Samuel Hulick is a user experience designer and runs useronboard.com. He is currently writing a book called User Onboarding. In this interview he talks about common UX mistakes, Jobs-To-Be-Done, and lessons learned from studying user onboarding processes.
If you would, tell us a little bit about yourself.
I’ve been a UX consultant for quite a while now. I did a stint at a local startup for a couple of years as the customer success lead.
As far as onboarding expertise—that was something that I really lived in at that company. It just wasn’t called user onboarding. It was more about seeing how people were able to log in and get through the setup steps, and how long it took.
Since it was more enterprisey, it was a lot more hands on. It literally involved watching people progress through the step. If they took too long, I’d send them an email. It was all hands on high touch kind of stuff.
Comparing business to business product and a business to consumer product, what are some of the differences that you see in the user onboarding processes?
Onboarding is such an inherently metric space. When you look at a lot of UX-oriented projects or engagements it is more about making the site more fun or more pleasant. When you’re specifically looking to apply UX to onboarding it is easy to say this is our conversion rate of trials to paying customers and this is how well paying customers tend to stick around.
There is a lot of natural overlap with churn rates, conversion rates, time to do adoption, time to activation, and other things like that. It allows you to look at the numbers in a really clean way.
User onboarding is about getting people to shoot fireballs as quickly as possible, not getting them the flower.
You asked about B2B, B2C and how those are different for onboarding. B2C is going to have a much wider funnel. You’re going to have a lot more numbers, and those numbers are going to have a lot more fidelity. If you want to run test, they won’t take as long. B2B, is more tied to money, which means the metrics are more focused on return on investments.
Let’s talks about useronboard.com. I think it’s a fantastic project. Why don’t you tell us a bit about it?
Right now, useronboard.com features slideshow tear downs of popular SaaS products. I make them by recording myself stepping through the process. Then I add live commentary. I’ll say things like “This is the first thing that I noticed and that’s good because it’s putting me in the right mindset.”
I’ll go through and take screenshots of all of those different steps. Then, I turn my commentary into annotations and try to come up with some major takeaways. Finally, I turn it into a slideshow and post it to the site.
The response has been really, really cool.
What you are putting out there is incredibly useful for founders and designers. From all the product teardowns you’ve done so far, what are the key lessons learned?
One of the biggest ones is preserving people’s attention throughout the entire process. I like to think of attention as a fixed quantity—like gas in a car. If you’re trying to get somewhere and you run out of gas, you would just run out of attention. If you don’t get to where you’re supposed to by the time that happens, then that’s not going to be a successful onboarding experience.
Show people the exact things that they need, and make the process as straightforward as possible, and no matter how much gas they have, they’ll get to where they need to go.
Have you seen any patterns or specific flows that you’ve seen that work really well?
I have nicknames for some of them. As I do more of them, I’d like to formalize them into a pattern library or something similar.
There’s two that really stand out. There’s the foot in the door technique where you’ll see “Sign Up Now” with name and email address field. You type your information in and then it takes you to a form that has six or seven fields, but those fields are preloaded with the information you already entered.
Not really specific to user onboarding, but one thing I constantly encounter the forgotten password pattern. You’ll enter your information a couple of times, and it doesn’t work. I click the forget password button and then there’s just an empty field that says enter your email address. That field shouldn’t be empty. I already entered my information three times.
Little tiny things like that reduce the cognitive load that it takes for people to accomplish what they’re looking to do.
As a suggestion for useronboard.com, I’d like to see the 500px lazy registration process. You get to use the product before you actually have to really sign up or put in your email. Have you looked at any products that use lazy registration? What do you think about the user experience with those?
I’m generally for it.
I think one really great example is Brennan Dunn’s Planscope. It’s a hybrid premium model that lets you start with a sandbox project right away.
As a general rule, only ask people to provide what they absolutely need.
On the flip side, what common UX mistakes you see constantly?
I hate submitting forms and not knowing what’s happening. You click the submit button and everything just sits there. You can have latency of three or four seconds, and then all of a sudden the page changes.
It’s a small UX detail, but ghosting the submit button, or putting a spinner inside it or having some copy that says we’re sending your information—I just want you to tell me what you’re doing.
The important thing about measurement isn’t the thing you’re measuring, it’s asking the right questions to begin with.
Another common UX issue is how blank states are handled. When you start using a product for the first time they take you to a dashboard and you start with no notifications or have no other other friends. It’s almost accusatory, and it doesn’t really show you how the product works.
I’ve seen this handled in two ways. You can pre-populate the account with dummy data, which I’m not totally against, but it doesn’t help people get a feel for the product. The other is a joyride where they’re basically putting labels on the interface that give directions telling you what each thing does.
Personally, I have not seen those work really well. Anecdotally, this has always been really frustrating to me. If your interface is confusing to people, I don’t think the answer is more adding more if it.
A long time ago Jason Fried posted a Flicker group called Signs on Signs. Somebody would have a sign that says watch your step, and then there’s another sign that says look at the watch your step sign. It’s a funny concept. They are literally pointing at bad designs.
If the sign is not working, just change the sign so that it works. Don’t put another sign pointing to the sign that doesn’t work.
That’s great. I love that.
Pre-populating information works for Basecamp, Trello and other products that are built on lists. How do you handle more complicated products such as YouTube?
Let me take a step back.
I think that the essence of onboarding is a two-step process. You have a marketing site where you’re just trying to communicate the value that you provide to people. You want to get them to sign up and that is when you’re entering into onboarding mode.
I would not mistake user onboarding with activating features. The onboarding process is completed when somebody receives the value they thought they were getting when they signed up.
For example, I just did a teardown on Vimeo’s onboarding experience. They’re establishing the value is that you’ll get to host your videos in this beautiful format, and share them with the people you love. That’s the goal.
As soon as I start the onboarding process, the quicker I receive the value that was promised, the more successful the onboarding experience.
So, the key to successful onboarding is to show the value of the product?
I would even take a step further back. I’m not looking at it from a product centric standpoint. It’s a slippery slope to say somebody has to do these six activities because that’s what our product requires.
What’s the shortest pass to get a user to think that this product is going to make them more awesome at a specific task?
Did you play Super Mario on Nintendo as a kid?
You see a flower pop up, you’re like, “Awesome!” Once I get that flower, I’ll be able to shoot fireballs and it is going to be rad.
A lot of products value propositions and things like that. Really talk about how awesome the user is going to be when you use the flower. The fireball is the thing that you should be really selling.
User onboarding is about getting people to shoot fireballs as quickly as possible, not getting them the flower.
That might be the best analogy I’ve ever had on this show. That’s awesome!
Let’s talk a little bit about user experience versus conversion design. Your blog on medium is called Designing for Results. Where should the line be drawn between conversion testing and user experience design?
What specifically do you mean by conversion testing?
A/B testing—this button should be blue instead of green. Is there a point where conversion testing might become deceptive, and take away from the user experience?
I would say that not to mistake the forest for the trees. I don’t think that they are separate disciplines. Scientific design is a really intelligent way to go about providing a better user experience.
If you aren’t paying attention to the impact of your recommendations, I don’t know how you can improve as a UX designer. I don’t know how you can have a high degree of confidence in the recommendations that you’re making on an ongoing basis.
To me, the user experience medium is people’s behavior—the behavior of people using the site, not the pixels of the site itself. Measuring and trying to make that behavior visible is a requirement.
I’ve got an example in mind that I think shows the difference between conversions and user experience, email sign up modals. People hate them. They provide a terrible user experience. However, I’ve done a good bit of conversion testing, and they perform incredibly well.
What is your opinion of the email sign up modal?
I think that it’s a question of what you value.
There’s short-term optimization and long-term optimization. Maybe you’re comfortable annoying people to get their email address. Maybe you can nurture a long-term relationship from there.
I think there’s a saying—the important thing about measurement isn’t the thing you’re measuring, it’s asking the right questions to begin with.
If you aren’t paying attention to the impact of your recommendations, I don’t know how you can improve as a UX designer.
The question in this example is if pop-up modals collect more email addresses or sign ups on page load. What about other metrics? Are you measuring how many of those people are sticking around for six months? Are others recommending this site to their friends?
You can create an experiment that will tell you something about anything that you’re doing. The key question is “What do you really want to know?” The next question is “How do you build an experiment that will show you what you want to know?”
I think that is fair. Perhaps modals drive signups in the short term, but cause poor retention.
Personally, I don’t like them. I’ve thought about doing it here on Signal Tower, but I feel like that impedes what people came there for. I figure if the content is valuable, they’ll subscribe anyways.
I made a very similar decision with useronboard.com. A big reason it exist is to build audience and get email subscriptions and things like that. It certainly occurred to me like having a persistent sign up form that was visible the entire time while you’re going through the slide show would probably lead to more sign ups.
I want to provide the best possible experience. I don’t want any distracting elements on the page. I also feel more comfortable doing something that’s align with my values. You’re asking someone to rank the experience before they’ve had a chance to look at it.
There’s a lot of people listening that are running startups. What kind of metrics or tests should early stage startups be putting together?
Early stage startups that are likely not going to have the traffic where you can roll out maybe test and get results back within a week or two. I think it is better for early-stage companies to look for more qualitative results than quantitative ones.
Go out and talking to people. Get a real clear understanding of the intent and the value that people perceive when they come to the site. Startups should be a lot more hands on, and lot more face-to-face than just putting out a survey.
So you need to go out and ask questions. What type of questions would you go out and ask potential customers? How do you begin that conversation?
My general recommendation is to ask as few as possible.
If you’re doing something as bold as starting up your own business, you’re probably pretty headstrong. You probably feel some pretty strong convictions that you’re on to something good.
It’s really a question of can we validate or invalidate our current leads more than anything. Try not to ask leading questions, especially if they’re your friends or people that you have some association with.
If you’re conducting an interview and you’re saying “Hey, I’m thinking about making this thing, you’d pay $40 a month for it, right?”
It’s a loaded question. You are putting words in their mouth.
You are selling, you’re not interviewing them.
It is important not trying to persuade people into thinking what you think. You’re really trying to get people to frame responses in their words.
Other than useronboard.com, what resources would you recommend for people that are interested in getting in to user experiment design?
I came across the two things that had the most impact through Ryan Singer of 37 Signals. One is the Jobs-To-Be-Done framework, which I can probably talk all day about. Getting a really strong understanding of the Jobs-To-Be-Done framework has certainly really informed my approach to UX.
Another eye opener for me is a talk by Kathy Sierra from 2009. I think it is called Creating Awesome Users or something similar. The general concept is if you want to create people who are evangelists for your company, just make them really good at the thing that your company exist to make people good at, all the rest will take care of itself.
It’s an hour long. I don’t normally tell people to watch an hour-long video, but that one I send every opportunity that comes up.
Let’s talk a little bit about Jobs-To-Be-Done. Everyone looks at the 37signals, because they take experience design very seriously. Could you talk a little bit about Jobs-To-Be-Done?
Sure. I think that the milkshake example is probably the most famous one, and that’s probably the easiest way of wrapping your hand around it.
The story goes that there is this fast food company that wanted to sell more milkshakes. They would get consumer panels together and ask them if they wanted them to be chocolatier, or thicker, or have different things in them, or whatever. They would get people’s opinion and then, go and make changes to their milkshakes.
None of the changes were working. So they did competitor analysis to see what the other fast food chains were doing. They tried to copy them, but that didn’t work.
Then they hired a consultant and he went out to one of their locations. He just stood in the restaurant all day long and recorded anytime somebody bought a milkshake. It turned out that half of the milkshakes were sold before 9am. This is crazy. Why would people be purchasing a milkshake in the morning? They send the guy back.
The next day, he stood outside the restaurant and took note of the milkshakes sold before 9am. The people were always alone, and got them to go. Anytime an individual person with a milkshake would walk out of the store, he would ask “What job are you hiring that milkshake to do?”
What they said was “I have this really long, boring commute. I don’t want to be really hungry before lunch comes along, so it’s nice to have this thing that I can hold in one hand. Plus it staves off hunger.”
The fast food company realized that this was the job people were hiring the milkshake to do, and completely reformulated their entire product around it. If it lasted really long, they would be better. So they made the milkshakes thicker instead of thinner.
Clayton Christensen is the name of the person behind the concept. He has similar quote along those lines that “being 40-year old white male does not cause me to buy the New York Times.”
If you’re building your demographics around attributes like age and income, you should be building your demographic around people who have problem X. It makes it a lot easier to align your entire business model, product, marketing, everything.
Basecamp’s not actually competing with other project management software. People are emailing themselves tasks and using email to communicate with one another—the same job as Basecamp. Thus, email is their real competitor.
Do you have any other thoughts or last comments to leave the audience with?
In your other interviews you ask what advice would you give to entrepreneurs or something along those lines…
[laughter] You can answer that question then.
I would say that my advice is to really learn to love being rejected, as weird as that sounds. Years ago I read a Chris Dixon blog post sensationally titled If you aren’t getting rejected on a daily basis, your goals aren’t ambitious enough.
Rejection doesn’t sound like a lot of fun but, only two good things can come from it.
If it turns out you’re not rejected, whatever awesome thing you thought you’re going to be rejected for happens. If you are rejected, it gives you an indication of what you can improve upon.
From a career standpoint and also from a business standpoint—putting yourself out there and seeing what happens is something that I really didn’t used to do and now that I am, I’m finding a lot of success.
It’s just like questioning, going back to validating, invalidating. Stop speculating on things. Just get out there and see what actually works and what doesn’t.
Samuel Hulick, where can people find you?
Did you enjoy this interview? Subscribe for more interviews with more designers like Samuel Hulick.