Sean Hsu walks through the design thinking process for building products from scratch: start with customer interviews to find real pain points, build and test a minimum viable product, and iterate based on feedback. He shares concrete tools like Mechanical Turk and SurveyMonkey, interview techniques to uncover authentic problems, and his own experience designing a product concept in ten days that landed a Y Combinator interview.
Chapters
Pretty much it I want to talk on TV. So today I want to just share about my experience with design thinking, where the phase you guys are, it's going to be like the whole process starts like this and then go all the way like here. You're just finding ways. And then sometimes you go backwards. And then you keep going. And then at some point you will start narrowing down and then to a point that you find out a plan to execute. This is typically what I feel from—
Can you use a darker marker?
Oh sure, thank you. What I feel from 0 to 1 is typically like from here to here. You first get one customer that they They really care about the value you bring to them. And at the same time, after that you can scale. So things here are like— have you guys had any other design thinking class before or like— no? So I just talk whatever I want?
Yeah. Okay.
The first thing I would definitely do is start with customer point of view to find out what are the pain points. Among 1,000 pain points, everyone are just bitching about the things they hate. And you can find out like, okay, maybe our technology can solve one of them. So you start with the idea of using— we call this like—
Matching.
technology, we build MVP, like most valuable product. Not MBA, MVP. Totally different. And then you just keep doing customer research. So more like user study. And then here it's almost like prototyping. And then you kind of You kind of lock down 3 more prototypes and then you lock down with one and then get like a PO or like a cash reward here or someone started to swipe their card online. And that's how you kind of feel I'm really validating this product to market fit and then I just keep going this and scale. So, but if in terms of company growth, you'll feel like this. It's all flat until here you start growing. So that's like, it's taking typically 12 months to actually get the first order.
So—
So who's buying the people and kind of that type of company?
Who's buying?
So this is like the first money you actually generate for the company. Like either from Kickstarter you can get paid from Kickstarter or you can run your preorder event online. So from here, typically what I did was from point of view, do you guys know where to get— I want to share some tools to you guys. Do you know how to get point of views besides asking your classmates?
Online surveys.
Online surveys. What kind of surveys are you using?
Mac CRM.
Mac CRM? CRM. This?
Yeah.
What else have you guys been using?
Proximity surveys.
Which brand?
Face-to-face.
Where did you get the customers?
Mechanical Turk.
Mechanical Turk is a good one. What's that?
Craigslist.
Craigslist is— I like Craigslist a lot.
LinkedIn, 42.
I think LinkedIn is good too. SurveyMonkey is good too. Have any one of you paid SurveyMonkey and buy a survey from them yet?
No, I just know that.
OK. How long is your entire program? 3 months? 2 months?
6 weeks.
6 weeks. OK, that's enough. Actually, so I'm going to share all of the tools that I use and the methods. And I can tell you that I spent 10 days to design a product concept, and I did all of the steps in a very fast And then I actually got selected by Y Combinator to get interviewed with their investors, and then they really want to invest in my product. The only reason I don't want to do it is I don't feel like 10 days is enough for a maturity of the project. But basically I found out if we start with here, right? So I can tell you all the pros and cons. I don't use Mac. CRM a lot, but mechanical terms, there's a— if you are developing a product for like household use, like housewife or someone who live at house a lot, then this is a perfect place. So for example, if I'm— I was designing a washing machine to replace hand wash. In that design process, I use a lot of here because those are like typically like—
Mechanical terms.
Pregnant wives, they are staying home or new moms, they are just staying home. So new moms are also very good if you are developing products for new moms. This is ideal. Most of the costs are the same, like 1 per survey. So it's kind of OK. This is also 1 per survey. This is sometimes like 10. Craigslist is sometimes hard because it's more broader scope. If you're looking for— so here if you're looking for age from like 20 to 40, you can use Mechanical Turk. Craigslist I feel like is typically 25 to 35. It's like a very niche range of people who are spending time online. And then LinkedIn, I think LinkedIn is weird. It all varies from like 20 to— 50, it's like a very wide range. SurveyMonkey, there's a pretty wide range. I tried to buy survey from 20 to like no limit and it end up half of my survey result is from 65 plus and then half of them are from like I think, 20 to 30 or something. So people basically are just not doing work and they want to get some other cash. So this is a good mindset that if you are developing an elderly product, you can go to SurveyMonkey and it's very cheap for you to buy the ad or buy a survey from them. So the cost here I'm trying to tell you is that this is the cost of— The actual person filling up your survey. This person, you have no clue what their background is. But they can tell you their race. They can tell you either they're Caucasian or Asian or others. They can tell you their income. They can tell you their location, these kind of details. Yeah, so I'm just going to tell you all of the content. There's no bullshit or no jokes in my my training, so it's going to be a little bit boring if you're going to sleep. But there's going to be a lot of content here. So for SurveyMonkey, there's also another trap. It's that not all of the users actually report their background honestly. So you typically find some difference. It's very obvious that if this guy can earn more than Let's say $200,000 per year. They don't care if the product is $200 or $100. But sometimes if I have a trap, it's like you prefer this product to be below $50, but they are earning more than $200,000 a year. I feel like this is not a valid survey. I'll just take that out of my count. So after this phase, you can actually find out what is their daily routine or what are the Pain points. Like another example from my experience, when I ask them what are the pain points in laundry, I will not just ask them what are your pain points in laundry. I will walk over the steps from can you tell me what was your last time to do your laundry? And then they will start to remember. We have to help them to paint the picture in their mind. So they will start to remember as like, OK, last time I did the laundry was maybe last weekend, last Saturday. And I'll also follow up with a question that, was it a sunny day or was it just a rainy day or something like that? So they will get more engaged with that actual question and they will feel like they are back to the Saturday. And go step by step. What's the first thing you do? What's the second thing you do? What's the third thing you do? whole process filled in. And now I started to go into what are the specific questions or pain points that actually bothers them. I don't know if I turn it on.
It's on.
It's on?
OK.
So then in that kind of mindset, they will already be building with the image and they'll just feel like they just finished their laundry Like 2 minutes ago. And then they'll tell you what exactly their pain point is. Besides that, I also already have my list about like I have 5 assumptions or hypotheses that I feel like they could have pain points. Like when women separate their undergarments or delicate dresses, will they actually do that? And I'll just Follow-up questions after they complain about everything. And so always keep your questions as just a list of their daily use and then go to your list afterward. Never start with the list because start with the list is very hard to actually get engagement. They'll just try to bullshit you and just spend 5 minutes with you and then get the cash and go.
