Showing posts with label IxD. Show all posts
Showing posts with label IxD. Show all posts

Thursday, July 21, 2011

Don Norman's Living with Complexity - notes for #uxbcldn

I'm reading Don Norman's Living with Complexity for a UX Book Club London meetup, but I am also trying to prepare for the Service Design short course at Central St Martins that I will be taking next month, so I'm going to focus on his comments on Service Design - especially since this is the first Don Norman book I've read where he discusses this topic.

In many ways the book is something old, something new. The old bit - and none the less true for it - is his job description for designers:
The designer's job is to provide people with appropriate mental models.
[ch 2, p29]
The official new element is the complexity of the title, and his view that many devices and services involve an irreducible complexity which cannot be magicked away, as he expresses in a quote:
Every application has an inherent amount of irreducible complexity. The only question is who will have to deal with it, the user or the developer (programmer or engineer). (Tesler and Saffer 2007)
[ch 2, p46]

Chapter 6, Systems and Services, and chapter 7, Waits, are the most directly relevant to service design. He discusses the relative immaturity of Service Design compare to Product Design in characteristically IxD terms:

The world of services is different from that of products, in part because they have not been studied as much as products. Although one would think that service providers should also adhere to the standard themes of good interaction design, that is, good feedback along with coherent conceptual models, in practice it is not that simple. Services are often complex systems, barely understood even by the service provider, with multiple components spread across many geographical locations and divisions of the company.
[ch 6, p144]
Don Norman has respect for the challenge: services are described in terms of complexity, back-stage and front-stage, recursiveness (a customer's back-stage may be a clerk's front-stage) and the systems thinking required.

Amongst other case studies, he describes IDEO's re-design of Amtrak's Acela Express: how IDEO declined an RFP for redesigning train interiors in favour of redesigning the entire service, starting with "learning about routes, timetables, costs" through 10 stages to "Continuing the journey" (how true - I would use UK long distance train services far more if I could hire a car at my destination as simply as I can at an airport), and Apple's iPod and iTunes, both illustrating roads to commercial success that involve thinking of the offering as a system and its delivery as an extended service.

I found an equally coherent but somewhat contrasting perspective in Journey to the Interface, a pamphlet published back in 2006 by Demos and Engine Service Design.

They take a more humanist approach:
Service designers do not see service as something that can be reduced to a commodity. They focus on how people actually experience services, in order to understand how large service organisations can create better relationships with their users and customers.

Experiences and relationships are the recurring themes of this pamphlet.
This model is not necessarily incompatible with the Don's - early on we get a little BUPA case study, which ends:
A tangible change that has emerged from doing this exercise regularly is that customers calling to discuss their hospital visit are now offered a checklist of things that other people in similar situations have asked. This was introduced after BUPA realised that people often didn’t know what to ask when the call finished with the question ‘is there anything else I can help you with?’ Alison, again: "You have to do as much as possible to manage getting into people’s shoes – psychologically, emotionally, physically"
Although the terminology is different, this example fits right in Don Norman's emphasis on ensuring that the user develops a mental model which includes understanding the most relevant future options.

Two things that distinguish the Demos / Engine Service Design approach from Don Norman's are,
[1] the emphasis on relationships, eg
The picture of service is no less complex from the interface as it is from a systemic perspective. What is different, however, is that the interface focuses on how people and services relate, not simply the shape of existing services.
and [2], the emphasis on co-production, eg:
Service designers focus on a specific kind of engagement:
engagement at the interface. Deliberation has to take place at the point of delivery to create the kind of engagement required for co-production – that where people are mobilised, coached, and
encouraged to participate in the ‘common enterprise’ of generating positive outcomes.
A point shared by both publications is the difficulty of measuring (and thus funding) the value of services which are designed to provide long-term or preventative outcomes - I would like to have seen some discussion of Social Impact Bonds here, as one apparently appropriate mechanism.

All in all, I'd say Don Norman's book is an excellent overview of various design issues from an Interaction Design perspective. There was a common feeling at the UX Book Club London meeting I was at (IG Index - thanks Jane and Chris!) that it would be most appropriate to an interested non-designer, and in fact the Service Design chapter was found by many to be the most interesting, perhaps because we weren't, in the main, Service Designers.

On Service Design specifically, I found it a helpful snapshot, if conservative in vision. The greater ambitions of Journey to the Interface are somewhat focussed on a specific British context, but I find it entirely plausible that changing the relationship between service users and service providers could have more radical effects than changing the relationship between consumers and their products or programs, especially when achieved by techniques like co-production.

Thursday, June 23, 2011

#AgileUX Agile Safari to MindCandy.com

AgileUX Agile Safari to MindCandy

Yesterday MindCandy were kind enough to host the Agile UX Meetup group for an Agile Safari where they showed us round their offices and discussed how they have combined User Experience Design with Agile development as part of their current success story.

MindCandy’s original hit product was PerplexCity, but they are now best known for MoshiMonsters, a games, virtual pets and social site for children aged 7-12 which has been described by some as the Facebook of kids, especially now that it has 50 million users worldwide. The product is directly funded by subscriptions, which not only provides a solid business model but gives immediate feedback on the success of new features.

What MindCandy gets from Agile

The story goes that before MoshiMonsters, MindCandy raised $10m, burned down $9m of this on waterfall-style development for PerplexCity, then, 2 months into their final 6 month plan, decided something had to change and introduced agile to make MoshiMonsters. From their perspective, the rest is history, and agile is seen as a core component of the company’s culture, from the top down.
MindCandy's Christmas Party, 2010, with the Big Board in the background

As MindCandy's main product is aimed at kids, and already has high penetration in many markets, they need to keep it fresh by adding features in order to [a] avoid letting their users get bored, and [b] avoid leaving a window of opportunity for their competitors.

How Agile works at MindCandy

A product board meets regularly and signs off projects, basically at a one-line level of abstraction eg “Add MiniGame Fun Park”.

Each project gets a team, some of whose members will be “ring-fenced” for that project (presumably others are shared as needed). The team works with a full-time scrum master (so developers can get on with development), and have their own four-column (Not Started / In Progress / In Test / Completed) section of the big board. There are some kanban-like features too, each team member can only put his name on two tasks. Teams deliver user stories in two-week cycles, the product manager (Dharmesh, in agile terms, the product owner) will focus on delivering something for each user story. Features are may be delivered as several user stories over several sprints.

MindCandy staff see themselves as being triumphantly team-focused in working practices and reporting structures. Co-location is seen as important and all UK staff currently sit together on one floor of Tea House in Shoreditch - they hope to keep the vibe as they expand to another floor.

They don’t do the “we will commit to two the following two weeks worth of work” thing, preferring to simply prioritise whatever’s still on the Not Started list. Likewise they don’t see themselves as deadline driven, though external deadlines such as school holidays are naturally part of the landscape and will help focus as decision time nears - “to ship is to choose”.

How UX works at MindCandy

Yael is the User Experience designer and works with Axel, the visual designer. They sit side-by-side in the middle of the development area. UX work is done within 2-week sprints as part of user stories. User stories are often implemented in two passes, the first (the first week of the sprint) concentrating raw functionality, the second (the second week) allowing for the UI to be optimised.

MindCandy don’t go in for big up-front research, nor do they currently use personas. They do have a good understanding of their target market from long  exposure to it, and from community managers.

Challenges overcome

MindCandy also work with an off-shore team who are fully agile and work on non-game focussed projects to allow the internal teams to work on the core games. MindCandy do company-wide sprint reviews (with team-level kick-off, stand-up and retrospectives), and the off-shore team takes part in these.

Off-shore teams are not part of the big-board but they use Acunote as a light-weight Scrum tool which allows them to do Scrum their way.

They do a lot of Skype - messaging even more than calling - and most desks (apart from marketing) just don’t have phones.

Looking forward

Their current agile and UX processes have clearly served them well so far. As their market gets more mature, MindCandy feel the need to get even more agile, both developing faster and learning faster.

They plan to introduce UX personas in order to target variations in their age 6-14 user base better, and hope to fit more regular user testing into the time and space available.

They now have a full-time business intelligence analyst to identify areas of concern which might otherwise be masked by their overall success, and are extending their use of analytics.

On the productivity front they have a full-time tools and environment expert who describes himself a the “force multiplier”, and who looks for areas of friction which he can automate and / or accelerate.

Finally

I’d really like to thanks Martyn, Raph, Axel, Yael and Dharmesh of MindCandy for a fascinating and exciting Agile Safari, and of course Johanna Kollman and Agile UX for organising it.

Sunday, October 10, 2010

Conflicts in quality for a London Man-and-Van Company website

This afternoon I moved half a garage full of books from Hampstead (in North London) to Earlsfield (in South-West London), a process that was made much simpler and less painful by using a prompt and helpful Man and Van company specialising in small moves.

Naturally I got them off the web, and since my work clients are typically rather large banks or building societies and my entrepreneurial friends tend to be in technology or information, I was delighted to find that Leonardo, the nice Brazilian driving me and my books across London, was in fact the founder of the company, and I asked him about his web site design.

From a design perspective their web site is a slightly bizarre mix of truly helpful and innovative stuff, like the fact that you simply book a 3-hour slot (for £40 - where do they get their profit?) from an online calendar, together with somewhat distracting Google-optimised content targeting the different London regions.

For me, as a budding UX designer, this was a bruising encounter with commercial realities. If they kept the site small and focussed, the usability, and indeed the whole user experience, would be improved. But their site has to appear in the first page of search results, and they pay good money to someone to ensure that it continues to do so.

I know very little about SEO, so the only suggestion I was able to make was to suggest they invite satisfied customers to review them online, in order to reduce their dependency on more artificial methods which [a] confuse the site, and [b] might be vulnerable to a Google re-interpretation.

Lesson [1] for some sites, you have to treat the Search Engine as one of your key users.
Lesson [2] even white-hat search optimisation can conflict with other UX goals.

Saturday, February 27, 2010

Notes on using WebSort.net

I recently had to design a top-level topic structure for a company-internal wiki. This wiki is intended to replace a predecessor which by now has a poor signal-to-ratio, being cluttered with obsolete articles and hindered by the absence of any kind of lifecycle-management, tagging or rating features. (The new wiki is being implemented using KwizCom's Wiki Plus, but that's not what I'm reviewing here)

But designing a good top-level structure for the mixed bag of topics found in a typical wiki is, like "go forth and sin no more", one of those tasks easier said than done. So I decided to try my first card sort, using 90 page titles from the old wiki as input to an open sort. I initially considered printing the list out and cutting it into physical cards, but this turns out to be alarmingly hard to do productively, especially to any level of quality, so it was time to check automated options. I came to WebSort.net from comments on a great card sort article, and, since it was free, and looked plausibly polished and complete, decided to use it.

Creating the test was simple - I registered for an account, gave the test a name, and simply copied the list of items from a text processor and pasted them into WebSort.

Running the sort was also simple - simply send a URL to your candidate card sorters. (And, if you want to learn from my mistakes, give them a more compelling reason to perform the task than the fact that it's convenient for them and helpful to you). When your users visit the link you've sent them, they get the instructions you left (I stuck with the default wording):



Performing the card sort is slick and sweet. WebSort provides a drag-and-drop interface for sorting the cards, and randomises the card order for each sorter. Unfortunately the slick interface is provided in Flash, and there are some gaps between the safety, transparency and reversibility of direct manipulation and the overall user experience - there was no way for users to print screens, or to save and resume. These are irritations, I'd say that the primary function here has been very well delivered.



Reviewing the results is slightly less polished. You see a list of all completed sorts, keyed by the sorter's email address. You select one or more names and hit "Reload" to load that particular data set. Once loaded the data set can be downloaded as a spreadsheet or in various text formats. The default display is "Category Summaries". This, along with "Categories * Items" is of limited usefulness in an open card sort where users invent their own category names, since each user typically invents different names. WebSort have helpfully provided this view with a "Merge categories" button to merge selected categories, but with no "Undo" or "Save-and-resume" functionality, I found this phase frustrating (of course category merging is only an issue for open sorts, not for closed sorts - WebSort supports both types).



There is also the mandatory tree diagram (aka dendrogram) which I found surprisingly unhelpful - this may have been a consequence of the low number of responses I was dealing with, but I have seen similar reactions from others.

All in all, I'd say that WebSort.Net is an excellent way of conducting and capturing a cardsort, with adequate analysis, but let down by weak category merging.

Wednesday, April 15, 2009

M364 Block 1, Unit 4, Activity 1

[Please complete the assignment on page 337 of the Set Book]
  1. Reconsider the HutchWorld design and evaluation case study and note what was evaluated, why and when, and what was learned at each stage?
  2. How was the design advanced after each round of evaluation?
  3. What were the main constraints that influenced the evaluation?
  4. How did the stages and choice of techniques build on and complement each other (ie triangulate)?
  5. Which parts of the evaluation were directed at usability goals and which at user experience goals? Which additional goals not mentioned in the study could the evaluations have focussed on?
Reconsider the HutchWorld design and evaluation case study and note what was evaluated, why and when, and what was learned at each stage?
There were several rounds of evaluation. The design was evaluated in terms of usability and scope. Evaluation of the first two prototypes, including the first series of usability tests, led the team to learn that there were problems with the initial scope of the project, so they changed the scope, limiting the 3D virtual world to just Reception, but adding support for asynchronous messaging, games, and locating approved medical sites. When the application was redesigned as a portal it was re-tested and further refined, allowing the team to learn about more specific usability issues. One goal of the new testing was to ensure that the system supported multi-user interactions. 

In general evaluation seems to have started with requirements and scope, then moved on to usability and user experience, in other words from whether the project had the right aims and goals to whether these goals were being reached.

How was design advanced after each round of testing?
There were two main rounds of testing. The first round of testing led to a fairly radical advancement of the scope of the design, from an intensively 3D virtual reality experience to a portal with support for email. games, medical queries and rather limited 3D functionality. The second round was  more focussed on detailed usability testing and led to incremental improvements in the design.

What were the main constraints that influenced the evaluation?
The team found it difficult to arrange testers rapidly for reasons inherent in their choice of user group - the potential users were sick and had limited energy and availability.

How did the stages and choice of techniques build on and complement each other (ie triangulate)?
The evaluation techniques ranged in nature from quite open and general (interviews, focus groups) to very specific (scripted usability tests). The more open evaluation techniques provided aims and scenarios which provided the context and goals of the more specific evaluation techniques

Which parts of the evaluation were directed at usability goals and which at user experience goals? Which additional goals, not mentioned in the study, could the evaluations have focussed on?
The initial interviews and resulting scope analysis were largely expressed in terms of user experience - one over-arching requirement, for example, being to reduce the social isolation of patients and care-givers. The formal tests were more focussed on usability, but some of the results are again best described in terms of user experience - for example, the problem with the purely synchronous early prototype never reaching critical mass is most simply explained by the fact that it is no fun being in a chat room on your own. Similarly there are UX explanations for the patients' preference for online games (fun), the search for recommended medical sites (helpful) and email (emotionally fulfilling). In addition to the carefully scripted usability tests in each round, there was a short questionnaire which asked both usability and user-experience type questions.

I think that the project could have focussed on the user-experience goals of motivating and satisfying, since these would highly relevent to socially isolated users suffering from energy-sapping conditions.


M364 Block 1, Unit 3, Activity 9

How user-centred was the approach taken by Tokairo? Start by listing the main stakeholders - beneficiaries, decision makers, gatekeepers and workers - and discussing their roles in the development process.

Then analyse Tokairo's approach against each of the five principles of user-centered development given on page 286 of the Set Book
The main stakeholders in the project were, at a corporate level, Excel plc as parent of Tankfreight and Shell, the project's client. Shell were responsible for allowing Tankfreight and Tokairo access to their depot. Tankfreight were probably the main beneficiary since they commissioned the project, and automating the delivery reporting process should reduce costs and increase accuracy.

The more immediate stakeholders were Tankfreight project management (Hugh Rowntree and Rachel Darley), Tankfreight's account manager for Shell, the driver foremen, trade union representative and the drivers themselves.

Hugh and Rachel were gatekeepers to the drivers and Hugh was presumably a decision maker in commissioning the project, and, together with Rachel, signing-off design options and prototypes.

Tankfreight's account manager may have been a gatekeeper for access to the Shell depot, and would have had some involvement in high-level project decisions.

The driver foremen and union representatives are not only gatekeepers to the drivers but possibly themselves users of the system, as workers, and possibly beneficiaries, either as users or because it makes the drivers happier.  

The drivers themselves are the primary users of the system, possible beneficiaries (if the forms are more accurate and they get fewer forms being returned to them) and are involved as workers.

Presumably others are also involved, either as workers or beneficiaries - those who used to enter the drivers' forms manually, those who manufacture and sell the kiosks, and those who service them.

Now let's evaluate Tokairo's approach to the five principles of user-centered development.

Users' tasks and goals are the driving force behind the development
Tokairo's initial input comes from Hugh Rowntree and Rachel Darley, who know the drivers, their tasks and their environment well, and Tokairo had already had access to the users before the project even started. They appear to have a good understanding of users' tasks and goals. But satisfying the drivers' tasks and goals is probably a necessary rather than a sufficient condition for the project's success, the final criterion presumably being that the system reduces costs and increases accuracy. 

Users' behaviour and context of use are studied and the system is designed to support them
As part of the system audit, "Treve actually went initially to the oil terminals and depots. Treve met the drivers and the driver foremen". The team appear to have a detailed understanding of what the drivers do, and of where they do it, in the cab and at depot reception, and it is clear that design decisions directly reflect these factors.

Users' characteristics are captured and designed for
Once again the team appear to have a detailed picture of their user group, of what they have in common (being literate, non-academic males, short of time, primarily interested in earning a living and going home) and where they diverge (level of interest in computers). And again, this is clearly reflected in the team's account of how they made design decisions. 

Users are consulted throughout development from earliest phases to the latest, and their input is seriously taken into account
This principle was defintely observed during the requirements phase of this project. It is clear that Tokairo would normally prefer to consult users during the design and implementation phases but felt that there were specific reasons they shouldn't do so in this case, although the form did get some early testing, which is a kind of user-consultation.

All design decision are taken within the context of the users, their work, and their environment
This principle seems to have been observed very thoroughly. In fact it seems that the team ascribe the success of the project to this factor.

Given that the project followed at least four of the five principles, and that the remaining principle was partially missed for specific reasons rather than lack of interest or lack of user focus, I would say that the approach to this project was quite highly user-centered.

M364 Block 1, Unit 3, Activity 8

Describe the approaches taken to user involvement in the Tokairo case study and discuss these using the issues you identified in Review Question 5 [List the (eight) issues you need to consider when deciding on the appropriate level of user involvement]. What alternative approaches might they have taken? You should refer to the case studies in Section 9.2.1 of the Set Book as examples.

What advantages and disadvantages might these approaches have had?

The set book lists the following issues that needed to be considered when deciding the appropriate level of user involvement:

Can you identify the users, or are they the open market?
In this case the users were already identified, as being the drivers.

How many users are there? Tens or Thousands?
We don't know exactly how many users there are, though common sense suggests it might be tens or maybe low hundreds. In any case, given that the users are, by the nature of their job, never in the same place at the same time, it's too many people for them all to be consulted easily or cheaply.

How long is the project expected to take?
Again, we don't know how long the project was estimated to take or how long it actually took, but in terms of user involvement the answer is probably that it was going to take too long to consider co-opting a real user for the duration, but not so long that any significent user practices or requirements would change before the project was completed.

Do you want a major contribution from users, or just advice and guidance?
There seems to have been a clear assumption, initially from the client but accepted by Tokairo, that the real users' main contribution should be to the requirements process, to alpha-testing the form and to beta-testing the entire system. Rachel, the systems analyst, acted as a proxy user for reviewing the design, but her contribution (apart from suggesting that the buttons be colour-coded) seems to have been mainly evaluation. 

How many users do you want involved with the project?
The team didn't want all users to be involved but consulted user and stakeholder representatives widely during the requirements phase. For the main design options, design and evaluation activities they mainly used Rachel, the user proxy. The notable exception is that they tested the form design on one driver area before beta-testing the entire solution, presumably having identified this as being at higher risk of failure (eg due to the driver's environment when filling in the forms) than the kiosk design.

Is consistency of user input important?
There was no continuous user involvement apart from that of Rachel, the user proxy, so consistency of user input does not seem to have arisen as a question involving the drivers. The consistency of Rachel's involvement is likely to have been helpful to the design process.

How important is familiarity with the system?
Given the general non-involvement of the drivers, familiarity with the system under development does not seem to have been a significent issue. Rachel, as the sole proxy user, was familiar with the system as it was designed, and this would have helped her contribution.

How important is it for involved users to be in contact with the user group they represent?
As there does not seem to have been any change in the drivers' environment or practices during the project, this was probably not an issue for Rachel with respect to the drivers. 

Comparing this approach with the Microsoft or OU case studies in the Set book, Tokairo could have co-opted a driver to work with the team, presumably part time. This would have been useful if the team had worries about the correctness of their scope and requirements (as the Open University appears to have had) but would have involved disrupting the availability and working practices of the driver in question, and would have risked inappropriate feedback due to lack of appropriate user motivation. In fact the team seem to have regarded this as tightly scoped exercise with well-understood requirements (especially given the comprehensive requirements phase), so the advantages of this step would have been low. 

They could also have conducted workshops and prototyping sessions, as the OU did. This would have had the advantage of reducing the risk of "requiring more changes during the prototype ... and even [the] pre-live stage", but it would have been disruptive to operations, the drivers and their management, and premature exposure would have risked "destructive feedback".

They could also have performed lab-based usability testing, as Microsoft does. This would have reduced the risk of the overall system proving unacceptable during beta-testing but the team seem to have felt that the risk of this failure was not sufficiently high to warrant the required level of disruption and expense. 

Sunday, March 15, 2009

M364 Block 1, Unit 3, Activity 4

I now want you to draw a design for an electronic calendar system that is radically different to the outline sketch in Figure 6.1. [which shows a page and book model] You only need to draw a single sketch in order to illustrate your design. Try asking yourself questions such as: how could a mobile telephone, with a small sceen, be used to access an electronic calendar? What are the characteristics of a magnifying glass, and how might these characteristics be of benefit when designing an electronic calendar?
[click on thumbnail for fullsize illustration - with apologies to Scott McCloud whose Understanding Comics I've just been reading for UX Bookclub London]