We made mistakes

In 2007, my DJ brother Gabriel asked me to build him a website. I didn’t know at the time, but it would turn out to be the best thing I ever did.

It took ages to build. I had to learn HTML, CSS and Expression Engine. Five years later Gabriel’s a full time DJ and I’m a user experience designer.

But the best thing about building your own site is the mistakes. Our biggest ones taught me unforgettable lessons about information architecture.

Names are not just about clarity

Global navigation is a great way to reinforce messages

Global navigation is a great way to reinforce messages

Our first mistake was naming. He plays music that people LOVE to dance to. Not shoe gazing indie – raucous, irreverent Jamaican dancehall crossed with UK club music. What do you call the section about their shows? 

Don’t call it Events (like we did). It’s clear, but it’s dull.

Events doesn’t sound like fun party music. Events doesn’t even sound like music. Events sounds like corporate functions in conference centres with delegates drinking terrible coffee and talking about ‘getting visibility’.

What should it be called? Parties. Raves. Jams. Gigs. Dances. Shows. Anything that communicates some excitement alongside the clarity. Your global navigation is on every page so what you choose to include, and the words you use, are a huge opportunity to tell your story. Use them wisely.

Not all content is created equal

Put your most valuable content on the homepage

Put your most valuable content on the homepage

Our second mistake was a classic. We structured the site to match our mental model by splitting the Music section into original productions, mash-ups, remixes, refixes, mixes, live shows and radio shows. Clear. Logical. Wrong.

Why? People only care about two categories. Good music. Bad music.

Only publish the good stuff. Don’t hide it in subsections. Put the very best on your homepage so people can get at it within two seconds of arriving.

Truthfully, very few organisations have enough high quality content to justify complicated hierarchies. Much better to publish a stellar subset and leave your users wanting more. Or you risk overwhelming them with choice.

Humans beat computers (sometimes)

Some content is best left to humans to edit by hand

Some content is best left to humans to edit by hand

Our third mistake – and this one is an all time favourite pastime of mine – was getting carried away with the content model. We designed our events to have titles, venues, locations, prices, addresses, concessions, web links, booking offices, artists and plenty more. We were exceptionally proud of our design.

This pride was misplaced.

Within two minutes of entering the first event we realised we didn’t have all the right information. We made some fields optional, which broke the visual design by leaving gaps where content was previously. We hacked the code with if/else statements for millions of data combinations. And it still never worked properly.

In the end, three years later, the solution was stupidly simple.

For rapidly changing content where you can’t predict the shape of the data, just have a page that a human can edit by hand. We’re good at that.

Huge mistakes can bring huge benefits

We saved the biggest mistake until last – last year Google killed the site for being infected with malware. We hadn’t updated Expression Engine for four years and deserved what we got, so we started over. (Losing a thousand pages overnight was easier and far more effective than a content audit!).

But this isn’t about Gabriel’s site anyway. It’s not even about information architecture. It’s really about how I learned to learn from my own mistakes.

Owning up to my bad decisions was horrible at first. It made me feel like I had given bad advice and often felt easier to argue back. The turning point was a conversation where Gabriel pointed out how much it took to maintain the site and I was practically shouting in denial. He, the client, was right.

Over time, I got better at admitting mistakes and started to relish finding flaws in my own thinking. Welcoming criticism is the hardest thing I’ve ever learned to do – and I’m still working on it – but nothing has improved my work quicker.

Let me know what you think on @myddelton. Thanks to Gabriel for putting up with me, @MagsHanley for encouraging us to share IA war stories and @Mike_FTW and @slowtext for their great podcast, Let’s Make Mistakes.

The hidden business of user experience design

User research often throws up problems beyond the scope of designing websites and applications. Awkward things like corporate focus, content freshness, customer service relationships and database quality problems.

All affect the user’s experience, yet addressing the business processes responsible is rarely seen as part of user experience design. Which is a shame because failing to address business issues can undo all our design work.

Redesign the organisation

My first exposure to user research was the CABE redesign in 2008. The big (unsurprising) finding was that people wanted to find content by themes like housing, health or sustainability.

It didn’t take long to design the information architecture, but three years later we still had problems with creating the content. Sections like sustainability, which had a dedicated internal team, had great content. Areas which required cross-team collaboration, like health or housing, were poor.

The lesson? It’s not enough to redesign your website - sometimes you have to redesign your organisation. CABE should have created new teams, or refined their editorial processes, to populate our shiny new information architecture.

Forget the front end (sometimes)

Another site that I worked on had serious usability issues. Frustrated at my failure to convince people of their importance, I did some user research to show the need for change.

But none of the research findings related to my usability concerns.

Instead, the real user issues touched on multiple areas of the business. Addressing them involved difficult conversations, serious data analysis, renegotiation of contracts and even culture change.

The lesson? User research throws up some issues that can’t be addressed with wireframes and prototypes. (Also, don’t do research to prove yourself right!).

Fight for better processes

You might wonder whether these issues matter to user experience designers. They sound suspiciously like things other people should be sorting out.

Maybe. But many managers aren’t digital natives, let alone advocates of user-centred design. They won’t make good strategic or operational decisions without good advice and, weird as it seems, our research is often the first time they find out what users really think. So, for now at least, it falls on us to fight for the better business processes our designs deserve.

And if we don’t? The beautiful websites and applications we design won’t work for users. No matter how good they are on paper.

Let me know what you think on @myddelton.

Concatenate rules

Normal people don’t usually thank you for teaching them Excel tricks. Unless that trick is the Concatenate function. Then they love you forever.

Concatenate joins together text from multiple cells. Let’s say you have two cells containing “Hidden” (A1) and “Gems” (B1). Here’s how to combine them:

results in “HiddenGems”

But Concatenate is a long unusual word which makes it hard to remember. So it’s good you can use an ampersand instead, just like you use plus and minus:

results in “HiddenGems” as well

You can add your own characters into the formula too. Insert a friendly space (or any character string) by putting it in quotation marks:

=A1&“ ”&B1
results in “Hidden Gems”, which is much prettier

It’s difficult to explain how useful Concatenate is. I use it to build greetings from title/firstname/lastname (mailouts), construct working URLs from unique identifiers (content audits), add HTML tags to list content (CMS uploads) and export quotes to Wordle to make pretty word clouds (data visualisation).

If you liked this you should read about PureText. Let me know what you think on @myddelton and follow @wizardofexcel to supercharge your Excel skills.

The hidden powers of SurveyMonkey

Twice I’ve joined companies to find professional researchers laughing at my use of SurveyMonkey for user research. They assumed it was inadequate compared to their costly enterprise software.

But most surveys don’t need advanced features. When they saw how easy it was to do surveys with SurveyMonkey, the researchers never looked back.

In their honour, here’s my guide to the lesser-known features of SurveyMonkey.

Super simple data sharing

The whole point of a survey is getting responses. But enterprise tools put so many barriers between writing the survey and getting the responses that it’s easy to forget why you were asking the questions in the first place.

This just isn’t good enough.

So my top reason to use SurveyMonkey is it only takes 30 seconds to create a password-protected URL for your survey results. Now anyone can access the data without being able to modify the responses or mess up the survey itself. 

And what does quick and timely access to user data equal? UX converts.

One survey, multiple collectors

SurveyMonkey lets you create a single survey with many different URLs (‘collectors’). You might put one on your website, another in a mass email to registered users and a third on your Twitter account. 

All collectors gather responses to one place, but the trick is you can view data in aggregate or segment it by collector. This lets you easily see differences between your audiences at the same time as gathering overall data.

And that’s just the start. Segmentation lets you test your hunches (do people respond differently when incentivised?) and compare feedback over time (one set of feedback questions with a different collector for each time you speak).

Conditional logic that my mum understands

You don’t have to create many user surveys before you find that you want to ask different follow-up questions depending on previous answers. 

The classic example is the satisfaction survey – ask one thing to users that failed to complete their task (‘what prevented you doing this?’) and something else to users that were successful (‘what do you most value about our site?’).

Although this sounds simple, in practice it’s dangerously close to programming. And, ouch, debugging. But SurveyMonkey has an interface for branching surveys so straightforward that my mum can (and does) use it.

The creepy part

Finally, you can pass information into the survey via a personalised URL. 

Using something like Campaign Monitor you send personalised emails to all of your registered users, each with a survey link that contains their email address as part of the survey URL. When users complete the survey their response is logged alongside their email address. Without them doing anything.

Now you can contact users about issues raised without ever asking for an email address. Fewer fields to complete, and no input errors either. Win-win.

Just don’t claim this is an ‘anonymous’ survey…

It doesn’t do everything (but in a good way)

Of course, many great features are missing from SurveyMonkey. (Although the refusal to bloat it with features just makes me love it more).

You can’t drag and drop when designing your survey. Or allow respondents to upload their own files. Or completely retheme your survey with CSS. If you want to do those things then check out Wufoo.

You can’t get an RSS feed of responses. Or do computation and scripting. Or use strings from previous answers in subsequent questions. If this sounds fun, you should look at SurveyGizmo

But SurveyMonkey has been ever-present in my arsenal for the last five years, and no other tool can make that claim. It’s solid, easy, powerful, friendly, usable and cheap. If you do user surveys, you should definitely try it.

I’m not affiliated with SurveyMonkey – I just love using it. Let me know what you think on @myddelton – particularly if you know something useful I’ve missed.

Update: SurveyMonkey introduced the ability to use strings from previous answers in subsequent questions in February 2011.