Tent Talks Featuring Caroline Jarrett: Who Cares About Forms? You Should—Here’s Why

[00:00:34] Chicago Camps: Could you share a common misconception about form design that you often encounter and explain the correct approach?

[00:00:40] Caroline Jarrett: I think that really goes to the heart of the, who cares about forms, why should you? Because people sometimes ask me what I do and I say to people, this is like everyday people in the real world, the people that we sometimes in UX refer to as users.

We say, okay, what sort of job do you do? And I say, I advise organizations on how to make forms better and easier to fill in. And they always say something to me like, “Oh, that’s good. You’ll never be out of work then. I didn’t know anybody did that.” And I think, yeah, I agree with you. I encounter crummy forms all day, every day.

Just today, I was trying to change my password on a website. And I discovered that some genius had designed their app so that the app would not accept spaces in passwords. Their website had changed, accepted the space in the password, but the app, no. Okay. So then I had to go back to the website, check the drill, and this is everyday life for nearly all of us.

And so that is like, why should you care about forms? Yet, your people will hate you if the forms are difficult, and yet everything we do is like mediated through the forms. But the common misconception seems to be among the people who I would describe as us, like not the everyday people, not the people out there.

The UX community, and especially that part of the UX community, or I might describe them as bosses or budget holders. It sometimes seems to me that the concept that forms could be improved is something that we understand, but is surgically removed from person’s brain when they’re given any money to spend.

I find it incredibly difficult to persuade anyone to give me any kind of fee or money for fixing their forms. Even though I can demonstrate stunning payback, like the amount of money you can save on a call center with a minor change to a form can be very impressive. But who cares about that because the call center people are on a different budget to the spending money on consultants budget.

Yeah. Did I sound passionate about this? That’s where I am often. Now that’s the biggest misconception I come across.

I think this is something that it’s bewildering to me, really, that. Something that is so necessary and essential just is overlooked. Recently, for example, I was helping a conference out by reviewing proposals submitted to the conference.

And one of the proposals that’s, it got me riled anyway, because it talked about human and non human touch points. And I resent an organization that thinks of its members of staff as a human touch point. I’m like, what? And they’re people, but there was this long list of touch points and it had brochures and it had advertising and it had everything apart from the one thing that is absolutely compulsory and it’s the forms.

Oh no, here we go again. Please, can’t anyone think of the forms? I’m here to plead with you. Think about the forms. Care about them.

Anyone can do a four. It’s like the same as the topic of my book. Anyone can do a survey. Bye. Anyone can do a form, there’s literally hundreds of thousands of form builders that will propose to you easy forms.

Anyone can put together a form. That’s not quite the same as putting together an easy form or a useful form or a form that works for someone who’s on, you just come back to the US from Scotland, over here we do a lot of train travel because we’ve got a pretty good train service. Okay, all of a sudden you’re on a train with intermittent dodgy Wi Fi and not much cellular service and now you’ve got an intermittent connection and it’s breaking up.

Oh, does your form freeze? Does it crash out? What goes wrong with it? So it’s like a form that is robust, that’s usable in different circumstances, that works for all sorts of people, not just people with brilliant vision who are in their twenties. Those are things which become a bit more complicated.

The other thing I do feel sorry about, as well, is that an awful lot of teams get thrown at them, “Oh, here is a form, digitize it,” right? And so I’ve had this discussion with people, I say, Oh, we’ve got to digitize it, but we’re not allowed to make any changes.

And I’d say, okay, one of the very first forms I worked on when we started doing everything like dominant internet compared to dominant paper. I’m rolling back to around 2000, 2001, that kind of era. And I remember being given this brief, Oh, we had to digitize it, but we weren’t allowed to make any changes.

And the first thing that came up on screen was “please write clearly in black ink.” I’m like, that’s not an inherently digital concept, writing clearly in black ink. You have to make some changes for the medium. And the minute you make any changes, you’re starting to change it. So where are the boundaries?

What can you digitize? What can you not digitize? And where did that form come from in the first place? Have people hated it for years? It may be a very bad and ineffective form. So one of the things that is consistently a problem for me is that many services, many forms have no idea what their error rates are.

But it, that the rule of thumb in the world of forms management is any even moderately complex form. That has not had a sustained, consistent user centered design, UX approach to it with lots of usability testing. Any like ordinary form out there in the world, when I look at it, it’s going to have around a hundred percent error rate.

That’s to say every paper form coming in will, I would expect to have at least one error. And it’s not unheard of to have error rates more than a hundred percent. It’s Oh, hang on a minute. You can’t go more than a hundred percent. Oh, yes, you can. If that form is going back, you have it’s coming into your organization and being posted back from all work, that’s a 200 percent error rate, and I’ve seen somewhat complex forms where you can have 600 percent error rate.

Like it’s taking people five, six attempts to get that thing even into being dealt with. But that’s terrible. And yet you come along and say, Oh, we’re going to magically digitize it. Here’s a pro tip. When you get a bad digital form, it’s probably that it was a bad form that was digitized. It’s not the medium, it’s the form that’s the problem.

So it’s the bad questions in a new medium will still be bad questions. I’m sounding incredibly old fashioned as if everything I’ve ever dealt with is paper. These days we are often purely digital. We’re not really thinking about paper routes at all. And the only challenge for us then is how do you actually find out what your error rates are?

So in fact, I probably ought to give a shout out to my friends at ZUKO, that’s Z U K O, Z U K O in American. They do analytics packages, which will show you where people are dropping out of your form. And there’s a kind of overlap into the whole thing of conversion. When people start, do they actually get through all the forms and end up making a purchase?

It’s just. Possible to do in digital, to track where people are falling off, just a bit harder. Whereas with paper, you can just look at the form and find out how much of a mess people are making of it.

[00:08:15] Chicago Camps: In your experience, how have modern technologies like AI and machine learning influenced the way forms are designed and utilized today?

[00:08:24] Caroline Jarrett: I love this question. People are asking me about how will AI change forms? And the reason I love it is because I got into forms in 1992, I started working with forms because of my experience in AI. You’re what, how could that possibly be? It’s because one of the artificial intelligence technologies that some of us know about now is like optical character recognition, otherwise known as handwriting recognition.

Or when you take a picture of a page of notes and you press the button and it turns it into text for you. That’s using optical character recognition technologies or intelligent character recognition. Those things are machine learning, they’re AI. When I was doing that stuff in the very early 1990s, we were all excited because it was moving from more of your basic machine learning type algorithms into using more neural networks, AI technology, and it was going to be great.

And it was all going to work even better and stuff like that. Put it like this. It didn’t work well in 92. What’s happened since then, it’s actually not changed that much. It still has the same problems that turned me from thinking, I’m going to use AI here into, I’ve got to make these forms easier to fill in.

And the problem was that when I went to look at the forms as to why the scanner and my super duper AI powered OCR technology wasn’t dealing with them, I discovered that. People were writing “see letter attached” on the form and AI in 1992 could not deal with that letter attached and AI today can’t deal with that letter attached.

You still need a human for it. So the AI has become in some ways cheaper, like all of us can fire up an instance of ChatGPT and play with it. In some ways it’s a lot more expensive because we’re killing the planet with the water and the electricity it uses, but that’s another story. But it’s now very accessible technology, but it’s not necessarily much better.

It’s still got a lot of those problems when we’re doing recognition technology. We still need to think about what sort of things do we have to refer to a human? What sort of things can we rely on the AI? We can’t rely on the AI, frankly. It’s basically a giant autocorrect that works on paragraphs, not words.

And anyone who absolutely outsources their chat text conversations to the autocorrect without ever thinking about what they’ve done is leading a very difficult and embarrassing life, I would guess. So we want to use the tools by all means. I’m not like a real dinosaur about this stuff because that’s how I got into the whole world of UX and forms.

And I do rely on my autocorrect. But I’m keeping an eye on it, I’m having a balance in terms of what sort of sensible things can we expect the AI stuff to do and what should a human really do. And I don’t know if you’ve seen, there’s a meme going around and I really should know who to credit this to, so I apologize.

But someone put up a note recently that said, it looks like AI is doing the writing and art for me, so I can do more folding of the laundry. And I’d rather that the AI did the laundry so I could do the art, thinking about what is life about, are we creating, not just trying to throw people out of work, but thinking about what good qualities they’re bringing to that work and which parts of this, does it make sense, to automate, and really understanding what the job really is.

I did not know that the job of typing stuff in from forms, was composed of hundreds and thousands of micro decisions about how to do a better job for a customer until I went to observe that job. And I think there’s lots and lots of jobs where we confuse the fact that we’re paying people a very small amount of money with the fact that they are actually using their skills as humans to make lots of micro judgments.

Another one is fruit picking, right? Fruit picking is a highly specialized job which it requires a lot of skill and training, but which is incredibly undervalued by society. We think it’s an unskilled job just because it’s really badly paid. And another example that really springs to mind, which often we don’t really think of as a form, but to me it’s a voice response form, is the process of actually having to call someone.

Do you remember that thing called a telephone, where we used to pick it up and we used to dial? And then someone at the other end would answer and we’d have a conversation. Okay. What’s happened is organizations don’t like paying call handlers. So they’ve tried to put an absolute mesh of press button B and go on our website and a zillion other things between you and a human on the grounds that they’ll save some time for the human.

But what it’s often doing to me, maybe I’m alone and I’m just irascible. It’s by the time I get to the human, I’m so angry that I’ve had to go through five or even 10 minutes of defeating press button B, that the poor human at the end is having me augmenting into the worst version of myself, but no one has thought about how can I actually employ more humans and deliver the customers to them in a better frame of mind, yeah, where is the augmentation?

Where is actually adding values through the technology for those luckless people who take call center jobs rather than just infuriating everybody. I can’t see where the cost savings are coming from by doing that. I really can’t.

[00:14:12] Chicago Camps: What are some key strategies you employ to ensure forms are user friendly while also achieving their intended purpose of collecting accurate and useful data?

[00:14:22] Caroline Jarrett: It’s like usability testing, basically just sitting and watching someone trying to fill in that form. Now, years ago, before we actually had Zoom meetings and all that cool stuff, we had to literally go and sit with someone and watch them fill it in, and that’s still a good thing to do.

I still think that doing observational studies of people in their own environment is a wonderful thing to do, especially when we’re trying to reach out to people from more disadvantaged communities, people in more difficult circumstances where they might not have great internet, people with various types of disabilities, actually getting yourself out of your comfy chair and your Zoom meeting into a space where they are can be incredibly instructive.

But a lot of this we can now do using our modern meeting technologies of one kind or another. Just say to someone, Hey, I’m going to ask you to try filling this form in, or you could say, Hey, I’d like you to apply for a loan or take out your car insurance or whatever it is that form is about. And just watch them, get them to tell you what they think about it.

Watch it line by line. See where with screen sharing now they can show you where they’re clicking or where they’re not clicking. And one of the things that you quickly learn is that the reading pattern for forms is very different. To the reading pattern for other types of material. And many of us have heard of the famous F shaped reading pattern, which is true where some content gets people’s eyes moving to it more easily.

That’s not the reading pattern at all for forms. The reading pattern for forms is, where’s that box for me to type into? Laser beam, eyes go, boom, to that box where I have to type. And then the eyes start circling around. All right, what’s the question I have to type here? And it’s a circular thing. And actually seeing where people don’t even see your instructions.

They don’t see the preamble. They don’t see that extra question you’ve put to the right. They’re just focusing on, let me get through this next question. And they don’t see the question where you’ve helpfully given them a default. Because they’re like, Oh, that’s filled in. I don’t need to do that. And they don’t read or see the placeholders or the labels in the box.

It’s like, why would anyone in their right mind put the label in the box where the poor person is trying to type. They click in the box, the label goes away. Now they’re confused. Crazy. You don’t do that. So you learn this stuff by just sitting with people or standing, if you’re a standing person, I don’t mind, but watching people do it, tell you what they find, where they’re having problems, where they’re not having problems most of us these days are pretty internet savvy.

If you do very boring predictable things they’re used to on websites, they’ll probably get through it. Is that question make sense to them in this context?

At the moment I’ve been concentrating a lot on asking the question about phone number. There’s nothing particularly difficult about asking someone for someone’s phone number. Apart from nobody wants to give you their phone number until they find out whether they’re going to be spammed on it.

So it was very interesting that, like the other day I was a bit stuck, I was about to do a meeting with a friend and we changed meeting technologies from one technology to another, and it just so happened that they hadn’t put the link for the meeting technology. And I tried to call them on my phone, it was, Oh, no, I don’t, I’m Gen Z. I don’t answer a phone number. I don’t know.

And yet enter your phone number is a very simple question. You only find these things out when you usability test that in your exact context with people who are trying to get the thing done. That the form is in the way of getting it done.

Yeah. Just, I love usability testing. I love it so much. The other thing that amazes me is so much fun. It’s the best. It’s fun in a terrifying way. Your cherished ideas get pulverized and, Oh, and another learning experience. But they’re pulverized in a way that helps you to make so much better decisions and improve quite quickly.

So that’s my absolute number one, get it out there and test it. And another use case, which I think is well worth people thinking about, is what are you going to do when that form is notionally being filled in by someone who is currently incapacitated, that they can’t actually do it themselves, but a relative or friend is got to do it for them.

How do they find out the answers? What do you do in those circumstances? Getting out to where people are, who don’t have the same level of privilege that we have is just a really, again, challenging, particularly the first time you do it. I’m not going to say you’re going to feel comfortable, but it’s a discomfort we have to get used to sitting with, because those are our people.

Those are the people we’re trying to make things straightforward for.

[00:19:23] Chicago Camps: Can you describe a particularly challenging project you’ve worked on involving form design and how you tackled the issues that arose?

[00:19:32] Caroline Jarrett: So one that’s really top of mind for me at the moment is that these days, I tend to work more with people who are creating design systems than with the forms.

Okay. So for example, I worked on what became the GOV.UK design system. And I worked on what became the NHS England design system for our health service here and a few others here and there. And so the challenge there is that we’ve developing really a lot of these design systems, including the GOV.UK one have got great components in them.

But how do we convey how to put those components together in ways that create good outcomes for our users? It’s not just a question, you can make beautiful nonsense if you’re not doing enough testing with your users, and if you’re not getting people to do those things around, for example, in the world of forms, one of the things that’s really crucial, if I go back to the phone number is to say, when you ask for a phone number, you really have to know how you’re going to use that phone number within your organization so that you can tell the person who’s going to trust you with their phone number, what’s going to happen with it?

That to me is a service design problem. Okay. And possibly a content design problem, like how do you write the words? But if someone is creating a product or service, they come to the design system. They want, Oh, I’m going to give people an open box.

Oh, there’s a sort of grab and go behavior where you come in, you grab the code you need, and you go back and you pop that code in wherever it’s going to go. And interrupting that journey to persuade the designer or developer to have a look at the other advice around it is a really challenging problem.

And I’ve observed the behavior myself. I’ve observed when I want to get something to pop into a prototype or website that I do exactly the same. I’m like, Oh, give it to me. I want to use it. I’m not in a mood to contemplate the other advice, and that’s something where, cool, I’m open to people discussing it and how do they manage it within their design systems and the advice they’re giving their clients and colleagues as how do we nurture busy designers and devs to, to take a moment to think more about context when they’re on deadlines, when they’ve got to get their jobs done and there’s that grab and go mentality going on.

And that to me is a problem that I really am thinking a lot about how do we create bits of guidance, bits of help, bits of material for our colleagues or designers and devs and managers come to that. How do we create the managers the space to ask their teams, what have you thought about? Did you look at, when do we create those moments to pause and reflect when busy people just need to get on?

So it’s perhaps it’s a slightly off answer to your question, but that’s like a current challenge for me.

[00:22:40] Chicago Camps: Looking ahead, what emerging trends or technologies do you believe will significantly impact how forms and surveys are created and used?

[00:22:49] Caroline Jarrett: I’m going to talk about surveys for a second because of the things I think we’ve all noticed in surveys is that they’re far too easy for organizations to send out.

We’re all bombarded with surveys and all the time you can barely buy anything or do anything on the internet now without being asked to review or rate or report. And what’s happened is that the method of doing a survey has become heavily overused and we’re being trained by all organizations all the time just to ignore it.

You have to be mad, cross in the American sense I meant, that if you’re cross with an organization or mad as I believe you say, then I’m going to tell them how much I’m fed up with it. But there’s also a thing like on eBay, like I either don’t give feedback or I give five stars because I know it’s going to really be very annoying for everybody if I don’t give five stars.

If I have a dispute, I’ll go directly to the person and review it. So the whole thing is becoming undermined because of overuse. Yeah. Now, that’s a survey thing, just massively overused. Where do I see we’re going on forms? I think one of the things I notice a lot in the world of forms is that browsers are trying to helpfully fill in forms a lot for us.

And I think that those kinds of agents where this is an AI thing, where an AI agent will come along and try and fill the form in for you, I think is going to be very prevalent. And then how do we actually let the user appropriately train and intervene on that browser thing that’s trying to second guess them.

So somehow or other the browser I use the most often has infected itself with the wrong spelling of my surname. This is really unhelpful behavior and I don’t know how to get rid of it. And then the other day, even worse, I had to renew my car insurance and somehow or other, like someone crashed into my car in December 2022, and somehow or other that information got transformed from when I put it in, and I know the right date, and I know I put in the right date.

But eventually the insurance company came back to me and basically said, you didn’t tell us about the accident you had in December 22. You only told us about an accident you had in December 23, on a different date. I’m like, no, I didn’t, but I haven’t taken the screenshots, which for some reason, I don’t know why, normally I’d take screenshots of everything.

I think that was a helpful browser agent. So we’ve got this potential for lots more errors creeping in because these cool, helpful assistants that I think a lot of us have started to rely on a bit, is handy when I click in an email box and the browser puts in the right email for me. But at what point does the error rate start to become really attritional for us all?

That’s something that I’m worried about at the moment.

[00:25:46] Chicago Camps: And now a question from our live studio audience. Neil asks, I’ve designed forms for really serious things where it is vital that people slow down and think about their answer. How can you, perhaps counterintuitively, add friction to a form in a way that is helpful and not just annoying?

[00:26:04] Caroline Jarrett: I think that’s a really cool question and it’s very much appropriate to something I was discussing today, which is topic of living wills, advanced directives, lasting power of attorney, end of life decision forms, which are deeply serious. You don’t just throw wills, you don’t just throw them off in a moment. You have to be extremely considered about them. And so building in some time for reflection and some time for really considering what’s going on.

For those forms, what typically organizations that publish them ask people to do is to explicitly have conversations with someone else about it. And I would really like to see now I think about it, an opportunity to say checklist, have you had a conversation with your loved ones about this topic?

When did you have the conversation? What was the main thing you learned from the conversation? So it’s almost like a journaling thing that people might have to go to. You can’t force people not to put in trivial answers, but at least the step would be there. And then another one that, that I have been writing on for a while.

So if you go onto my blog and look for Effortmark is my business. That’s E double F O R T M A R K. If you put in the word Effortmark and search for signatures and signing ceremonies, Signatures and signing ceremonies are like the traditional way in which we’ve asked people to pause and reflect.

What’s the difference between a signature and a signing ceremony? A signing ceremony is where you give people some sort of sequence of actions that’s designed to give them a pause. So it might be, okay, if you agree with everything you’ve done, please tick this box, and now I want you to type in your name.

So it’s emulating some of the formality of a signature without actually being the technicality thing of a digital signature. So there’s various technologies that emulate that these days, where it’s not what’s known specifically as a digital signature, but it’s a kind of signing ceremony process. So it’s thinking about how to create a pause in the transaction, which isn’t just an, are you sure?, where people are just going to click, yeah, I am sure.

Gives them a question that they have to consider and then create a response to. So it’s very contextual. And I agree when it’s a really serious thing, really serious things are really serious in ways that are different to each other. That’s the friction that I would add. It has to genuinely feel like kind of diversion.

And in general, we have not yet really, as an industry developed good patterns in our work for situations where more than one person is involved in a decision. We have a very kind of one on one concept and individualistic, possibly I might say that’s a cultural thing and stray off into another topic, but we have this individualistic concept of one person, one form, one computer, where it’s an individual’s responsibility.

For example, writing a will is typically not just your own responsibility. It actually reflects a whole lot of family and other human relationships that need more than one person involved. So that kind of workflow. Another one that I did a presentation with Karen Fernandez of Cambridge University last year, where she’s looking at applications for academic grants and Which are very rarely one researcher, one grant giving body.

It’s much more common for there to be consortia of researchers across different departments and a lot of coordination between different people about who is responsible for what aspects of the work. A lot of our patterns are not really standing up to that extremely well. The only one that kind of comes close is something that I worked on a bit with, again, UK Government Digital Service, which is a task list pattern where you explicitly front the form up with a list of tasks involved, and people have to go through those tasks as a sort of almost like a mini hub for their form experience.

And some of those steps could be offline ones of one kind or another. So that’s partly where my thinking is going on those more deeply complex type of forms.

[00:30:40] Chicago Camps: Another question from our live studio audience. Jane asks, can you talk about where the line is between a form and a survey? When do you realize people are talking about the wrong thing?

[00:30:51] Caroline Jarrett: Oh, Jane, I love that question and I’m so excited to see you here because I follow you on Blue Sky and it’s fun having you there on Blue Sky. Hi. And the answer is that to me, it’s a form when we intend to use the answer individually, one on one. And it’s a survey when we intend to use the answer in aggregate.

For example, I was looking for a client the other day, I was looking at a pricing product survey. They called it a survey where they were asking people, this was a product which has an annual purchase basis for various reasons, and they were sending people a survey because it was delivered in a survey tool and felt like a survey to them, asking each user about their purchase intentions this year. Okay.

So the answers were, yeah, I’ve bought it already. I’m not buying this year for whatever reason, not decided. And then they intended to do follow up based on those individual purchase decisions. So to me, even though it was built in a survey tool and delivered as a survey and kind of optional, like a survey, to me, it was much more of a form because they intended to use each answer individually.

Now you might say, actually, there’s often a bit of an overlap, like quite possibly with that purchase survey thing, they were probably producing report for a manager. To say, Oh, this year we’ve achieved 27 percent more purchase indications at this point in the cycle than we did last year.

Yeah. Okay. So sometimes you can get a bit of value out of aggregating forms, answers. That’s to me is how I think about them. And so I’m looking through and people send me some kind of, as the survey methodologists would call it an instrument, but we might talk about a form or questionnaire looking at it.

So how are you planning to use these answers? Are you going to use them individually? Or aggregate to make your decisions. That’s what creates the distinction.

Having said that, that may not be what the customer or the user or the person filling it in thinks, they can make their own judgment. They might say, Oh yeah, this looks like a form because I’m trying to get something done and it’s in my way.

Or this looks like a survey. It doesn’t look like I have to do it. So we can have different views internally and externally.