all the monkeys at the bottom see abunch of arseholes who are always shitting on them?


Wednesday, April 02, 2003

*  Recent Topics

*  Fog Creek Home