Per some recent suggestions, I've decided to elaborate a little more about some of the things I learned at the SXSWi panels. First up: the keynote by Kathy Sierra of the Creating Passionate Users blog. Of course I can't recreate Kathy's talk or her immense knowledge on the subject, but I can give the you gist of what she talked about. If you'd like to listen to Kathy's speech, SXSW has a podcast or her talk available for download. There is a edited video of her talk available as well. And, Kathy writes a little bit of her experiences on her own blog.
Phew! And now, on to the talk.
Kathy started out asking, what's that "one thing" about human interaction that makes it so meaningful? And how do we convey that emotional meaning in our software? Some people think it's simply the smell of other humans that makes it meaningful - but we can't put that in our software. Really, nobody knows for sure. But no matter what, a more human experience makes for better software and more passionate users.
When you are frustrated or confused, you make a face - usually furrowing your eyebrows and tilting your head. People around you see that, and they respond. They ask you questions, they help you out. Problem is, software can't do that. The software has no idea what the user looks like or is feeling. It can't tell if the user is confused, working well, or even if they're totally lost.
Thus, our software needs a way to have the user tell us they're confused. Sounds simple enough. And when the users is at this confused stage, it's a very critical time. If we don't pick up on the users confusion early on, we're going to loose them and their passion. Nobody is passionate when they suck at your software.
When someone is interacting with your software, there are some perceived milestones people achieve. First, there is the "suck threshold." People here say to themselves, "Wow, I suck. I'm not good I'm not bad...I just suck." Secondly, there is the passion threshold. Here, the user is actually getting really good at the software and is really enjoying it. Quick tip: all things being equal, he who gets their users past the suck threshold the fastest wins. The faster we can get people past that threshold, the more passionate they'll be about our software.
So what do we do? How do we get them to stop sucking and move past the passion threshold? Let them tell us. They need a "WTF?" button. It lets the user tell us, clear as day, when they're confused. So back when the user is making that face, but now at the computer, they have a way tell us. Click the "WTF?" button.
But, your usual knowledge base or FAQ system isn't going to cut it. You might ask, "Isn't that what they're doing when they click FAQ? Aren't they saying "WTF?" Not really. As Kathy says, FAQs are written for a user who is mildly in love with the application and likes the product. The people who actually need and use the FAQs are usually new to the software and totally lost. Help documents and FAQs are meant more for the higher level users, those who are closer to the passion threshold.
So what should happen in our software when someone presses the "WTF?" button? Well, Kathy suggests to "start thinking like a human." In the real world, when you're talking with someone and they look confused, you simply ask them, "what's wrong?" You need to emulate that in your software. Ask if the user is even in the right place? Maybe they went too far in the process and actually need to be moved to a different location? In your help documents, Kathy suggests using top level phrases like, "I'm lost," or, "I'm confused" after someone clicks the "WTF?" button. They're real user questions real people might say, and help the software identify what's wrong.
Being human is very important - especially with the writing style of your help text. Make sure you talk like a person. Kathy notes several studies which show tiny changes in help language - making it more conversation like, personalizing it, using the word "you" - can have huge benefits. But why? Well one of the theories is when your brain reads conversational text, it doesn't know the difference between reading conversation and having conversation. And so your brain turns on and says, "Crap, this is a conversation, I need to pay attention!"
Kathy then gave a bunch of examples of software, and some that got it right and some that got it wrong. She pointed at Excel, and it's painfully un-human help system. All these examples helped to solidify the main point of the talk: rather than creating passionate users through marketing, where you can usually out spend the other guys to win, you should instead focus on out teaching your users better than the next guy. That's the true key to success, and creating passionate users.
What I learned at SXSWi - Kathy Sierra
Events
The author's views are entirely his or her own (excluding the unlikely event of hypnosis) and may not always reflect the views of Moz.
However, you can go too far, and make an Office Paperclip thing. Human is good. Annoying isn't.
Don't patronise me, just solve my problem without making me feel/look like an idiot.
I think that's why the help button should be named the "WTF?" button, like Fluxx said - it's very human without trying to be a cute customer service representative, like the Paperclip.
I had posted a review on my blog about her Keynote -- suffice it to say, I was very impressed and became an immediate fan of hers.
Just so I'm not all link spammy with my site and force people to see it just to read what I wrote, I'll just paste it here:
You're more than welcome to read my 4-part review on my blog.
By the way, her gratuitious puppy pictures in the presentation were great.
Yeah. I never fully heard why she inserted those. My guess is that our brains remember ideas and concepts (her presentation) better when we associate it with a plesant thought (the puppies).
Close. My guess would be precisely because it's got nothing to do with the subject. Whenever I give presentations, I use random things, be they images or music or whatever to make people go "ooh!"
:o <- "ooh" face.
Make people sit up and take notice, and that will trigger a chemical reaction in the brain that makes whatever they're seeing/hearing/doing memorable. Associate important things with importance using emotion. In other words, make them do this:
:o
and then do this:
:)
but NEVER this:
:s
unless this is coming very soon after:
:D
Don't confuse people unless you're going to clear up that confusion straight after. People don't like being confused or feeling stupid.
Thanks for sharing more of the details Jeff and I'm looking forward to more from some of the other sessions.
I think the idea of making help files and FAQa more human and conversational is an excellent idea. One of the things I've often noticed about both is that you sometimes already need to know the solution to your problem before being able to phrase your question in a way to find the documentation. There's a definite art to writing help files.
Did Kathy have any suggestions for collecting information. Would it work better to collect information through a form? Maybe am ever present link within the help files or help menu to send a message via email?
One things you see a lot is a FAQ and/or help system built as a forum where questions can be responded to directly by developers and other users. I've always liked help systems set up this way.
ideally we'd be able to anticipate user confusion and eliminate it before it happens, but I think we'd all agree that's impossible so being able to collect the info is a huge benefit toward preventing confusion for other users.
Kathy didn't really make mention of how to collect information the best way. I think a form with "conversational" style text would probably be just fine.
Forums are good for the pseudo-power user who is looking to get better, but a lot of times you're not really sure the question you're asking.
That's true and I think it's also the case with a lot of help systems. I'm thinking of the times you can describe what you want to do, but don't know the terms that the help or FAQ uses to refer to it.
My thought with a forum is that you can ask your question as conversationally as you want and chances are someone will know what you mean if you aren't using the standard industry jargon.
But at the same time I know a lot of people would have an issue with posting a question in the first place so maybe it's not the solution for everyone.
Great post Jeff. This was one of the Keynotes that I was really bummed out to miss. That's the great thing about SXSW, you can go back and see/listen to most every panel.
Yup...though, they need to get the podcasts up faster. I want my "Getting Unstuck"panel audio!
I remember how we struggled withthis at the software firm I used to work at. Combining help with natural language was actually *very* difficult and predicting what verbiage a user will use for a query is a lot like being psychic... a lot like... well, keyword research *smiles*
Excellent post Jeff. I actually don't subscribe to Kathy Sierra's blog, but I'm going to start now.
This sort of ties in with good web design. You want to answer your users questions before they even have the question.
Something like, "This is where the search is" vs. "Where the hell is the search, wtf?"
Software developers are notoriously bad at this. I remember once when I was talking to a client and was trying to explain to them how bad their on-site search feature was, a developer told me that the user should adjust the way they search to fit his site.....I had to stop him right there and explain to him that it was he who needed to mold the search to fit how his visitors use it if they wanted to be successful.
Great customer service on the web begins with your site, which is basically your software. Have answers and solve problems before the questions come up and you will have a returning visitor for sure.
Good points and I think it goes to the idea that the best web design remains hidden from the user. When people can easily find what they need and are able to interact well with your site or software the design gets noticed less, which is really how it should be.