RFID for bartenders

Sorry I have been away for so long, but I was waiting for something good to discuss. Boy did I find it.

There is a new RFID device that attaches to liquor bottles. By measuring the angle and duration of each pour, it measures how much was added to a drink. It is connected to the cash register, so it actually knows which drinks are being poured too. It can determine if bartenders are under or over pouring and it even has a warning message for consistent errors. The warning is just a beep, so they need to work on the human factors of this. But talk about instant feedback!!!!

There is some good math behind the device for those of you in doubt. It knows what the shape and size of the bottle is, so it can change its predictions based on how much has been poured from the bottle so far. And when the bottle is empty, it can recalibrate its estimates to increase accuracy even more - although this info can only be used for long term modeling of bartenders, not the immediate feedback (unless the whole bottle is poured at once).

From a HF point of view, this is cool in terms of instant feedback. To learn, telling bartender once a month that he/she overpours doesn't help. But telling him/her that he just overpoured the Creme de Menthe bottle is much more useful, because he/she can adjust the way he/she pours that particular bottle shape.
From an IE point of view, this is cool in terms of real time performance measurement and feedback - with the goal of improvement. It also adds some security because the bartender can't pour free drinks for friends (unless there is a button for that on the cash register).

I would have never thought of this one - what will they come up with next!!!

net neutrality

This topic is not centered on Human Factors, but it is something that is critically important to the development of Internet innovation so it will affect most of us in many ways.

For those of you not familiar with the basic premise, here are the two sides:

Telecoms: The fairest and most efficient way to fund the development of bandwidth is for companies that use more bandwidth to pay more for that use. So if your site has high-res graphics, video, audio, or simply lots of visitors, you will pay more for their access. A text only site with three visitors will pay the minimum. This is the basis of the free market system. And anyway, they own the bandwidth, so even if you think this is a bad idea, it is their private property on which to make such mistakes.

Internet companies: The greatest innovations of the Internet age have come about because there are minimal start-up costs to create the next greatest thing. I you think of the Web 2.0 services and other new companies (MySpace, YouTube, Skype, even Google), they were created by college students in their dorm rooms. If they had to pay for every bit, they could not have afforded to get started. And even established companies would be hesitant to create new services (think online banking) if the bandwidth used would not be covered by the extra revenue. If they charge for the new services, no one will try them out and great services will be killed at the outset.

Social organizations: It is very rare for MoveOn.org and the Christian Coalition to agree on anything, but here is one. If they (and other non-profits) had to pay for every visitor who clicks on their sites, it would kill the circulation of their ideas. Independent bloggers (especially Vloggers) would disappear. Local governments would not be able to afford to support their services. Charities would lose their online fundraising capabilities. So we need net neutrality.

And most people realize that government often screws up regulation of technology. They move too slowly to regulate something that moves so fast. Even if the regulation was good to start, technology would change and it would take 10 years for the government to update the regulation. And the regulations they do enact are generally flawed because of the influence of lobbyists, special interests, and a general ignorance of technology by the average politician.

So you are probably asking whether I am in favor of government regulation of net neutrality. To be honest, I am torn. I wish there was a free market solution. It is possible that competition in the telecom industry would create net neutrality because anyone who tried to charge more would lose business. But I don't think that there is enough competition at present to do this. So some government regulation is necessary. But how to do it, I don't know.

how much does Human Factors contribute

I was watching an interview of Alvin Toffler (the author of Revolutionary Wealth) and he said something that really struck me. His example of the problem with measuring national GDP was that when Microsoft spends millions creating a new version of Windows that is only marginally better than the old one, the measure of GDP goes up based on the money spent on it. But when thousands of programmers contribute to the development of Linux for free and thousands of companies download and use the software, again for free, there is no recognized effect on GDP.

I wondered about Human Factors in this same way. For the past few years we have been struggling to make the business case for Human Factors by proving that it increases sales, etc. But are there other effects that we are missing? I think that a usable product, even if it doesn't make me more productive at work, still makes me happier, and maybe more productive overall because of my mood. Or maybe just the happiness should be measured.

Overconfident experts

There has been a lot of writings lately about how experts are not nearly as expert as they think (Arnold Kling, James Surowiecki). The Arnold Kling article specifically discusses economists and how this overconfidence can harm public policy when governments rely on their recommendations. This is a considerable problem.

Since Human Factors has so much to say on naturalistic decision making (of which economics is a prime example), such as the fantastic research on recognition primed decision making, I think it behooves us to think about applying some our work to the public policy domain. Yes we are helping the military, law enforcement, etc., but why not macroeconomic and industrial policy at the federal level. Not only can we have a huge effect on our nation(s), but we would also increase the visibility of our field.

The Japanero Room

Another example from my stay at the conference. I had a Board meeting in the Japanero room. The hotel meeting rooms were arranged alphabetically, so the Japanero room was easy to find, right between the room beginning with 'I' and the one beginning with 'K'. A colleague asked where we were meeting. I said "the 'J' room". He had no clue what I was talking about.

This illustrates a pretty good example of how an incorrect user model can lead to a poor design. I made an incorrect assumption of his domain knowledge. Because he was a Board member, I assumed he got the same tour I did, which explained the alphabetical design (or he would have figured it out on his own by then). I also made an incorrect assumption about the quality of my message. I also assumed that saying "the 'J' room" would be enough of a semantic link to the organization of the rooms that he would understand my meaning. Because the alphabetical organization was a salient part of my schema, the statement was enough to activate it. But that is no guarantee for other users.

salient design

As I checked into my hotel to attend the IIE conference last week, I was struck by their use of Human Factors in selecting a white bedspread. I am sure that most of you have heard the nightmare stories of how disgusting hotel bedspreads are and the variety of bodily fluids that they find in stains.

So how does a white bedspread help, and why do I consider it a Human Factors application?

1. If the bedspread is white, it provides a visibly salient cue that the bedspread is clean, unlike those stories.

2. They assume that this is important to people, which implies a domain knowledge that includes those stories.

3. There is a semantic link between a clean bedspread and a holistic impression that the hotel cares about cleanliness in general, and therefore even other linens, carpets, etc may be assumed to be clean as well.

Thus, the money spent on cleaning the bedspreads, rather than hiding the stains in colorful patterns, may be worth the investment, but only if it is made clear to the user - which can be done elegantly with a white bedspread. Much better than a sign above the bed declaring that they are clean, which would just call out attention to the kinds of stains that could be present elsewhere.

customer made

I have been reading the trendwatching newsletter for just a few months, but it has become a monthly religious experience for me. It takes maybe 10 minutes to read, but I find myself clicking on lots of the links to check them out first hand, so it is more like 15-20 minutes. Not bad for such a great tool. Basically, the site identifies an emerging trend and discusses it, citing many examples. It is hard to describe briefly, so go to the site and read one or two of them for yourself. You will see what I mean.

This month's trend is "Customer-Made", discussing all of the companies that have created some way for customers to participate intimately in the design process. The companies range from high tech to simple retail and all over the world. I plan to use this as a motivator to the students in my Human Factors class to pick their own design projects in the Fall, instead of giving them one myself. It is not a trend I didn't already know about, but reading all of the ways it is being implemented was very informative and really amazed me both at how innovative the participating customers are as well as how great this is for the companies that do it.

Inspiration v perspiration

Sometimes it amazes me how places to apply HF are conceived. I generally promote the scientific method. Wade out into the tides of the user population, observing and collecting ethnographic, naturalistic data that illustrates their true needs. Then quantify and prioritize the needs through surveys, stratification, and statistical analysis.

But the inspiration method works pretty well too. I was reading an article in Industrial Safety and Hygiene News and I had a great brainstorm (helped along by the author's idea). If workers with different skill sets have different color uniforms or at least parts of the uniform, it could be very helpful in situations were faces are not recognizable. This could be in smoky facilities, over far distances, when workers wear goggles or face masks, etc. The uniforms (or patches) could signify many things that would be useful. In an emergency, it would be fast and easy to recognize who has partcicular skills like hazmat or medical. People who are from other departments could be recognized, either for security or safety reasons. An application of HF that I had never thought of before materialized right before my eyes.

Of course I would verify the need and marketability of the idea before investing my life savings in a start-up venture. My quantitative bones couldn't handle the uncertainty otherwise.

Defaults as expert opinion

There was an article in psychological science this month investigating the reasons that when public and organizational policies have a default (such as organ donation defaulting to opt-out), people generally stick to the default. The research concluded that people assume that the default is a recommendation from the policy makers of the best course of action.

I wonder if there is a correlate in usability. Does the typical user, who notoriously leaves settings such as virus protection, operating system defaults, etc. at the default, also have this unconscious assumption that the software maker recommends the defaults? That would be a much different conclusion than it is simply laziness.

Successors in politics or business.

Listening to a story about the King of Nepal gave me a thought. When a new leader, either in politics (President or King) or business (CEO or even plant manager) follows a particularly loved or hated predecessor, there are a few things that seem to come up.

1. They often stake out policies that are more extreme than they would based just on their original position, in part to make it clear that they are "their own man" (or woman).

2. They also sometimes are less willing to respond to criticism or failure.

The first one makes sense from a Human Factors point of view. In order to differentiate yourself from your predecessor (and have any hope at being remembered in history), you have to make the distinction clear, or stakeholders' schema will lump you together with the predecessor. If he/she was loved, this lumping will make you more likely to be loved, but less likely to get any credit or for people to follow you as the new leader. So staking out a new direction makes sense. If he/she was hated, then the distinction is even more important.

But what about this second one? I have a theory. It is possible that the policies of the predecessor are strongly connected to "policies of other" because a) they are not your policies and b) you are trying to differentiate yourself from them anyway. Policies of critics are strongly connected to "policies of other" because a) they are not your policies and b) criticisms always get connected to negative affect. So it makes sense that the same response will occur.

Failures could work the same way. It makes sense to try to disassociate yourself from your failures by blaming them on external factors. This is a common decision making bias in HF research. So they get connected to "others" as well, even if they are your own fault.

So this could explain the King of Nepal's (and president Bush) intransigence in the face of failure.

What do you think?

Safer cell phone use while driving??

This article presents an interesting concept, but I am not so sure it would be as effective as they think. The basic idea is that your cell phone will chirp (the same sound as the pedestrian crosswalk warning) when you get within 100 meters of a traffic light.

The idea is that many drivers who are talking on their cell phones don't notice it turn red and hit pedestrians or stopped cars. It is based on GPS, so you don't need any fancy technology, just a GPS enabled phone. It is connected to a database of traffic light locations.

But this means it will chirp even for green lights. So my questions are:
1. Will drivers habituate to the sound and start ignoring it? This could happen either for drivers who pass traffic lights a lot - habituation based on frequency of red and green lights - or when drivers start to ignore chirps because they often warn about green lights - habituation based on proportion of false alarms. The Human Factors research on trust in automation could probably predict what will happen, but focused research is probably needed.

2. Will drivers get annoyed by the chirp and either turn it off or just ignore it for aesthetic reasons? After all, it would have to be salient enough to draw attention away from the conversation for it to work. So the annoyance is virtually guaranteed for a large proportion of the user group.

What do you think?

Long term and short term behavior

I read an interesting article today (to show you how widely my interests vary!!) on how watching violent media (TV, movies, comic books, video games, music) affects aggressive behaviors. The relevance to Human Factors is interesting.

What they found is that adults are affected more in the short term, presumably because they already have schemas on how to respond to aggressive feelings, which become activated through the media exposure. But kids have long term effects, presumably because the exposure causes the development of new schemas.

Just some speculation, but I wonder how this would work with training or exposure to a new work system or technology interface? It could have some implications for new v refresher training sessions and how to design and deliver them.

Description error strikes again. In my constant effort at metacognition, I noticed another good one this morning on my way to work.

I left a water bottle on the kitchen counter to remind myself to grab my lunch on the way to work. As I passed the kitchen on my way out, I saw the water bottle and remembered about the lunch. So I walked into the kitchen and noticed my multi-vitamin bottle. I had forgotten to take my vitamin that morning, so I grabbed one and swallowed it. As I walked out the door, I realized that I had forgotten my lunch again. This is a description error because the vision of my vitamin bottle caused my activation to go there and it dropped below threshold for my working memory of grabbing my lunch.

Amazing !!! The author of this article (http://tinyurl.com/zacvh) is shocked that some web sites that focus on clean design over cool aesthetics are preferred by users. Duh !!!! I hate to be critical, but I think it is pretty well accepted that usability is often more important than cool looks, except in artsy, trendy kinds of sites. I am not making the reverse mistake by thinking that aethetics are not important, but you need to have usability first. Just look at Google. Their interface is basically spartan, but it is the biggest, most used site around. I don't think it is the cool logo.

Human Factors and political partisanship.

The NYT has a great article today (www.nytimes.com/2006/01/24/science/24find.html?_r=2&pagewanted=print) that tells us something about confirmation bias. When people listened to political speech that they disagreed with, they rejected it. But this rejection was not based on logic, it came from the emotional center of their brain and excited their pleasure area. In other words, when we hear something that we disagree with, we don't even give it a chance. We just reject it outright. And even worse, we get pleasure out of rejecting it, so we are more likely to reject similar situations in the future.

This explains a lot to me about the popularity of reality TV, talk radio, and lots of other media that I always find ridiculously asinine. Even people who hate it, watch it so that they can reject it. So the ratings are high and it stays on.

Human Factors and wayfinding

On my way to Stanford Business School for a meeting last week, I got off the Caltrain at the Palo Alto station. It was a typical commuter rail station layout. There were tracks on two sides - running lengthwise past the station in both directions. On one of the ends was a bus station. On the other end was the parking lot. So to get to a street, you had to walk about 500 feet or so either across the parking lot or past the bus stop. With my Google Map in hand, I looked around for a sign to tell me either which street was in which direction or at least north or south. No luck.

One of the basic principles I use in my HF design is to minimize risk to the user. In this case, there is significant risk that increases the need for good signage.
1. There is time risk: it takes a long time to walk across the parking lot, only to discover it is the wrong way and you have to walk all the way back and then past the bus stop.
2. There is effort risk: all this walking is tough with two suitcases.
3. There is safety risk: my train arrived at midnight and there were some shady characters around.
4. There is independence risk (see previous post): In order to find out where to go, I had to approach a stranger and ask.
5. There is dignity risk (see previous post): I looked very foolish walking around the station for ten minutes trying to find a sign or at least a street name, all the while studying my Google Map.

Context is also important in this case. When designing a train station, I think it is obvious that many users will be tired from traveling (the Caltrain line connects to three airports as well). The trains run late, so many passengers will be doing this at night, and probably hungry (I know I was). And train stations at places like Stanford are even more sensitive because there are lots of people arriving for the first time (conferences, new students, visitors, etc.) who will be unfamiliar with the area.

Independence and Dignity in product design

I was reading a great ethnographic paper on how people use products and what happens when products don’t meet our needs. In one case, an elderly user had several clocks in one place. A digital clock was big enough to read, but she didn’t know how to set the alarm. An analog clock was not readable, but she could set the alarm. The things we will do to accomplish what we need!! Interestingly, she did not blame the product designers, but instead felt incompetent.

Another example was a short person who could not reach the cabinet over the stove. So she grabbed a broom and knocked the item she needed out of the cabinet. In the process, she knocked a few others out too, creating a fire hazard when the stove was on (or if she forgot to remove it).

The purpose of the paper was to highlight two very important user requirements that we often forget about:

Independence – the product should allow you to complete whatever tasks you need to without help. Referring to a manual or calling customer service is not an acceptable solution. Even asking a friend can be a problem, especially for demographics like elderly or children who are afraid of being considered incompetent.

Dignity – the product should not make you look foolish when completing whatever tasks you need to complete. Even when someone can accomplish a task quickly and accurately, the product fails when the user looks foolish.

How many of us even measure these when we test our products? How many of us have our own stories like the two I described above? I know I do !!

If it ain't broke, add more features.

I heard this quote last week at the HFES Conference. It is so typical, but I like the contrast with the other cliches. So many systems add more and more features for every new version/model/or whatever come out. Its like the Peter Principle. If the system is usable, more features are added. When it is finally not usable anymore, they stop.

I heard another quote which was from a doctor during laparascopic surgery. He said "move it down, up the axis." What the heck does this mean? The other doctor didn't know either. The presentation was a study of communication during surgery. There were lots of other examples like this. Even simple statements like "move it to the left" are ambiguous because you can't tell if he meant his left, your left, or the patient's left. It makes me very afraid to get sick.

Isn’t networking great? When I got to the Human Factors Conference last Monday, I accomplished more in the first 24 hours than I usually do in a month. For example:

1. I bumped into a usability engineer I had met at a previous conference and got a lead on a consulting gig for his company.
2. I met an old colleague who is writing a book and we discussed me perhaps co-authoring it with him.
3. I met a contact from NASA who may be able to use my help on a project in January.
4. I bumped into a usability engineer who I have known for several years and mentioned that I would be in his city next month. He asked me to give a presentation to his group, which could possibly lead to a future collaboration.
5. I met a Canadian consultant who may be able to introduce me to some Canadian companies with expertise that I need.
6. In a hallway discussion with my grad school advisor, he gave me the contact information for a Korean lab that is working on something I am also trying to do. I never even thought to ask him.
   

I also helped out a few people:

1. I found out a student chapter needed help creating some marketing materials and I offered to help.
2. I asked a colleague to apply for a job at my institution, and he decided to apply.
3. I found a summer internship for one of my students.
4. I helped a student at another school decide on a grad school to attend.
   

And all this was just on the first day!! On the other hand, there are some things that go nowhere. At one lunch I attended, I spend 90 minutes talking to a guy about a potential project and he wouldn’t even give me the time of day. And who knows what will happen with the 10 items listed above. But it was worth the five days in Orlando if even one or two come through !!

search repositories

A have recently become interested in the concept of search repositories. The idea is that when we are searching, a lot of information is found and then lost. How many times have you skipped a link that didn't seem relevant only to go back later and try (unsuccessfully) to find it. This is also true of links that we find in content pages that we navigate through during the search. It can even be true of content that is not a link.

Wouldn't it be great if there was some unobtrusive repository somewhere that we could review when needed that would show an organized display of everything we have seen relevant to a search activity, even if it spans several sessions? Even better if we could use it to create reports, divide it into subsections for later use, conduct relevance analysis, data mine it for internal consistencies, share it with others, have it automatically updated and set up alerts when new content is added.

Anyone know of a product that does this? Or want to help me create one?