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 Y Combinator experts took place:
Second part of the course
Kevin Hale: So when I talk about creating products that users love, I mean, in particular, “How do we produce products that users will care about, which they sincerely want to see as successful, as in the case of products that we produce, so in the case of the companies behind them? " We will now analyze a lot of information; try not to make too many records - try to listen more.
')
In my account on Twitter, I will post a link to the slides, and, having passed through it, you can leave comments to the slides [
approx. trans. - all this you can find on the link to the source of the translation ]. So ask questions, and if we don’t get to them, I will answer them after the lecture.
Well, friends, over the past few weeks you have heard a lot about growth, but it seems to me that growth is quite simple. This is the interaction of two concepts or variables: conversion values ​​[
eng. conversion is a coefficient that shows the number of site visitors who later became buyers ] and “turnover” [
Eng. churn is a coefficient that indicates the number of customers who have refused service ]. The gap between them largely reflects how fast you are developing.
Most, especially business people, try to view this interaction from a purely mathematical, calculated point of view. Today I want to talk about it from the point of view of relationships with people, because when in a startup you communicate with your users, your relationships are rather close at an early stage; and therefore I believe that this is a new look at how we create our products. We consider many examples on this topic and how to implement them.
My philosophy behind everything that I talk about startups is that the best way to make a billion dollars is to focus on the values ​​that help you get that very first dollar, to attract that very first user. If you learn this, everything else will turn out by itself. This, so to speak, is a matter of faith. I became a YC partner, being a graduate of the program. I took part in it in the winter of 2006 (it was just the second program) and created a product called
Wufoo .
Wufoo is a form builder that helps you create feedback forms, online surveys and the simplest forms of payment. In general, this is a DBMS application that looks as if Fisher-Price was involved in the design. However, it is interesting that since the application was easy to use, we had customers in every market and in every area that we could imagine, including most of the Fortune 500 companies.
I led the business for five years, and later in 2011 we were acquired by Survey Monkey. At that time, we were considered a very interesting acquisition. Then we were only a team of 10 people, and while we received funding here in Silicon Valley, the company was managed from Florida.
We did not have an office, everyone worked remotely, and we were quite different from other companies. Here [
on the slide above ], each point represents a startup that went IPO [
eng. Initial public offering - initial public offering ] or has become someone's acquisition, the point in the distance to the left is us. Below is the amount of funding received, and the vertical axis - the value of the company at the moment. As you can see, the average start-up collects $ 25 million of investment and returns around 676% to investors.
We in Wufoo collected a total of 118,000 dollars, with a return on investment of approximately 29,561%.
In this regard, many ask what makes Wufoo different from other companies, or how our management is different. We paid a lot of attention to the product. We were not interested in developing an application that people would just like to use, which would remind you that you work in an office, since, in fact, this is just working with a database.
We wanted to get a product that people would love, to which they would like to be involved, and we are just fans of our approach to this idea and almost raised it to the degree of a scientific approach. Therefore, we said to ourselves: “In terms of start-ups, from the point of view of creating products that will be loved by users, the following is interesting: love and other sincere feelings are what we find very difficult to achieve in real life. And in a startup we are forced to push them on a larger scale. ” And we decided to start with the question: “How are relationships built in the real world, and how can this knowledge be applied in business and product creation?”
We will consider two such metaphors: attracting new users, as if you were trying to meet with them, and current users, as if you were in a happy marriage with them. When it comes to dating, much of what we have discovered is related to first impressions. Each of you often talks about the beginning of a new stage in your relationship. You tell others about the first kiss, how you met, how you made the offer. This is what we talk about again and again, these are stories about relationships that are passed from mouth to mouth.
Similarly, the situation with companies. People are creatures that constantly build relationships. We cannot not create and not animate what we interact with again and again. Whether they are the cars we drive, the clothes we wear, or the tools and applications we use; ultimately, we attribute our own characteristics to them, note their individuality, and we expect these things to work in a certain way - this is how we interact with them.
First impressions are important at the beginning of a relationship, because we talk about them again and again, aren't we? There is something special about how we treat stories about the start of our relationship. I will give an example. If you are with someone on the first date, and you have a good time at dinner, but have noticed how he / she picks his nose, you will probably not go on the next date. But if you have been married to someone for about 20-30 years and see how he / she is “looking for gold” lounging in his chair, you will not run to call your lawyer (you understand what I mean) and you will not say: “ We have a problem here, you need to prepare papers for divorce. ” You shrug your shoulders and say: "But he has a heart of gold."
Therefore, such an initial interaction is characterized by the fact that the possibility of not passing further is extremely high. When working with applications and most online products, first impressions are usually obvious: you can see what things many companies pay close attention to in terms of what (standard) tasks their marketing specialists work on.
Although experts, who really know the product well, find many other things that can be remembered: the first letter that came to your email; what happens when you register; links; advertising; first interaction with technical support. All this - the ability to hook the client.
So how are we going to make the first impression? In fact, we borrowed the idea from the Japanese. In general, they have two concepts to describe the quality of the finished product. These two notions of quality are atarimae hinshitsu and miryokuteki hinshitsu. The first means already inherent qualities, that is, in essence, functionality. The second means "attractive" quality. Take, for example, a pen. She has miryokuteki, if the weight of the pen, then how ink flows out of it, how people perceive written by this pen, gives pleasure to both the person who uses the pen and everyone who perceives the result, taking the product assessment to a new level. Let's start with a few examples.
This is a link to the registration in Wufoo, and there is a dinosaur on it, and I think it's great! It has an additional function: when you hover on a link, a pop-up hint appears that does not contain data on how to register, or what happens next - it simply says: “RARRR!” In the early phase of usability of the site, we noticed that this effect makes people smile, without any labor, and absolutely everyone.
I think, very often, when we evaluate products, we do not think: “And what emotions appear on the face of those who interact with this?”
The Vimeo start page seemed to me particularly attractive. She makes it clear that when you are going to travel with them, it will be unusual - this is their style. If you start the search for the word “fart”, sounds scroll up and down as if someone was farting. This is something new, as if the site is interacting with you, a little bit of magic, and this is a bit unusual. This wants to share.
It is not necessary to form such interaction through design. This is a registration form for Cork'd, a project conceived as a social network for wine lovers. It says: “The email address is also your name when you register, and it must be real. The name is what your mother calls you. The name is what your colleagues call you. A password is something that you yourself will remember, but others will not be able to pick it up. Confirm the password - enter it again, consider that it is just a test. ”It sounds [
in English ] literally as verses. And filling out the form, you say to yourself: "And what do I like the people who are working on it, I will like it here."
Now take a look at the Yahoo registration form: what does this form say about the service, what will its “face” look like? I’m also upset that any Yahoo product or service works with the same registration form.
I believe that Flickr came up with one of the best calls to action. It sounds like this: "Come to us!".
This is the Heroku registration page. I think this is one of the latest versions. The following is amazing (and you get used to it): scaling backend services is no more difficult than dragging sliders to and fro. Adjusting Heroku's scaling is really very simple.
Here is a site for those who study computer science; I think you will like it. This is Chocolat, a code editor. They have only one call to action: when time runs out, everything remains in place, except that the font changes to Comic Sans: they seem to say: “Hey, we know who our users are and who our customers are. These are people to whom this makes a difference. ”
This is Hurl, a website for testing HTTP requests: sometimes the places where errors occur are opportunities to create a first impression. If you stumble upon a 404 error, you will see a unicorn that is sick of a rainbow.
Often we create pretty colorful marketing materials, but when documentation is required, we save on design solutions. This happens very, very often. MailChimp is a company that understands this. They redid the manuals in such a way that they looked like magazine covers, and suddenly — actually after this innovation — the readership increased, and the number of requests to the mailing list support service dropped.
Speaking of documentation, Stripe is an API-based business that does not have a UX. Generally speaking, the UX is just documentation, but even in the documentation there are opportunities to attract and surprise users. What I like about Stripe are their excellent examples, however, when you enter the application, the real headache for most people is to retrieve authorization data and keys for your API. We were shocked by what they said about this: “When you enter the application, we automatically put your authorization data into examples, so you need to copy and paste them into the form only once, instead of trying to remember them” .
When Wufoo got ready to release the third version of the API, we decided: “Well, in the end, it's good enough, and we would like people to create projects based on it”. We tried to figure out how to design its launch in such a way that it would express our individuality, because many usually hold something like a contest for API-based application developers and distribute iPads and iPhones to winners; it makes you look like others.
Here is one strange thing about our company: our founders are simply obsessed with the Middle Ages, and we all go to the Middle Ages every year to celebrate the anniversary of the company's foundation. Therefore, we said to ourselves that we need to do something in this style. We contacted the guys from armor.com and asked: "Can you make a battle ax to order?". We said: if you win a programming contest, then get it.
In the end, people wanted to tell others about it. They wanted to tell that they are participating in our competition in order to inform others: “I am programming for weapons”. It's great that we ended up with more than 25 applications, the quality and quantity of which we could not afford with our budget and lack of time. We have an iPhone app, an Android app, a Wordpress plugin, and in fact all we did was change how people perceived the story of familiarity with one of our services.
In conclusion, I’ll say that you should just subscribe to
Little Big Details . Generally speaking, this resource contains a huge number of screenshots of software, which show that everything in these programs is arranged as it should, and their creators are honest with users and customers.
As for long-term relationships or marriage, the only research that we eventually managed to read was John Gottman's materials. He took part in the program “This American Life,” mentioned in the books of Malcolm Gladwell, etc. He studies marriages while working in Seattle, and has a special talent — he can view a 15-minute video recording of a couple who quarrel for any reason and predict with 85% accuracy whether the spouses will stay together or divorce for four years.
If the duration of the video increases to an hour, and he asks them to discuss their hopes and dreams, the accuracy of the forecast increases to 94%. The same video was shown to experts in the field of marriage, couples in a successful marriage, sociologists, psychiatrists, priests, etc. They cannot even predict by chance whether the spouses will stay together or not.
John Gottman understands something very important about how relationships develop in the long run - the way we quarrel in the short run can determine the appearance of the entire system. One striking fact that he discovered was not that the successful couple didn’t quarrel at all; it turns out, everyone is quarreling, and we all argue about the same things: money, children, sex, time, and so on (for example, jealousy and relatives from the husband / wife can be attributed to “other things”).
To make it clearer, you can relate each of these points to the problems that we encounter in the technical support service when creating a new product: Money is a problem like “too expensive” or “I have problems with credit cards”. Children are customers of our users. Sex - performance, how long you work, and how fast everything happens. Other - I talked about jealousy and relatives, so in this case it is competition and partnership programs, all that is strange, what is happening in the company, and what you will write about.
And the reason why I like to talk about this from the point of view of technical support is that when studying the conversion funnel, technical support is what happens between each of the steps; this is the reason why people do not move to a new level of the funnel; what prevents conversion.
And while we reflected on all these ideas and developed the company, we realized that, as a rule, creating a new business or a development team is fraught with big problems. They are associated with the absence of cyclic feedback. In ordinary life, people are divorced because of the consequences of their actions. This is the result of the natural process of creating companies, in particular, technical co-founders.
Before the launch of the product there is time for bliss, Nirvana and opportunities. Everything you do - you do the right thing. Each line written by your hand (which seems to be “God's hand” to you) and any code written by you seems perfect; you think it's brilliant.
After launching the product, you enter the real world, and all these new tasks appear that need to be solved; everything you need to deal with. Technical co-founders usually try to return to the initial stage, and we very often see how the company shelves everything that makes a startup a real business, and shifts this work to others. It seems to us that these tasks are secondary, that we have other people in the company who can deal with them.
Personally, we wanted to understand how to change software development in such a way as to instill in the end result some values ​​that we speak about quite seldom, such as responsibility, discipline, moderation and modesty. We call it Development through support [
eng. Support Driven Development, SDD ]. This is a method of developing high-quality software, and it is extremely simple: there is no need to soil a pile of notes for notes. All that is required is for everyone to provide technical support. At the exit, you get a streamlined feedback. Software developers are engaged in support, and as a result you get all the relevant benefits.
One of them is the formation of responsibility for support among developers and designers. When people develop something, they can provide better support. And we are not the first to think about it. Paul English did a lot to support this idea at Kayak. He set up a hotline for support calls in the middle of the floor where the engineers worked, and they just took those calls.
People often asked him: "Why do you need to pay engineers $ 120,000 or more to do work for which you can pay only a part to other employees." He answered: “You know, after the second or third call, when the engineer faces the same problem, he stops doing what he did, fixes the bug, and calls with such a problem cease”. This is a kind of elegant solution that provides quality control.
In addition, John Gottman talks about the reason why we often part with each other, and which arises from four main sources. These are warning signs. He calls them the Four Horsemen [
cf. Four Horsemen of the Apocalypse ]: criticism, contempt, defensive behavior and blocking.
Criticism, generally speaking, arises when people try to focus not on this problem, but on more general ones, constantly repeating "You never listen to users" or "You never think about us." Contempt arises when someone intentionally tries to insult another. Protective behavior is a disclaimer, or a situation where a person tries to come up with an excuse for his actions. Blocking is simply detachment. Blocking, according to John Gottman, is one of the worst things that can happen in a relationship.
Often we don’t care about things like criticism and contempt when providing technical support. Protective behavior: we constantly observe it, especially as the company grows. But blocking: I constantly see how it happens in startups. You get a lot of incoming support calls and you just think: “I don’t need to react at all, I don’t need to answer”. Not going back to these questions is the worst thing you can do, and for sure this is one of the main reasons for “routine” in the early stages of a startup development.
Here's how the support worked in Wufoo. At the time of the takeover, we had 500,000 users in the system, 5 million people used Wufoo forms and reports, and they knew it or not, but all these people received support from those same 10 people: usually they were assigned every day for support on any issue one man.
As a result, there were about 400 calls per week, which is about 800 emails. From 9 am to 9 pm the user waited for a response from 7 to 12 minutes, from 9 pm to midnight it took an hour, and on weekends the answer took no longer than 24 hours. We worked in this mode and as far as scaling.
What many people say and often forget to say about Airbnb is their interesting move when they came to New York and offered to take professional photos of the interiors. The founders of the company went and actually took pictures of the rooms, helped the owners to hand them over better, taking into account the conversion. Most people are not aware of this: at the beginning of the formation of Airbnb, I very often saw Joe in headphones with a microphone, as he provided customers with telephone support non-stop.
[
The second part of the translation of Kevin's lecture ]