My favourite design articles of 2015

The only annual post I actually look forward to is Dan Saffer’s list of his favourite design articles of the year. Go and read his list first and then see what you think of these…

Messaging Is Just Getting Started
On Conversational UIs
Facebook M’s bet and the future of Web interfaces
The radical idea that a simple messaging interface coupled with narrow AI could end up dominating interface design is the most interesting design topic of 2015. I didn’t see it coming and have no idea how to design for it. But it makes total sense and has a weird feeling of inevitability.

The AI Revolution: The Road to Superintelligence
The AI Revolution: Our Immortality or Extinction
A monster two-part article that lays out the terrifying prospect of advancing from the narrow AI we have today to a world-dominating superintelligence. I have doubts whether this is possible, but this is a phenomenal overview.

An Interview with Rochelle King, Global VP of Design at Spotify
Spotify’s design used to be a car crash. But something changed because over the last couple of years they’ve pumped out a slew of user-centred design improvements. So when Rochelle King talks about how to make design and research work together, I listen.

What Is Code?
Be one of the people who read this, rather than one of the people who meant to but didn’t. An amazing story, beautifully told.

How Discover Weekly cracked human curation at internet scale
I loved ThisIsMyJam more than anything but I’ve hardly missed it since Spotify released Discover Weekly. Astonishingly, it turned out that Matthew Ogle was involved in both. This marriage of music with user-centred design and algorithmic data is the most exciting thing I can imagine working on.

4 things I learned from Breaking Smart about the future of work
Treat work as an infinite game. Do whatever is the most interesting thing to do. Stop trying to reach closure and focus on opening things up instead. Interesting, counter-intuitive but strangely liberating ideas.

Harsh Empathy
Not many articles shock you into re-evaluating your own actions. This one, which links empathy with bad behaviour, made me think twice about mine.

I didn’t read as much in 2015 as I would have liked to, so I’m hungry for your recommendations. Send links to @myddelton or point me at your own list.

How to get your client to say yes

7 things to do that hugely increase the chances your client will sign off your designs when they’re supposed to.

Every client-facing designer struggles with sign-off. It doesn’t matter how good you are, how perfect the work is, how watertight the research was. If there was a quick fix you would know it already, but there isn’t.

I’m about to leave client-facing design, but before I go I want to share my list of things you can do if you’re tired of wasting time with wasteful sign-off processes. None of these things are tricks. All of them are based on deep respect for your client. And they all work.

These are my 7 steps to getting your client to say ‘yes’.

1. Research your client’s needs

If you want to get your client to say ‘yes’ then start with understanding what your client needs in the first place. It’s user research, but with clients.

There are two killer client research methods you should be using.

The first is the kickoff meeting. It’s your responsibility to leave with a clear understanding of exactly what the client needs from this project. Prepare in advance by going through any brief or proposal materials and write out that huge list of questions you have. Then do good some detective work and make sure you leave the workshop with your questions answered. You would be amazed how many people fail to do this.

The second is the stakeholder interview. A stakeholder interview is a one on one discussion behind closed doors with someone in your client’s organisation. This privacy gets you to the deep needs of your client, not just the surface-level ones that people talked about at the kickoff meeting. Prepare well. Speak to lots of different people. Make them feel safe. Ask tough questions. Every design project should have stakeholder interviews.

If you pay close attention, this combination of the kickoff meeting and stakeholder interviews will give you a fantastic, nuanced understanding of what your client needs from the project.

Client research is always the first step to getting your client to say ‘yes’.

2. Over-communicate like crazy

People hear what they want to hear. If you have 10 people at the same meeting, it’s likely that all 10 of them think something different happened. This is a disaster for getting your client to say ‘yes’ because everyone is working from a different understanding of the problem.

The solution is to over-communicate, to say things more often than you think you need to, and here are my two favourite ways to do it.

The first is to post notes from every meeting into Basecamp (or whatever you use). Write them in plain English and make sure you cover everything you agreed in the meeting. Then say to your client, look, this is what I think happened, and this is your chance to tell me I’m wrong, because after this I’m going to be treating these notes as the binding truth. It takes a bit of extra time but it’s worth it. Clients respond well to these notes because it shows you were listening.

The second is Jared Spool’s short-form brief (go and read his article if you haven’t already). Taking the time to write a short-form brief is good. Signing off the brief with your client is better. But the real magic is the act of reading it aloud at the start of every client meeting, because this brazen act of over-communicating guarantees that everyone has the same understanding of the project, throughout the whole project.

Only one thing matters with over-communicating: are people still saying random things that sound as if they’re talking about a different project? If they’re not, you’re already closer to getting your client to say ‘yes’.

3. Show work before (you think) it’s ready

Take a problem, lock yourself away for weeks to solve it perfectly, and then rock up to your client’s offices asking them to sign it off.

That’s how you get your client to say ‘no’.

You make a million tiny decisions that feel obvious to you but are invisible to the client. It’s fine to make these decisions but you need to find ways to make them visible to the client before you build a super-structure on top of them. Because if they’re wrong your super-structure is coming down.

Instead, show your work before you think it’s ready. I’ve shown sketches, wireframes, comps, prototypes, Word documents, spreadsheets. This is all about framing it well, knowing what to listen to, and knowing what to ignore. It’s a skill you need to learn.

Here are three reasons this helps with getting your client to say ‘yes’:

1. Showing clients things that look real prompts them to give you more information about their real needs. It’s amazing how the quality of the feedback improves when you talk about artefacts and not abstract ideas.
2. Talking through your design patterns lets you work through their objections and fears one by one. This is so much easier and more productive than trying to persuade them in one big final presentation.
3. Sharing work in progress shows your vulnerability. Asking for feedback shows you are listening. Both these things build your client’s trust in you, and in the end sign-off is all about trust.

Showing early work feels dangerous and risky when you first start doing it. But it’s probably the single best tool you have to get your client to say ‘yes’.

4. Build the ultimate sign-off process

None of this matters if you fuck up the actual sign-off process. Over the last five years I have fucked up the sign-off process more times than I can count.

This bitter experience has taught me two truths about sign-off:

• Clients always need more review opportunities than you think they do
• Clients don’t care if their new feedback contradicts their old feedback.

The ultimate sign-off process must live with these two truths if you want to get your clients to say ‘yes’. Here are two ways to do this that work.

First, schedule (many) more reviews than you think are necessary. On a recent project we worked in 3-week design sprints and the client needed 6 reviews in each sprint before they started signing off things in-sprint. Your goal is simple: get to the point where your client isn’t feeling rushed into making decisions about design, because rushed clients don’t say yes. It means you have less time to do design work, but that’s the cost of getting your client to say ‘yes’. No one said it was easy.

Secondly, control your client’s access to review materials like a maniac.When you plan 6 reviews every 3 weeks there is a HUGE risk their feedback will be late, they’ll comment on previous versions, or they’ll reverse their previous decisions. We avoid this with hardcore workflow management:

• upload the designs to InVision two hours before the review
• go through the designs in InVision during the review
• leave the designs open for comments for a few hours after the review
• remove the designs from InVision completely and close comments
• work on the next iteration based on the feedback from the review.

It’s (nearly) impossible for clients to give late feedback, or to give feedback on the wrong versions, with this system. You might think it would be impossible to get your client to work in this way, but it’s how we work.

And it works like a dream in getting your client to say ‘yes’.

5. Bend, but don’t break

How you respond to client feedback in design reviews is critical to getting sign-off later on. You should bend, but not break.

The ‘bend’ part is to concede on the stuff that doesn’t matter. Fighting for every single design decision is exhausting and counter-productive. Here are some reasons why it’s OK to concede stuff:

1. There are many ways to do things. This means there are hundreds of details where it doesn’t matter which way you solve a problem.
2. Your approach is often based on a hunch. Unless you have bottomless pockets it’s impossible to base every single design decision on research. Why go to the wall for a hunch about something that doesn’t matter?
3. Clients are right more than you like to admit. You’re attached to your design because you made it, but often their approach will be just as good if not better. You do not have a monopoly on being right.

You probably hate this advice. But if you concede on stuff that doesn’t matter then your clients not only feel listened to (which builds trust) but when you DO push back, they are much more likely to listen.

The ‘don’t break’ part? Never concede stuff that matters without a fight.Make the case for why it matters and convince your client. Use every tool at your disposal which might include the hairy arm technique, psychology studies, an appeal to ‘Apple does it like this’, crying, pointing out how much extra it will cost, trying to get the client’s boss to overrule them, sneaking it in through the back door. Whatever it takes.

This is a core skill of being a good client-facing designer. Can you judge which parts of your design are really important and which parts are not?

Getting this right is critical to getting your client to say ‘yes’.

6. Kill ambiguity where you can

Design needs ambiguity. We don’t know what the answers are, what they look like, how many there are, how big they will be, what shape they will take, or really anything about the answers at all before we start.

Clients hate ambiguity. It makes them nervous. They worry about paying you a ton of money when you even can’t tell them what they’re going to get.

We must convince clients to live with the ambiguity needed for good design. But in return we should remove ambiguity from everywhere else to stop our clients feeling nervous, because nervous people struggle to say ‘yes’.

Here are some places where you should remove ambiguity:

• Cost. Make sure you are clear about how much things cost. Alert clients to anything that changes cost and make it clear how big the change is. Don’t hide your costs and think you will get away with it. You won’t.
• Scope. Agree a clear scope upfront and keep returning to it. Don’t let the client make it bigger without proper agreement. Don’t ignore it when you come to deliver. Make sure that you deliver what you agreed to deliver.
• Time. Set times for meetings. Set times for sending over work for review. Set times for calls. Set times for milestones. Keep to all the times. If you can’t keep to a time you have set, tell your client well in advance.
• Amendments. Make notes of what the client says and what you agree to do in design reviews. Type up these notes and send them to the client. Address all the points in your notes by the time you do your next review.

You earn the right to ambiguity in the design process, where you need it, by removing ambiguity from everywhere else. Too many design agencies are ambiguous about too many things. Don’t be one of them.

If anything sums up my whole approach to getting your client to say ‘yes’, it’s being clear at every step of the way.

7. Sign it off before the end

This is a short one. It’s a hack really.

If a project ends on a Friday, set the final sign-off for Wednesday. You know there will always be final amends. If you get them on Wednesday, you can get your client to say ‘yes’ on Friday.

Hello weekend.

Good luck!

It’s taken me five years of client-facing design to work a lot of this out. It still amazes me how little of this is common practice, but this bread-and-butter stuff is never as exciting as articles about new design patterns (notice that none of these 7 tips are about doing better design work!).

Why should you care? Why should you bother? Well, designers that get stuff signed-off finish their projects on time, which means happy clients and happy agencies. These two things do not go unnoticed.

I’m off to work at the Government Digital Service now. Good luck!

This is how we ran design projects when I worked at cxpartners.

How to be a good client

10 simple rules to follow that will dramatically increase the quality of work you get from any agency you work with.

Have you ever bought work from a design agency? I’ll bet you weren’t 100% satisfied. I know because I used to buy work from design agencies and was never 100% satisfied with what I got.

Perhaps they didn’t do what they said they would do? The designs looked great in the presentation but terrible at launch? That increase in conversion never happened? Or you got to the end only to find the agency needed more money to finish the work?

All these things happened to me when I was a client.

I always blamed the agency. But while there are some awful charlatans out there, now that I do agency work myself I’ve found another reason why I might not have been 100% satisfied. One that’s a bit closer to home.

It’s not possible to get truly great work out of a user experience agency (no matter how good they are) without being a good client yourself.
 

Be clear about what you want

The first rule is to assume that we, the agency, are utterly stupid. Also that we are terrible mind-readers. Be extremely clear about what it is that you want us to do for you. Write it down in black and white and make sure we understand it.

If you don’t do this, we’ll spend all of your money doing something else, and you’ll be annoyed with us.

Tell us how much money you have

This sounds like a terrible idea to you. But you’re going to have to trust us to get great work and that starts with the money.

Tell us how much you have and we’ll design a project to get you fantastic results. Hide it away and we’ll respond by either scrimping where we don’t need to (compromising quality) or, more likely, by spending way too much early in the project (making it more expensive, lowering the quality, and in some cases even killing the whole project).

Send us the documents we need

What could be easier than emailing some files? It all sounds so simple. We meet, agree which documents we need, and you send them.

But, in a twist (which I find mind-numbingly bizarre) we almost NEVER get all these documents from you when we ask for them! And if we don’t get the things we need to do the job, the job won’t be any good.

Give us access to the people we need

Your people are all busy, we get that. They have a million things to do.

But we need to talk to them, at length, because they’re the ones that understand how everything works in your business. You don’t know how it all works (honestly). We don’t know how it all works (clearly).

And if the people making the new stuff don’t know how everything works, the new stuff won’t work.

Reply to our emails and phonecalls

This one sounds needy. Designer-no-mates waiting for the call back. But you would not believe how often we get ignored by our clients!

Again, we know you and your people are busy, but if you don’t get back to us it simply means we are spending your money working on things we’re not sure we should be working on. It’s a waste of your money.

Call us back. Please.

Review the designs we send you

Reviewing design is tricky. You never feel you have enough time. We always feel you’re taking too long and making too many changes.

There is no getting around this but it helps when we send you the designs on time (hold us to this!) and then you review them as planned. Otherwise we end up with feedback on feedback on feedback and everyone hates everyone.

Make binding decisions

This is my number one rule of being a good client.

Successful projects always have a client with the power to make binding decisions and the gumption to actually make them.

Without binding decisions about scope, direction and signoff projects descend into a spiral of mediocrity, over-runs and appalling compromise.

You will hate this when it happens, so don’t let it happen.

Don’t change your mind too often!

Everyone changes their mind. It’s human and natural.

But on a design project changing your mind leads to a cascade of rework that eats into the project budget and threatens delivery dates. It usually means more money, or less scope, or both.

So although every client has the right to change their mind as often as they like, the best clients rarely do.

Challenge our work and our processes

Good clients make us better designers by challenging our work, our methods, our outputs, our timings, our costs, our designs, our everything (we might grumble, but it’s good for us).

Not-so-good clients leave us to get on with things without ever challenging us and then moan about us afterwards when it’s too late.

Challenge early, challenge often, challenge loudly.

Don’t give us work that’s not important

The sting in the tail: all the rules above are useless if you are giving us work that you don’t care about.

You won’t send the documents, make the people available, review our work, make decisions or challenge us because you just don’t care.

User experience design is expensive. Don’t waste your money on unimportant projects that you don’t care about.

Good clients get great work

That’s it. It’s not hard to be a good client, but it is surprisingly rare. I can count on one hand the clients who have followed all of these rules.

But that handful of clients got incredible work. They saw their businesses improve dramatically. They got promoted, and promoted again, and promoted again.

What if you’re already doing this stuff and still getting bad work? Then you need a new user experience agency…

Let me know what you think on @myddelton. I wrote this working at cxpartners in London. Great projects are always a partnership between the agency and the client.

The change curve

Let me tell you a story about a thing called the change curve.

It’s a story I can only tell now, months after the events unfolded, because it’s taken that distance to work out what the hell was going on.

And to start we’re going to have to slip back in time…

A time before the fall

Rewind my life one year. I’d just survived the first 12 months of working for a proper UX agency. No one had found me out as an impostor. I’d done some good work and developed more as a user experience designer than I ever expected to in that time. In fact, almost exactly a year ago, I gave a talk at UX Lisbon to a group of designers about my approach to prototyping. It went down really well.

I was on top of the world and hungry for more responsibility.

In short succession I started line managing people for the first time in my life, got promoted to lead the user experience designers in our London office, and began leading a project team of multiple designers on a huge, complicated, high profile, and interesting project.

This sequence of events propelled me onto the change curve. So let’s quickly explain what the change curve is and then we can get on with the story.

The Kübler-Ross model

Elisabeth Kübler-Ross was a Swiss-American psychiatrist who did a lot of influential work related to death and dying in the 1960s. She was all about helping people to confront the reality of terminal illness and death, which is pretty high on my list of worthwhile life missions.

You already know Kübler-Ross’s work because it’s the five stages of grief. Denial. Anger. Bargaining. Depression. Acceptance. She unveiled these stages in 1969, but over time they’ve come to be used for much more than modelling the stages of grief.

These days the five stages have evolved into the change curve below, which is used as a method of helping people understand their reactions to significant change or upheaval.

Riding the curve

Before I dropped into the change curve I’d spent five years improving my design and research skills with tangible results. As my performance increased over time, so did my confidence in my skills and my future ability to keep improving. A year ago I was confident, happy and hopeful about the future. 

But that was all about to change.

First came the shock, a major change that sent me hurtling into the change curve. In my case this was taking responsibility for other people’s output, their happiness at work, and their development as designers. Up until that point I’d only ever had to focus on my own output, happiness and development - but when I started focusing on supporting other people, the tools and methods I had used on myself were clearly insufficient and I had very little idea what to do.

The denial came next. The honest truth is I never even registered the shock at the time. For months and months I thought I was doing a wonderful job when in fact the quality of my project work was slipping and I was treating my colleagues like children. The thing about denial is that it’s incredibly, invisibly powerful - you would think you’d spot it in yourself but you don’t. In the end it took some very direct feedback from my colleagues and my girlfriend to help me realise just how much I was struggling.

Then flowed the anger. And boy did I get angry. With the job. With the project. With my colleagues. With my girlfriend. With the prototype I was building. And, on one unforgettable late night, with a web font that simply would not work when I uploaded it to the server. The thing about the angry stage is it’s happening right when your performance is falling off a cliff, so it’s a powerful front to avoid taking responsibility for the way things are going wrong. It’s the part of the change curve that I am most ashamed of when I look back.

The depression quickly followed. For me this set in once the crunch project was over, and it sat with me for a while because I went away to New Zealand for a month. It was not my happiest holiday ever. During that time I started to glimpse what had really been going on: I was out of my depth and didn’t know what to do about it. They say that the depression stage is a kind of early-acceptance but one where you are still emotionally attached to the things you are accepting, which is what makes it so fucking difficult. I kind of feel like I don’t need to tell you about depression. It’s the very bottom of the change curve for a reason.

And, eventually, the acceptance. After New Zealand, and after another painful four weeks on the project, I started to privately and publicly accept the scale of my failure and own up to the many things I had done wrong. Honestly, this all started from some brave, direct feedback from my colleagues in a sprint retrospective. I gradually saw the scale of the task ahead of me, which was that taking responsibility for more than my own work required a radical rethink about how I approach the act of work in general. I stopped being in denial. I stopped being angry. And although it sounds like this would make you more depressed, actually holding my hands up and saying these things are my fault was a huge relief, so I left the depression behind as well.

Finally, integration. This is where you experiment with new approaches and then adopt the ones that help you raise your performance in your new, changed, situation. I’ve been doing a lot of thinking, reading, training, and experimenting in the art of leading teams. I’ve been doing a lot of listening too. Some of it works, some of it doesn’t, but my future feels interesting and hopeful again. It’s not the same future that I envisaged a year ago - lots more about working effectively with other people and lots less about interaction design patterns and research methods - but that’s the thing about change: it does involve making changes.

Mental models for mental problems

It’s been a rough year if I’m honest.

But I can pinpoint exactly when things turned around. At about 4.00pm on Thursday 16 April someone that I trust scribbled this change curve down on a piece of paper and said, look, Will, this is what you have been going through.

In an instant, my whole year made sense. I saw my life on that curve - shock, denial, anger, depression, acceptance - and it gave me a sudden burst of hope that there was a way forward.

I love a simple model that helps us understand a messy situation. I love being able to put the right words on complicated things in a way that brings a sudden clarity.

That’s why I wanted to share this story. Maybe it’ll do the same for you.

Thanks to everyone who went on this journey with me, but in particular to those who gave me the feedback I needed. You know who you are.