There is a school of thought within usability that asserts that during
facilitation, moderators should not speak to the participant as this interaction affects behaviour, and so invalidates the research.

It’s similar to the idea in ethnography that the very presence of an observer will lead to modifications and unnatural behaviour.

There is also the idea that people may not have conscious access to the real reason for their behaviour. In trying to explain their actions to the moderator they will introspect and provide an answer that they feel is rational, but is effectively made up.

Malcolm Gladwell sums this up something like this (I’m paraphrasing here): Basically—we feel about a thing, then act.

And then, the moderator asks us why.

Faced with this question, we try to think up a plausible, rational-sounding explanation for our actions. And then—here’s the thing—we alter our future behaviour to match that rationalised thinking.

Wilson and Schooler investigated this phenomenon in depth, concluding: “We come up with a plausible-sounding reason for why we might like or dislike something, and then we adjust our true preference to be in line with that plausible-sounding reason.”

So if we had never been asked why we did something, we might continue doing things differently.

However, without entering into conversation with a participant we can only say what happened; and with no insight as to ‘why’ we can’t make decisions and can’t improve.

The problem then is that we need to know ‘why’ but can’t ask ‘why’.

Here’s a few methods that we use at Flow:

Sometimes more open interviewing will tell you what you need. If you need to know why a participant clicked on that link (or didn’t), questions such as “Tell me about the kinds of things you have looked for in the past on a site like this?” can tell you about the keywords or visual elements that a participant is searching for.

“What are you interested in finding out at the moment on a site like this?” can tell you what it is they haven’t found yet.

Ann Light continues in this vein: “An undesirable, but common, way of interrupting evocation [the flow of recall, in this instance] is to invite the interviewee into a judgemental mode. To avoid this, there is no use of questions starting ‘Why... ?’ Instead, carefully manipulated ‘How... ?’ and ‘What... ?’ questions cover the same ground: ‘How did you know that X?’ ‘What were you thinking at the moment when X?’ This does not interrupt the recounting process. So ‘tell me how it was that you came to be looking for this site that day’ does the work of ‘why were you looking... ?’”

Of course together with the different ways of asking why, the facilitator needs also to combine high degrees of empathy and observation. Interpreting what the participant does and says, and ultimately understanding those things will enable us to make better decisions about what to do next.

References

Wilson and Schooler (1991) Thinking too much: introspection can reduce the quality of preferences and decisions, Journal of Personality and Social Psychology 60 (2), pp181-192
Light, A (2006) Adding Method to Meaning: a technique for exploring
peoples’ experience with digital products, Behaviour & Information Technology 25 (2), pp175-187

Even as online retail struggles, you can keep sales figures alive by optimising the site you have. One area to look at is improving what happens to your customers when things go wrong.

A classic and expensive error

During my previous life as a developer, I lead the prototyping for enhancements to a clothing and home wares website. Like any sensible UCD practitioner, I wanted to get measurements to help me understand where customers were going on the current and revised site.

Just one error message

When it came to errors, the need for more detailed measurements was particularly urgent. We didn't know which pages or processes were throwing up errors. We had no detailed analytics and only one error page for the whole site: "An error has occurred. Press Continue."

I knew that we needed to write individual, polite messages which would reassure people enough to carry on, but we didn't even know yet what those errors were.

Once the analytics were up and running, we could quickly see which pages occured before and after the error page. This let us identify the user journeys where the errors were triggered. So we were able to write custom messages, things like "We're sorry, we've had a problem processing your order. Your card hasn't been charged yet. Please click checkout to try again." We also provided a customer care number together with a code for continuing the transaction offline.

Return on investment

Within a month the percentage of completed purchase journeys increased a modest 0.5%. Putting it in some perspective that 0.5% was worth £27,000 a month on average - or over £250,000 per year.

And all this was hypothesised, diagnosed and implemented over 2 weeks, at a staff cost of approximately £8,000. That's a great return on investment from paying attention to the detail of the customer experience. Each one of those customers had a chance to complete their purchase, and to remain a loyal customer with a chance of purchasing again in future.

6 error message tips

If I've inspired you to revist your site's error messages, remember these golden rules as you go:

• Be polite: Don't make your visitor feel like they've done something wrong; accept responsibility for the problem, say "sorry", "please" and "thank you".
• Use plain English: Remember that the people visiting your site aren't developers, and don't know Java from JavaScript or 404 from 500 (unless you're writing error messages for Slashdot or the MSDN that is).
• Reassure: Especially when the visitor's money is at stake, it's important to make sure that people understand that nothing catastophic has happened.
• Make it brief: Your visitor isn't interested in the exact whys and wherefores, be succinct.
• Help people recover: Provide them with a clear path out of the woods - even if that means calling someone on the phone.
• Tag them: Track where people go and what they do before and after experiencing an error on your site, and track what happens as you make changes.

Optimising is just one of three design-based strategies for beating an economic downturn:

• Innovate
• Optimise
• Cut costs by improving the customer experience

The web still has the capacity to delight and disappoint me in equal measure. Recently I experienced examples of both extremes on exactly the same day.

Extremely good

Threadless  sells t-shirts, and sells them well. Limited edition t-shirts, designed by anyone who wants to design them, and voted into production by the Threadless community.

On their site, as on many others, I often use the basket as an ongoing wish-list, collecting the stuff that I might buy if and when the conditions are right. But if you do this with Threadless' basket you stand a fair chance of missing one of those limited edition t-shirts. So Threadless have come up with a nice email to let you know when this is going to happen, and here it is:

The call to action is strong, the tone of voice cheeky and familiar but still polite ("thank you from your pals at Threadless.com”).

Extremely…. well, bad

That very same day I received a film and ink pack for my photo printer; I'd ordered them online having done a little research and ordered a pack of 100. However the pack that I got in the post was only a 50. Frustrated, I telephoned the supplier:

10 Man on phone: "What was the product code on the invoice?"
20 Me: "CO3548"
30 Man on the phone: "that's the 50 pack"
40 Me: "Well your website and the email confirmation says its the 100 pack"
50 Man on phone: "What was the product code on the invoice?"
60 GOTO 20
RUN

In the end the query was passed on to someone else and while waiting for the call back I thought I'd check my order online and log into my account. Perhaps I was flustered from the phone conversation, but I made a mistake with my password. This is what greeted me:

Okay so now I'm not just a little miffed but in fact somewhat teed off, let's just read this out loud together:

"Internet fraud is a serious offence..."
"we record IP addresses to help trace the location of fraudulent transaction attempts"

Wow

Adding insult to injury, I now have the wrong item in my hands and I feel just a little bit criminalised to boot. I waited in anticipation of the call to come and the website left me in quite a self-righteous, unhappy customer kind of state.

In the end the call back was really good: there was an apology, a reason (well an excuse - data entry error), and they sent me the right product out in exchange. Even so, the experience was unsatisfactory and inconvenient and I'm unlikely to use the site again.

If only the site could have reflected their (eventually) helpful manner in resolving the problem; and if only Threadless sold photo paper.

http://www.jamesandjoe.co.uk/

This is the finest Google mashup I've seen since the last Google mashup, and possibly the most original use of the Google maps api I've seen. Why? Because it's not with maps. Instead, they use the functionality to zoom into images and let you get a closer look at their work.

Which makes me think: the Google maps api is free, there's no limits on commercial use that I can find in the terms of use - so why do retail sites continue to pay for Scene7 and others for their zoom functionality?

Anyway back to James and Joe. You need to go into Projects to get to the mash up goodness: these guys are very funny, very engaging, very clever, and very creative; I'm particularly fond of their dairy council diet coke parody, possibly because I'm a new dad.

http://oakland.crimespotting.org

Crime database websites are nothing new, and I wonder how helpful they really are to people. Does access to this information empower you, or does it create an atmosphere of terror which within which the state is free to curtail civil liberties?

Whatever, the interaction here is simply grand.