Tag Archives: estimates

I estimate you suck

This is yet another work-related post in a long series of work-related posts. Sorry, sometimes work just has to come out of me, usually in the form of vomit and/or poop.

The boss came to me a few weeks ago and said he wanted a company-only “wiki.” Yeah, just like that famous encyclopedic one. He explained it would be a good place for everyone on the team to document critical information. We’d all benefit by having searchable information at our fingertips.

Even I had to admit that sounded like a logical good idea, if everyone chipped it and actually used the tool effectively.

I should have smelled a rat.
Continue reading →

For the boss: I’m here to help

Pinhead bosses, perhaps intelligent in other aspects of their of their lives, become cripplingly stupid in the workplace. I theorize it is because the greed and power centers of their brain become so enlarged and aroused that they squeeze the thinking part right out.

Even the most mundane tasks, ones that you and I take for granted, can loom like a Death Star in that wide open galaxy of space they call a brain.

Boss: (whining) We’re not sell any of the THX-1138 widgets!
Employee: That’s because you never listed them for sale on the website.

It’s not their fault, really. They’re too busy doing important stuff like telling you what to do, like micromanaging how many rubber bands or pieces of tape you use. They’re only human and can only do so much!

That’s where I come in. I’ve decided to do the mature thing. I’m going to take the high road and forgive their foibles. I’m going to try to help.
Continue reading →

Who Wants To Be A Bouillonaire?

I have good news and bad news.

The good news is that I write my own material. The bad news, well, look above and see for yourself. 🙂

I wrote the joke above while thinking about odds and probabilities, especially in terms of events like the Indiana State Fair and the nuclear disaster at Fukushima Daiichi.

We often hear things expressed in probabilities. For example, “This luxurious home is offered at only $999,995. It is located on a 100-year flood plain.”

According to Wikipedia: “A one-hundred-year flood is calculated to be the level of flood water expected to be equaled or exceeded every 100 years on average.”

“A 100-year flood has approximately a 63.4% chance of occurring in any 100-year period, not a 100 percent chance of occurring.”

Interesting. I did not know that. I assume that such estimates are based on a lot on factors like recorded history and an evaluation of as many guessable variables as possible. But who knows how accurate such things are?

In the case of something like a nuclear reactor, I always wonder how they can estimate the probability of something for which no history yet exists. Truth be told, it sounds a lot like total theory and guesswork to me.

Is it the height of hubris to think we really know the odds of what might happen, what might be possible? And do we always tend to err on the side we favor? And if so, what is the cost of this bias?

Project estimating: You want it when?

My boss used to give me projects. Lots of them. That was his specialty. It was pretty much the entire basis of our relationship. Him telling me what he wanted all day long. Delicious.

As a web site programmer, I pretty much had to “invent” a solution for every single thing he asked. No two tasks were ever exactly the same. The process was simple. Imagineer a solution, plan it, then do it making any necessary adjustments along the way.

Of course, the boss wanted to know up front exactly how long it would take. And that is where estimating came in.

Bosses love estimates. They may not understand the magic of what you do or how to do it themselves, but estimates are something they can understand. Then they can make “management decisions” based on what their “gut” tells them. Again, this is done without the luxury of actually knowing how you do what you do.

Let’s take a look at a typical example:

Boss: I want the web site to email customers a different email depending on if they order product A, B, or C.

Programmer: That’s easy enough. We’ll replace the standard order confirmation email with one based on what they order.

B: Great. How long will that take?

P: Unknown. It depends on several things. You need to give me a detailed project specification sheet (AKA “spec”). What will the emails say? What happens if a customer orders two or more of those products?

That’s usually where the process breaks down because, of course, the boss isn’t going to give you any of those damn things. He just wants to know how long it will take. Asking follow-up questions regarding your understanding of the task just makes you a pain in the ass.

Then, once you do give him an estimate, comes the inevitable follow-up question: What day will it be done?

P: It’ll take about an hour.

B: So you’ll have it done by Friday then?

P: That I cannot say.

B: You just said it will take one hour! That’s plenty of time to have it done by Friday!

Sure, but that means absolutely nothing about what day it will be done. For some reason, bosses seem to have a real problem grasping such a simple concept.

It works like this. You, the boss, control my entire day. You pull me off task to produce reports, work on other tasks, answer phone calls, drive downtown to pick up your mail, work the retail counter, ship packages, and attend bullshit staff meetings where you expound on motivational topics like we can all be fired, you make no money, and how you browbeat a poor elderly couple who run a hotel into accepting half their normal rate because times are hard. (True story.)

The math is simple and works like this: A one hour project (assuming that’s a valid estimate) will take three months if I’m given five minutes a week to work on it. Actually, probably longer, since there is overhead associated with switching on and off tasks so much.

A good analogy for illustrating “task switching overhead” is painting a wall. You have to get the color of paint needed, open it, set up your equipment and grab a step ladder.

Which method of painting do you think is more efficient?

  1. Paint an entire wall one color then move on to the next.
  2. Paint 5% of a wall one color, then switch to another wall in a different color, paint 5%, then return to the original wall. Repeat ad infinitum. (Add about 10 minutes of “switching” time each time you relocate.) Bingo! A one hour project now takes 10 hours.

For some reason, bosses always seem to think option #2 is the best one. Then, at the end of the project, they always demand to know, “What the hell took so long? What the hell is the problem?”

Indeed. The only problem was you, goddamn Bossholio.