Hewlett: So you think you understand your users.

March 13, 2008

Phoenix Wang introduced the session on usability and the challenges given by Patrick Whitney of the Institute of Design in Chicago.

Patrick talked about the issues that commercial organisations worry about and some failures and successes: failures included an Apple cube computer, Newton, Microsfot Bob, Web Van – of these were actually driven by sound user research. And some successes that had come from individual ideas – iTunes, OXO Good grips, the Novo(?) keyboard that clicks through identities. Moving from push to pull – but actually we might be dealing with anonymous users. Knowing how to make something does not help us match to patterns of daily life. He set out 4 principles for helping user centred design.

1. Activity-centred research

Patrick described several approaches:


  • Product-centred research
  • Culture-centred research: look at market segments
  • Activity-centred research: focus on activities and the context


The sort of studies to carry out include – disposable camera studies [sounds like Bill Gaver’s cultural probe studies]. Good things to look for are work arounds and structured field-notes that then allow


Insights matrix to find clusters by sorting the insights matrix. Rather than hierarchical sorting. A database that links back to the image and cases then allows illustration and deeper analysis.


2. Desirable, possible, and viable.

He gave examples of spotted opportunities: iTunes, FaceBook, Pearsons.


3. Look for user patterns, and then build systems and platforms.


Look at the context in terms of dimensions Abstract -> Real, Analysis -> Synthesis. Studies can move from real to abstract amd then from an analysis to synthesised options, and then make fast implementation.


Patrick answered some questions about whether design led could work. His examples included Home Schooling that cried out for aggregation.

How do we know more through mining data? This is not design – but monitoring behaviour as you don’t know the context of people’s lives around this: and those lives are often messy and complex and needs qualitative data around it.


Designing for participation (Vijay) – we have hoped on adoption and adaption. Need to design without being minimalistic but still offer options. The answer is around platforms – Steve C commented to me that this is a content v delivery tension.


No way around spending a lot of quality time with the customer – and then look at the alternatives to implement v temptation to just ask “tell me the product you want”.


Will there be a “killer app” for learning? Answer he is looking at is to develop a platform to be Kid-centred v curriculum-centred, beyond formal, assuming content is free.

4. Prototype a bold future

Plant yourself in the future and work you way back (cf Michael Porter activity systems). He showed a bank with interconnected books: RFID chipped so taking them out, a giant plasma that links to Google Earth, dream drawers. “Only going to the dentist is worse than going to a bank”. The approach was to just build and then use


He offered to send a poster with more principles if we email him to request an IIT methods poster.



Qs: What makes somethings that Apple does a success and others a failure.

A: Apple accepts failure

Q: How to tell poor ideas from poor implementations

A: E.g. of webvan v home delivery from Tesco – link to physical

Q: How to take lessons into virtual world

A: Actually the two worlds come together quite often

Q: How to carry out the research in Maine where they are taking all the yr 7&8 kids and giving them devices to gather data in the field.

A: Start with cheapest most available, and investigate, go with them, experience what they do. Do software experiments with people: whenever the user asks for information the person acting as the software will fetch the information. E.g. free advice for visitors – set up a stall at attempt to answer all questions. Follow similar activities.

Q: Designing OERs – end users could be anywhere.

A: Get data from distributed users and collate. Look for extreme users (e.g. for laundry find those who use multiple products)

Q: Is there a protocol for involving non-expert users?

A: Patrick explained that there are papers released about his approach, though they have not released the software they have developed. In general their approach is to hardly ever do direct observation, but instead get information from people in the field.

Q: We are not challenging the dominant mode of teaching: the course. We should be challenging that if following the principles here.

A: Yes. What if we move away from Fordist idea that there are things that we should learn?

Q: Facebook

A: Just build it and put it out there. Think big, start small, move fast.

Q: We can’t just rely on geniuses. Christienson models.

A: True innovations start small.

Q: Around remixing – capturing observations. How to capture that? Focus is on virtual site.

A: Video enthnography is hard – giving people cameras is easy. Getting them back will tell surprising stories. If you take the pictures then will introduce bias.



2 Responses to “Hewlett: So you think you understand your users.”

  1. simonfj Says:

    What can I say? Very nice Patrick. If yo had a mouth (and ego) as big as Stephen Downes we would see some real nice stuff come out here. (I like big egos)

    Prototype a bold future. That’s nice, Sounds like a semi scientific approach to philosophy. You’ve seen the janetcollaborate prototype, yes? Just one thing as the web centric don’t seem to think that ICT (Info AND Comms Technology) means anything. Picking a pattern from half the elephant means we’re still left with (well) half an elephant. (Say) Data is all IP – that does include the real times comms (RTC) stuff, which most institutionlists don’t bother about because the taxpayer pays the bills.

    So would you take a squiz at the Accessgrid (AG) in a bit more detail. It really is an “ensemble of resources”. Could you give some thought of the Speed of EVOLUTION – of (say) msg through to a peta pixel AG session and recording and broadcast. It’s all about layering protocols and adding bandwidth between IP addresses, which logically will (appear to) be the same ones as the pretty Web pages. Skypecast is almost out of beta, so it’s probably time to look at how multicast networks make sense for academic communities. (unless you like travelling)

    And if you want to take the “set up a stall at attempt to answer all questions” approach, perhaps you might want to consider a collaboration between OCWC and Questionpoint. The reference desk for global libraries.

    God i wish you had a open forum at OCWC that was designed well and stocked with Lauras (one from each silo). We might be able to bring so many threads (smokestacks) together.

    Teacher! What does this mean?. Q: Is there a protocol for involving non-expert users. A: Software is not available as research, papers are. Hardly ever do direct observation. And the next Q&A

  2. Patrick Says:

    Yes this was good stuff.
    I agree that some of my Q&A notes are a bit cryptic. I will edit the one that you comment on – this was a question asking Patrick Whitney what he actually did to reach out and bring in information from non-expert users. His answer was that they had written quite a few papers about that (which were available) and developed some analysis software (which isn’t available as they have developed them for themselves) but the general rule was that there are few situations in which he finds that he can actually carry out observation of people using the systems. Instead then effort needs to be spent on getting people to give feedback.

    Hope that is clearer – and I suspect in tune with some of your thoughts, Simon!

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: