Wednesday, December 24, 2008

Sony Ericsson cell phone

I just bought a new Sony Ericsson cell phone and my first impression is that the usability leaves a lot to be desired. And that's saying a lot considering my previous phone was 4 years old.

The problems are not learnability - everything is learnable. It just takes more steps to do everything. Accessing menus takes 1 or 2 more steps than the same things on my old Motorola. And getting out of a menu requires steps too. On my old phone, closing it did this automatically. And hanging up also requires a keypress when its on non-auto mode.

What a pain.

[update]

And there's more. There is no way to quickly put the phone on vibrate. You have to go four steps deep into the menu. Even putting it on silent requires opening the flip. This is often done under time pressure (one person's phone rings in the middle of a meeting and everyone else realizes they forgot). A basic tenet of HF is to create simple shortcuts for anything done under time pressure. Hello Sony? Anyone there?

Tuesday, December 23, 2008

The Human Factors of Bernie Madoff

The Bernie Madoff scandal presents a very interesting illustration of confirmation bias (for his early investors as well as for him). I am assuming some of what went through his mind to make the point. And some is based on rumors that spread around Wall St in the 1990s about him.

In the beginning, Madoff engaged in a perfectly legal (although ethically questionable) practice of front running trades. When any large investor makes a big trade, they force up the price of what they are buying because they are sucking up the available shares. Its simple supply and demand. If they are selling, the same thing happens to force down the price. Madoff basically paid floor traders to buy shares for him ahead of large purchases or after large sales. He made a few extra points on each trade. This way, he was able to make profits that were a few points above the market. This was legal throughout the 70s and 80s. His customers got used to the great returns and were rewarded for sticking with him.

But eventually, Arthur Levitt, chairman of the SEC, was able to get this practice severely limited (although there are still loopholes from what I understand). Madoff could no longer make the gains he used to. Imagine you go from making 12% a year to only 8%. This doesn’t happen in a straight line; some months he probably made 12% (annualized) and others 5% (annualized). But each month that he made 12% his confirmation bias could convince him that he was “back on track” and that the 5% months were exceptions. But rather than disappoint his investors, he fudged a little bit what he reported to them on their statements on the assumption that he would make it up later. As long as they didn’t all try to withdraw money at the same time, this would work (if he actually got back to 12%, which he was sure he could do). Each time he had a 12% month, he was sure he was “back” and convinced him to continue the “little bit of fudging” on the statements. Since he also owned the trading operations and had an accounting firm with only one employee dedicated to him, he probably was able to convince these few people too.

For his customers, there were two types. The naïve investors didn’t really know about the frontrunning or what the probability of a consistent 12% return really was. All they knew was that he delivered. Each month, their statement showed a 12% return. There was the evidence. Maybe they heard rumors that he was doing in unethically, but it was legal so why complain. And when frontrunning became illegal, they may never have realized anything had changed. As long as their statements had a 12% (annualized) increase each month, there was no reason to question anything.

For his professional customers, they knew how he was doing it, and were happy to get the returns. And when frontrunning was limited, they figured he was exploiting the loopholes or he had found another way to game the system. After all, their statements had 12% returns. Why question it? If he was willing to be unethical in the 1970s and 1980s, of course he would do the same in the 1990s and 2000s.

Some interesting evidence for this is based on the amount of money people are claiming to have lost with him. If you give me $1 and I promise to magically turn it into $5, but instead I put it in my pocket and run away, how much did I steal from you? If you say $5, congratulations, you are a Madoff customer. Their last statement had $5 on it, so they assume that’s how much they lost. But in fact, the $5 never existed. All he really stole from them was their original $1 investment, which he never really turned into anything. I saw a couple on TV last night who had been a Madoff investor for decades. If you add up their original investment and subtract what they have withdrawn over the decades, they actually withdrew more than they invested. But their last statement said $1.65 million. So they accused Madoff of stealing the $1.65 million.

Monday, December 01, 2008

The value of a new innovation

Amar Bhide asks in his new book (excerpt) how we estimate the value of new innovations. he describes a study that asked participants how much would they would pay to kiss the famous person (movie star, musician, athlete) of their choice? Try this for yourself. If is really hard to put a value on things that we have never done.

Part of this problem is directly related to Human Factors. What is the true need that the innovation is addressing? What is the (behavioral) cost of the current method(s) used to accomplish the goal. How much better is the innovation at doing it? How hard will it be to learn? How hard will it be to use? What side effects/interactions will there be?

There are some famous quotes that under- or over-estimate the value of new innovations. "Who would want a computer in their house?" "Subprime mortgages will allow millions of additional people to own homes." (at least until they are foreclosed on :-(.

This is one of the great challenges of HF. First, we need to understand the real needs that users have. Then we need to understand our designs and how they will be used in the real world. Unfortunately, this is so much tougher than it seems.