Adora Chung (part 1): product and honesty curve

Original author: Adora Cheung
  • Transfer


Stanford course CS183B: How to start a startup . Started in 2012 under the leadership of Peter Thiel. In the fall of 2014, a new series of lectures by leading entrepreneurs and experts of Y Combinator took place:


First part of the course

Thanks for inviting me. Today I am going to talk about how to go from scratch to acquire a large number of customers. I believe that you already have many great ideas, and you are thinking about what your next step will be.

I prepared the text of the speech this morning and almost everything I will talk about is based on the mistakes that I made in the past. So, as Sam noted, I joined Y Combinator in 2010: I doubted for three years, dropped out and started all over again, and I determined what I wouldn’t do if I started a new project after Homejoy. Most of the tips are the result of failures, and I'll just tell you what you shouldn't do and summarize what you need to do.

I must say that you should not take all the advice as a guide to action, since each business is different from the others. You are different from me, and I - from you, keep this in mind.



When you start a startup, you need to have a lot of time available to concentrate on it. I’m not saying that you should skip classes or quit work, I’m saying that you should have a lot of time to devote to the idea and immerse yourself in it, look for solutions and try to solve problems. For example, if you are studying, it would be better to devote one or two days a week to work on an idea than to spend two hours here and there every day.

Since our course is technical, we can draw a parallel with programming. The ability to fully concentrate and immerse yourself in work is very important due to the large number of distractions.



As I mentioned, when I recorded all this, I thought that most people, when creating a startup, are doing wrong. Beginners think: “I have a great idea, so I won’t tell anyone about it. I’ll work, work, work, and then maybe I’ll tell one or two people about it, and then I will publish it on TechCrunch or somewhere else, and I will have many clients. ”

But what really happens is a consequence of the lack of feedback. Maybe you will attract many visitors to your site, but none of them will stay there for a long time, and this is because you did not work with them properly. If you are lucky enough and have a bank account, you can “buy” customers, but they will leave in time, and you will give up.

This is a kind of vicious circle. Once I already did that, and then I repeated my mistake while I was at YC. When I went through YC school, by the results I didn’t even launch the product. I did not publish the release on TechCrunch, and it definitely needs to be done. You should not fall into such a loop because it will not end in anything good.



The next point: you have an idea, and you should seriously think about what problem it actually solves. What is this problem? You should be able to describe the problem you are trying to solve in one sentence. And then think: “Does this problem relate to me? Does she really care about me? ” And again: “Let it be a problem that excites me, but does it excite others?” This can be verified by interviewing others.

One of the biggest mistakes that I have made affects me and my cofounder - my brother. In 2009 or 2010, we founded the company and named it Pathjoy. We pursued two goals. The first was to create a company that can make people truly happy, the second is to create a company that would be very, very effective.

A good solution is to simply create a large company. The problem we are solving is how to make people happier. At first, we came to the question: who are the people who make people happier? And they decided that these are personal growth instructors and therapists. Obviously, we needed to create a platform for personal trainers and therapists.

As a result, when we started using our own product - and we are not cynical, in any case - it turned out that therapists and instructors are not the people with whom we would like to deal. Everything that we created turned out to be useless to us. So, we did not try to solve our own problem, we were not sincerely passionate about what we were doing, and yet we spent almost a year trying to implement this project. And if you start from scratch and think about it before embarking on the creation of any product, you, it seems to me, will avoid a big headache and will not do what you do not want to do.



Let's just say that you have a problem, and you are confident in it, so where do you start, and how do you think to solve it? First of all, you should think about the industry in which you are going to do business. Whether it’s a large industry or even just a huge one, you must immerse yourself in it. There are several ways to do this.

One way is to briefly become a “cog” in a large production machine. This may seem somewhat illogical, because it is believed that if you really want to bring something new to the industry, you should not plunge into it for a long time.

Those who spend 20 or 30 years working in one place, follow the usual routes, are happy with how things are going, and never think about what may be ineffective in their work or things that can “break” the familiar pattern . However, as a beginner, you, plunging into the industry, should spend only one or two months just to understand the components of this system and the principle of their work. Because, delving into the details, you begin to notice that you can use something that works inefficiently and costs too much, and how you can save on it.

As an example: the story of Homejoy
When we launched Homejoy, we started with the cleaning industry and decided to work as cleaners ourselves. We started cleaning the house and very quickly found that, as cleaners, we were so-so. As it turned out, we needed to learn more about cleaning, so we went and bought books that helped, but just a little bit. We learned more about cleaning accessories, but it’s akin to playing basketball: you can read and learn the rules, but you won’t play better if you don’t practice throwing the ball in the ring.

And we decided that one of us would go to learn how to clean. Or, in any case, learn from a professional. Surprisingly, we began to look for work in cleaning companies. Of course, it's great that I learned how to clean up in just a couple of weeks that I spent at work, but even cooler is that I learned a lot about the work of local cleaning companies. This helped me understand why local cleaning companies cannot grow to today's Homejoy size. This is because they are quite old-fashioned and do many things inefficiently. For example, customer registration and optimization of the cleaning schedule are implemented there extremely inefficiently.

If your situation is similar to mine, and in the business that you want to do, there is an element of the provision of services, you must learn to provide these services yourself. If your idea is related to the restaurant business, you should become a waiter, if it is connected with drawing - an artist: you need to look at your customers from all possible sides of the business that you are trying to build.

Among other things, you should be slightly "crazy" in your work. You need to be obsessed enough to strive to learn everything that others in your industry do. This also includes compiling a list of all potential competitors and similar companies. You must click on each link and read each article about them from the first in the search results and the thousandth.

I found all potential competitors, large and small, and if these were public companies, I read their quarterly financial statements and attended their newsgroups. Most of this you already know, and you won’t extract anything important for yourself, but among all this information, insights will begin to come across to you with a certain frequency. But you will not be able to find them until you yourself go through all this and put this information in your head.

You must become an expert in your field. When you start working on your project, no one should have doubts that you are an expert - then people will trust you.

Next, you will need to determine the target audience. Ideally, you should create a product or service that absolutely everyone will use. But in real life in the beginning you just want to reach a certain part of potential customers in order to further optimize your business for them. It's all about focusing on the target audience, whether you are serving teenage girls or mothers of guys from a football team in a restaurant: you should be able to concentrate on their needs.

And finally, before creating a product or starting to write program code, you should describe the intended UX: how the user solves his problem with your help. To do this, it’s not enough just to create a website, you need to make sure that buyers know about you. It can be an advertising campaign or just communication with those who could visit your site and learn more about you.

  • What will be written on the site, how will you communicate with customers when they register or purchase a service?
  • What do they actually get from your service or product?
  • When they stop using the product or service, should they leave feedback or comments?

You have to go through all this and imagine in your head what the ideal user experience with your product will be. Then write it down, use it when writing code, and start creating the product from this very point.



So, you have an idea, you partially know the target audience that you want to reach, and you know everything about your industry: what do you need to do next? You start to create your product. Today, most say: "You have to create a minimally viable product."

I emphasized “viable” because I believe that many people miss this part and immediately begin to create “additional options” for the product, which is poorly perceived by customers. A minimum viable product is the smallest set of functions that must be implemented to solve a problem. I think if you look at already running projects, you will understand this very quickly. But then again, you have to talk to customers, you have to keep track of what already exists in your field and what you create should correspond to the current needs of the audience.

Also, before you offer something to the buyer, you must understand the positioning. By this I mean that when you approach a person, you should be able to say: “Look, this thing does this, this, and that” in one sentence. For example, at Homejoy, we started with something very complicated. We were an online platform for providing home-based services; their list began with cleaning. Then it was possible to choose one, another, third. The entire description was stretched over several paragraphs.

When potential customers visited our site, they became bored after the first few offers. So we realized that we need a short slogan. Creating a slogan is a very important component. It describes the functional virtues of what you do.

When in the future you will form your brand, you should be able to describe the advantages of the product in terms of user emotions. But if you start from scratch, you need to tell them what they will get as a result. After we changed the positioning and began to tell us that we do the cleaning for $ 20 per hour, our offer immediately became clear to users, and we had customers.



So, you figured out what MVP is: how now to attract the first customers? The first few users should obviously be the people you know. You and your co-founder should use the product, you should involve your parents, friends and colleagues. This will help get even more feedback.

I made a list of the most obvious places to go, depending on what you are selling. You can choose one of them. Use online communities: Hacker News now has the Show HN block - it's a great place for this kind of thing. Especially if you are creating developer tools or something like that. Work with local communities if you produce consumer goods. There are many local community mailing lists where members can become opinion leaders about your product. Especially the community for parents — they might also be worth it.

At Homejoy, we tried all of these options.
We used our services on our own - this is normal. We were the only cleaners, so it was pretty easy for us. Our parents lived in Milwaukee, and our office was located in Mountain View, so we could not go to them. Friends and colleagues were scattered in different cities - someone lived in San Francisco, someone else, so that we could only count on a few of them.

It turned out that we were at an impasse due to the fact that at first we could not convince a sufficient number of people to use our services. What we did: we were in Mountain View, and some probably know that on Bastro Street street bazaars open there in the summer. We went outside and literally chased people and tried to get them to schedule an appointment. Almost everyone said no until one day everything changed due to the weather. It was a very hot and damp day, and we noticed that on such days more and more people are reaching out to shops with food and drink.

We realized that we needed to be right in the middle of the crowd, so we stuck bottles of water, cooled it, and began to distribute them. And people just reached out to us. Having done this, I realized that people began to schedule cleaning only out of gratitude. But as it turned out, they did not refuse our services. Well, some refused, but most took advantage of our help. I thought that this was a way out of this situation - let me then have to clean up these people, but we solved our problem.

One of the startups of the last set of YC, whose founders were engaged in the sale of packaging goods for transportation, acted as follows: they searched the post offices for those who wanted to send the parcel, removed one of them from the queue, offered this person to use their goods and then sent a parcel for him. So, you just need to find crowded places. Your customer turnover will be very small, but to move from zero to one, three customers, and so on, such things are simply necessary.



Now you have clients, so what do you do with them? First of all, you need to make sure that people have a way to contact you. Ideally, this is a phone number, and if you register it, you should make sure that you have voice mail so that you do not have to pick up the phone for each call.

In any case, the ability to receive feedback is good, but in fact you have to communicate, talk with customers. Get up from your desktop, go and do it. At first glance, this is not easy - it is really hard, grueling work, but that’s how you get the most comprehensive feedback about your project. And that’s how you will understand which options or services you need to completely change, which ones to remove or add.

Another way to get feedback is to send interviewers to conduct a survey among people who have used the product. This is in the order of things, but more often than not, people agree to leave reviews only if they really love you or strongly hate you. And you will never get a more balanced opinion.

To collect additional information, you need to meet with the client yourself. I saw people who met with clients and interrogated them, both in the laboratory or at the Inquisition court - this will not give you good results. In fact, you should build a conversation, try to get to know them, make them feel comfortable.

You must gain their trust in order to improve something. I realized that taking someone to a bar, for example, is a great way to achieve such a confidential atmosphere. I'm not sure that all of you are old enough for such maneuvers, but you can invite them at least for a cup of coffee.

Another indicator you should keep track of is how things are going in terms of perspective. The best way to do this is to record the number of your customers. The number of those who came to you during the day, the number of people who will return tomorrow, the day after tomorrow, and so on. Usually, over time, you look at the number of visitors and notice that those who come today return in the next month, again and again.

The main problem that arises in the case of this parameter is that data collection can take forever, but sometimes you do not have a month or two or three to calculate. Therefore, the most effective indicator remains the collection of reviews and rating building. It can be a count of reviews in which users rate you 5 out of 5 or 4 out of 5 or collect information about the consumer loyalty index (NPS) - for this you just ask users to rate on a scale from 0 to 10 the likelihood that they will recommend you to your friends, and calculate NPS.

Over time, you will see that as the number of new features increases, both the number of reviews and the number of customers increase. This means that you are doing everything right. If this does not happen, then something is going wrong with you. If everything stays the same, maybe that means you need to go and find out what new features you need to implement.



Another thing to consider is the honesty curve. Some people will just lie to you. Look, there is your mother, your friends, friends of friends and completely random people. Your mother will use your product and will be proud of you in any case, so that she will be honest with you. Your friends will be fairly honest with you and will give you recommendations because they care about you - if your product is free. And then you will receive more and more random in nature feedback from those who know nothing about you. These are people who are completely not interested in wasting their time and energy on recall. So take that into account.



Say now that this is a paid product. If we are talking about a paid product, the opinion of mom will be in last place, here, below. She will just lie to you and say that the product is wonderful. But then the graph goes something like this (draws an increasing graph). Your friends will support you and give you useful advice, but in fact the most important thing for you will be the opinion of those strangers, because if they don’t understand what they are paying for, they will tell you about it: their money is at stake .

This is another way to say that you will receive the most effective reviews and comments when working with a paid product. This does not mean that you should make it so from the very beginning. This means that if you are going to create a product that you will ultimately have to pay for, whether it be software, hardware or anything else, you need to come to the moment when you really need to pay. Because from now on, you will be able to create more useful features, which will attract more people willing to pay.




So, you get feedback, and what do you do before the official launch of the project? I always want to do everything quickly, constantly improving the project at the growth stage. You can work with 10 clients at the moment - it makes no sense to try to add features for another 10 million people. You need to optimize the project for the next stage of growth from 10 to 100 clients. Add the features you really need and move on. From my own experience I understood the full significance of this approach to scaling.

Do not try to automate everything and create software for controlling robots that could do everything for you. To understand what you need to do, you first need to start working on these tasks yourself.

for instance
When we started working with professional cleaners, we tried to ask them a lot of questions by phone, and then continued to ask them questions personally. After that they conducted a test cleaning and only then, if the quality of their cleaning satisfied us, they began to cooperate with our platform. According to the results of all these surveys, we hired only 3-5% of candidates.

Over time, we realized that certain questions that we asked turned out to be good indicators of whether a particular person would be able to work through our platform - we could determine this by the results of data collection and viewing the online form that the candidate filled out. Only after that we created an online application with which candidates submitted applications, and then, if necessary, asked them several questions during a personal interview.

If you try to automate some tasks too quickly, you may find yourself in an unpleasant situation where you cannot quickly respond to new data and develop processes iteratively - similar to how we gradually moved from personal questions to creating an application.

The next point is that temporary poor performance is better than permanent paralysis. By this I mean that at this stage it is useless to engage in bringing the product to perfection. When you enter a new stage of growth, what you tried to polish at the previous stage ceases to matter. Therefore, when creating something, do not worry about all extreme cases, work on the most likely scenarios, starting from an understanding of who your main user is. As your company grows, the volume of such extreme cases will only be more and more.

And finally, beware of the “Frankenstein approach”: the fact that you collect user feedback and ideas is certainly good, and your first desire in this regard will be an attempt to satisfy all of them, do everything that they ask, and show them the next day change and make them all happy.

Definitely, you need to listen to user reviews, but if someone advises you to do something, it’s not necessary to run away and fulfill this request. What you really need is to get to the bottom: why in reality the user asks you for these or other changes. As a rule, what your audience offers you is far from the most optimal solutions.

What they really want to convey to you is either that they have problems in the process of using your product, or that the product in principle does not solve any of their problems, because of which they may not be ready to pay for him. Find out before you put together all the offers from your users, for which, in fact, you can’t see the real problem.

So, you have a product that you are ready to supply - many at this moment continue to “finish” it and delay the release. I believe that the very idea of ​​remaining in secrecy and improving the product to infinity is, in fact, an imitation that is “cheaper” than innovation in terms of money and resources.



I believe that everyone should generally understand that if your idea is really good - it doesn’t matter at what stage the release will happen - someone will definitely try to use it and make their product as fast as they can to surpass you. There is no point in rejecting the opportunity to acquire potential users and get a lot of feedback from them simply because you are feeling paranoid and afraid that someone will do it for you.

I hate to repeat about this, but it is precisely these moods that I see among the founders of the companies - I myself went through something similar. I think that unless you create something that requires a multi-million dollar investment at the start, there is absolutely no point in waiting for something and delaying the release.

[ Second part of the translation of Adora’s lecture ]

Also popular now: