• Save
It looks like you're embedding user assistance - would you like help?
Upcoming SlideShare
Loading in...5
×
 

It looks like you're embedding user assistance - would you like help?

on

  • 810 views

It looks like you’re embedding user assistance. Would you like help? ...

It looks like you’re embedding user assistance. Would you like help?


Delivering integrated point-of-need help information in a software application.(Roger Hart, Dominic Smith)


A presentation from Technical Communication UK 2011


The death of Clippy, Microsoft’s much-maligned “Office Assistant” in 2007, did not mean the death of user assistance. It just meant that technical authors needed to work harder with developers to find genuinely useful mechanisms for assisting their users.


Embedded user assistance is a mainstay of technical communications at Red Gate, saving our customers time, and making our products more usable.


We’ll teach you everything we’ve learned about embedded UA, and show you how to design, test, and write for it. Plus, we’ll do it for real – in the workshop you’ll design your own embedded UA, and test it out.

(There's a bunch of extra info in the speaker notes)

Statistics

Views

Total Views
810
Views on SlideShare
809
Embed Views
1

Actions

Likes
0
Downloads
0
Comments
0

1 Embed 1

http://twitter.com 1

Accessibility

Upload Details

Uploaded via as Microsoft PowerPoint

Usage Rights

© All Rights Reserved

Report content

Flagged as inappropriate Flag as inappropriate
Flag as inappropriate

Select your reason for flagging this presentation as inappropriate.

Cancel
  • Full Name Full Name Comment goes here.
    Are you sure you want to
    Your message goes here
    Processing…
Post Comment
Edit your comment
  • Title cardAm I the only one who thinksClippy’s eyebrow waggle was kind of sleazy?
  • Who are weWho are Red Gate
  • Video of Clippy. Dom to talk through his modes of action:Asking Clippy a plain-English question, causing the help topic to load (reactive mode);Clicking somewhere where you’re not allowed, making Clippy tell you what to do to be able to click there [not very helpfully] (proactive, non-Bayesian-logic mode);The infamous ‘It looks like you’re writing a letter’ scenario (proactive, Bayesian-logic mode);In place of an information dialog (spell check complete);In place of an alert (document not saved).
  • Discussion of what’s wrong with Clippy, leading to what’s potentially right.
  • What is eUA?Point of needNobody wants to read the helpeUA is all about not interrupting workflow.
  • You have to choose the right typeContext. Workflow. UI positionConfusion
  • Splash screens. Possible that nobody read them, so re-use existing content. More of a “nothing to lose” approach.Aim at tips and “getting started”
  • Simple static textFilling blank space“What next?” moments when users get stuck.
  • Richer “Getting started” info.Visual order implies a workflow.More graphical, taking more screen real-estate, so needs design & UX input.
  • Pop-outs “Green Widgets”The “WTF?” Button.What’s this? What does that mean? What happens next? Wider contextual info.Simple HTML panel.Question marks performed better in UX tests than info icons, which users found more optional.Include links out to web help.
  • Dynamic, on-hover panels.Like a much richer tooltip.Great for product options.Always place to the side of, not on top or below the thing they refer to. (where people look naturally).
  • Tooltips. Delicious, simple.Can include graphics.Use line breaks for readability.
  • Hyperlinks with HTML pop outs.Feature usage research suggests people are massively more likely to click these than almost any other UA.You can control the text, steer the call to action.
  • UI controls as UA (the workflow buttons at the top).Visual order is workflow.
  • Extreme UA.Started as a joke, but users loved it.Big, bold, for when you’re really stuck.
  • Introduce the tasks1) Add UA to our deliberately bad dialog2) Convert Microsoft web help for adding an index in Word 2003 into embedded UA on the dialog.
  • Task: Add UA to make this suck less. You can include design feedback/changes.
  • Video. Show paper prototype. Explain technique and limitations.
  • UX prototyping exercise. 2 sets of groups – improve wireframes with UA, test, come up with further improvements based on test feedback.
  • The best and simplest UA is UI text. It’s easy to get buy in to change this. Makes developers’ live easier too.RESX files & abstraction.
  • Over to Dr Dom.
  • Analyses events, the order they occur in and time takenUser specifies the threshold at which they want helpIdentifies long-term needs and suggests further reading at end of sessionHorvitz et al. (1997)
  • Proactive assistantsProactive assistants evaluated more positively than reactive.Proactive help evaluated positively even if it’s wrong most of the time!Users want proactive information before they make mistakes.Reactive assistantsUsers blame their queries if incorrect / inappropriate answers given for reactive help (provided not repeatedly incorrect).Xiao et al. (2003a)
  • Agents are gender-stereotyped: female voices are considered less friendly than male voices.If an agent is praised by another computer with a male voice, humans consider the agent more competent than when the same praise is delivered with a female voice.If a computer asks you to rate its efficacy, humans are more likely to praise the computer than when rating the same program on a different computer.Humans volunteer to help a ‘helpful’ computer.Humans are more intimate with intimate agents.Nass (2000)
  • Responses to user agentsWebsites are rated positively if an agent tells jokesAn agent introduced as ‘fun’ is much less favourable to users if it doesn’t tell jokesJokes encourage people to use a website againA website with jokes is considered more reliableBut:A website is rated even more positively if there’s a joke in a UI label (not an agent)People buy more if a website agent is introduced as ‘useful’ (not ‘fun’)Swartz (2003)
  • Making users more likely to trust the agent now…Beautiful and helpful (but beautiful and unhelpful is worst)Looks similar to the userFor health advice, making the agent appear overweightMaking users least likely to use the agent in the future…For males only: Looks similar to the user and was unhelpful last timeVan Vugt (2008)
  • Beginners prefer to rely on more experienced usersUsers want to become self-sufficientSwartz (2003)
  • Embodied agents have no effect on the users performance (versus plain labels)Achieving the task is more likely to affect satisfaction with the software than the appearance of an agent

It looks like you're embedding user assistance - would you like help? It looks like you're embedding user assistance - would you like help? Presentation Transcript

  • rough agenda
  • www.red-gate.com
    Image credit: http://www.flickr.com/photos/25285370@N02/4369967691
  • to bury,
    not to praise?
  • Image credit: Flickr user – Jovike
  • Image credit: Flickr user – Lemsipmatt
  • Embedded UA
    at
  • try it
    yourselves
    Image credit: http://bwog.com/2009/11/01/costume-contest-its-clippy/
  • time for
    a break
    Image credit: Flickr user – Jinny.Wong
  • paper prototyping
  • try it
    yourselves
    Image credit: http://dpnation.wordpress.com/2008/08/14/sensitive-skybil/
  • 1 thing
    to take away
  • clear interface text
    matters
    Image credit: Flickr user – Drab Makyo
  • Bonus:
    Embedded assistantsa research-based perspective
  • The Lumière Project
    Horvitz et al. (1997)
    http://www.flickr.com/photos/clagnut/2865788674/
  • Responses to user agents
    http://www.flickr.com/photos/philandpam/2190067385/
  • Social responses to agents
    http://www.flickr.com/photos/dm-set/3559804855/
  • Should an agent tell jokes?
    http://www.flickr.com/photos/checiap/49390797/
  • Responses to user agents
    ‘Beauty is a dangerous treasure to cherish. It is a catalyst of extreme judgements.’(p. 79)
    ←This man has kept me in his basement for 3 weeks
    http://www.flickr.com/photos/77043400@N00/2661695451/
  • What users want
    Swartz (2003)
    http://www.flickr.com/photos/chris_radcliff/1295389897/
  • Should we use assistants at all?
    ‘We found that performance with both reactive and proactive assistants was equivalent to that with printed help. Proactive suggestions made by the assistant did not improve performance but were viewed as being helpful by study participants.’
    Xiao et al. (2003b)
  • Impolite software
    Don’t pre-empt user choices
    Don’t dominate the interaction
    Make it easy for the user to have control
    Don’t interrupt the user, unless necessary
    Be brief
    Help users make desired choices
    Ask before acting on interface resources
    If in doubt ask; if you ask, remember
    Keep the user informed of your actions
    Don’t pester
    Offer choices relevant to the user context
    Offer useful choices
    Say excuse me, thank you and please
    Whitworth, B. (2005)
  • Bibliography
    Horvitz, E, Breese, J and Heckerman, D, et al. (c.1997). The Lumière Project: Bayesian User Modeling for Inferring the Goals and Needs of Software Users. Redmond: Microsoft Research
    Nass, C and Moon, Y (2000). Machines and Mindlessness: Social Responses to Computers. Journal of Social Issues, 56:1 (pp. 81-103)
    Swartz, L (2003). Why people hate the paperclip: Labels, appearance, behavior and social responses to user interface agents. Dissertation submitted to Stanford University.
    van Vugt, H (2008). Embodied agents from a user's perspective. PhD thesis submitted to VrijeUniversiteit, Amsterdam.
    Whitworth, B (2005). Polite Computingin Behaviour and Information Technology 24:5 (pp. 353-363)
    Xiao, J, Stasko, J and Catrambone, R (c. 2003). An Empirical Study of the Effect of Agent Competence on User Performance and Perception. (Atlanta: Georgia Institute of Technology)
    Xiao, J, Catrambone, R and Stasko, J (2003). Be quiet? Evaluating proactive and reactive user interface assistants. (Atlanta: Georgia Institute of Technology)
  • Twitter: @RMH40
    @DNAS2
    Email: Roger.Hart@red-gate.com
    Dom.Smith@red-gate.com
    Blog: http://tinyurl.com/RMHBlog
    http://tinyurl.com/DNASBlog