Alex Schultz (Part 3): An Introduction to Growth Hacking
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:
Second part of the course
First part of the course
- Sam Altman and Dustin Moskowitz: How and why to create a startup?
- Sam Altman: How to build a startup team and culture?
- Paul Graham: An Illogical Startup ;
- Адора Чьюнг: Продукт и кривая честности;
- Адора Чьюнг: Стремительный рост стартапа;
- Питер Тиль: Конкуренция – удел проигравших;
- Питер Тиль: Как построить монополию?
- Алекс Шульц: Введение в growth hacking [1, 2, 3];
- Кевин Хейл: Тонкости в работе с пользовательским опытом [1, 2];
- Стэнли Тэнг и Уокер Уильямс: Начинайте с малого;
- Джастин Кан: Как работать с профильными СМИ?
- Андрессен, Конуэй и Конрад: Что нужно инвестору;
- Андрессен, Конуэй и Конрад: Посевные инвестиции;
- Андрессен, Конуэй и Конрад: Как работать с инвестором;
- Брайан Чески и Альфред Лин: В чем секрет культуры компании?
- Бен Сильберман и братья Коллисон: Нетривиальные аспекты командной работы [1, 2];
- Аарон Леви: Разработка B2B-продуктов;
- Рид Хоффман: О руководстве и руководителях;
- Рид Хоффман: О лидерах и их качествах;
- Кит Рабуа: Управление проектами;
- Кит Рабуа: Развитие стартапа;
- Бен Хоровитц: Увольнения, повышения и переводы по службе;
- Бен Хоровитц: Карьерные советы, вестинг и опционы;
- Эммет Шир: Как проводить интервью с пользователями;
- Эммет Шир: Как в Twitch разговаривают с пользователями;
- Хосейн Рахман: Как в Jawbone проектируют hardware-продукты;
- Хосейн Рахман: Процесс проектирования в Jawbone.
Alex Schultz ( Alex Schultz ): Other tactics to ensure growth: virality, the use of SEO-Optimization, SEM, referral and affiliate programs. It seems to me that virality can be considered from two points of view. Adam L. Penenberg has a great book called Viral Loop . In this book, he examines the many case studies of companies that have grown through viral marketing. I highly recommend you read it if you are interested in viral marketing and advertising. I think Ogilvy's book on advertising"Is also an excellent reading on this subject: in the 7th chapter of the book there is an example - glue the car to the billboard with super glue, and everyone will buy this super glue from you. There are several really interesting, creative examples in the book.
So, let's talk about virality. Sean Parker created a great model that he told us about when I joined Facebook: it is about talking about virality, about a product in terms of three approaches. First, think about the “efficiency” of your program - how many people will be affected by the viral campaign. Secondly, consider conversion rates, and thirdly, frequency. These approaches form an understanding of how “viral” your product is.
Hotmail is a model example of a great viral marketing campaign.When Hotmail started, there were many email services on the market that threw a ton of money into traditional forms of advertising. At that time, people could not get free access to the mail client - it had to be tied to their Internet service provider. Hotmail and a couple of other companies made access to the mail client possible from anywhere in the world. You could log in by accessing the Internet from a library or school. For those who wanted to access mail, this opportunity was of great value.
Most of the companies that provided such services carried out large-scale advertising campaigns on TV, bought ads on billboards or in newspapers. However, the Hotmail team didn’t have such tools, so they had to brainwash to find a way to make themselves known. As a result, they added to each letter sent via their system a small text: “This letter was sent using Hotmail. You can get a mailbox for free here. ”
Interestingly, the efficiency of their program was quite low - at the selected time you could send a letter to only one person. Of course, you could send spam, but in this case, the respondent would hardly click on the link. The frequency, on the other hand, was high - you write letters to the same people again and again, which means that the recipient sees the link one, two, three times a day, which increases the likelihood of impressing the recipient. The conversion rate was also high, since many did not like that their mail service was tied to an Internet service provider. As a result, Hotmail became an exclusively viral product, because the frequency and conversion rates for the described campaign were high.
Another example is PayPal.PayPal is interesting because in this case you have two types of customers - buyers and sellers. Another interesting point is that eBay provided PayPal with a viral growth. For viral growth, you can use many things that are not necessarily directly related to the term “virality”. If you tell the seller that you are going to send him money, the conversion will be extremely high. Both the frequency and efficiency of the project were low. But PayPal conducted a campaign in which the company paid those who invited friends to become customers of the service - thus the campaign became viral from the point of view of the buyer.
They did not need to do the same for sellers, because through PayPal they already received money from buyers. But the project has become viral even on the consumer side - if someone tells you: “Register, and you will receive ten dollars,” will you refuse? So, they were able to provide viral growth, because the conversion rates for their campaign were high both on the side of the buyer and on the side of the seller, and not due to large values of efficiency or frequency. This is a good way to evaluate a product for virality.
Facebook virality was not provided by mailing list or the like.It was provided by the users themselves, passing on product information by word of mouth. Interestingly, in order to start using PayPal or Hotmail, you should have received a message from someone who has already used the service. Facebook did not have the built-in ability to send a message to those who did not use the product. Everyone thinks Facebook is the product of a successful viral marketing campaign, but that's not why it grew. In our case, there is virality in the “word of mouth” format, because we created an amazing product that users really wanted to tell their friends about.
Question:[On the example of Hotmail] for the first round of the campaign, low efficiency is a relatively normal thing. Will it increase subsequently as the campaign grows and with the number of people sending others letters about your product?
Schultz: First of all, I think that at first they sent letters to a small group of people. Even though you, as a user, repeatedly sent letters with links to all your mail contacts, the campaign efficiency was much lower than what can be achieved now, when you can import someone’s entire contact list and send letters to all these people, or write a message to all the friends of a person on Facebook. But in general, the idea is true: the more you attract users, the more they send letters, and the more they send letters, the more successful the product grows.
Question: Does conversion play an equally important role in this process?
Schultz: Of course. Take the same example: in the case of Hotmail letters, you just had to click on the button, but if it was about advertising on a billboard, you would have to remember the address of the site, go to it, find the registration button and register. Do everything to reduce the number of operations that the user must perform to achieve the goal. The transition from billboards to online advertising significantly reduces the number of such operations.
Q: Are conversion rates and conversion rates related?
Schulz:Of course. If you send a person a letter with the same content several times, or if he stumbles on the same banner ad several times in a row - these rules apply to any channel of user interaction - [so here], the more often you [for example ] showing the user the same advertising message on Facebook, the less likely the user will click on it. Therefore, we rotate content: this is true for banner advertising, and for news in the feed. If you are seeing an article in the news feed for the fiftieth time, you definitely won’t want to read it. The same thing happened with the Hotmail mailing list. If you send out letters inviting you to join the service to the same people several times, your conversion rate will drop. This is a fundamental truth
Another way to look at virality is to follow the example of one guy named Ed. Ed leads the growth team at Uber, and he has worked on a similar team on Facebook. He received an MBA from Stanford and took a course similar to this, in which students discussed virality and tried to create viral products. This is interesting here: if you look at Uber , it is obvious that their main focus is drivers. This is a bidirectional market, so they need drivers. Working with them is extremely important for the team, even though it contains the best viral growth specialist in the world.
So, regarding virality: you need someone to, say, share their contacts with you. Then the question will be: how many people do you need? How many do you ask to do this? How many will respond positively? And how many people will eventually provide access to their contacts? It is very important that first people need to register on your site. You want them to then send out invitations to join your service to their friends, ideally to all their contacts, and not just a few of them. And then those who receive this message should, as planned, click on it and also subscribe to your service.
You multiply all numbers (percentages) at each step - it is important to understand the value of the K-factor. For example, a person using your service sent messages to one hundred of his friends, 10% of them followed the link and 50% registered, and among these remaining only 10-20% provided access to their contacts: in this case, the K-factor will be from 0.5 to 1, and your service will not become viral. Projects like Viddy have excelled in viral marketing. Their K-factor was above 1, which is quite achievable. But if at the same time your product does not have high retention rates, then all this is pointless.
Even if you figured out the K-factor, you need to understand that retention is a much more important thing than virality, so start working on the latter only after most people who subscribe to your service continue to use it.
Now we’ll discuss something else: SEO, mailing lists, SMS mailings and notifications. Regarding SEO, there are three things you need to keep in mind. The first is keyword selection. Many do it not well enough. When I launched my site about cocktails, which I already talked about, I spent a year optimizing it in the search results for “creating a cocktail”, but it turned out that almost no one in England was looking for sites using the phrase “creating a cocktail” - such people approximately 500 per month. But I was the first to issue this phrase - and I felt cool! I had as many as 400 visitors a month! In fact, everyone is looking for similar sites with the words “cocktail recipes,” and with the States, with the words “drink recipes.” I did the optimization based on the wrong premise. So you first need to choose the right keywords, and then do the rest.
The selection consists of answering the question what (from what is related to your site) users are looking for, how many of them, how many companies use the same keywords, and how much all this has value for you. Supply, demand and value. Therefore, study your keywords to understand which ones are worth working on in the future. In my opinion, it’s best to work with the Google AdWord Keyword Planner Tool for that purpose.
Once you do this, you will need to work with the links. A place in search results is a central concept for SEO, and Google in this regard does a lot so that the system cannot be fooled. The information about whether users are specifically looking for your site, what the link to your page looks like, and so on, is embedded in their delivery algorithm, so if you try to “trick the system”, you can be considered a spammer. Repeated white text on a white background no longer works.
But the most important thing in terms of high place in the search results is links to your site from other visited resources and effective internal linking of pages. We launched SEO in September 2007, I joined Facebook in November 2007. However, when we launched SEO, there was no traffic from public pages (we considered them as a starting point). When I tried to go to a similar page myself, I realized that the only way to do this was to click on the About link in the basement, then open the list of articles from the blog, find some author and go through his profile in search of those friends who have there were public pages.
It turned out that according to Google, we tried to hide these pages, so their ranking in the search results was very modest. We changed the approach and added a new directory so that Google could quickly find any of these pages on our site, and our SEO traffic grew 100 times.
In my opinion, mailing list for people under 25 years old is a useless product.These people use WhatsApp, SMS, Snapchat, Facebook, they do not use mail. If you are targeting an older audience, mailing will be effective. It works when it comes to disseminating information, but let's be realistic - this is not a product for teenagers or even students. You yourself know very well how often you use applications to send instant messages, and how rarely you use mail. But you are one of the people who use mail most often for your age, because you are in Silicon Valley.
Given this, you need to understand that both mail, SMS, and notifications work the same way. To all of them there are questions on the possibility of their delivery, so this should be considered first. Your letter should be in the mailbox of the recipient. Therefore, if you send out a lot of spam, use dirty IP addresses or share the server with those who send spam, your letters will end up in the Spam folder and will not reach the recipient. Letters from you in general can be blocked or sent back.
With regard to letters, you need to carefully study the feedback from the servers to which you send letters, pay attention to the errors of the 400 series and the errors of the 500 series, you need to carefully monitor how they are fixed. If letters from someone come back, try to send the message one or two more times, but not more, because if you besiege someone else's mailbox, the mail provider will put you in the Spam folder, which will be very difficult to get out of. If you are listed as spammers at Spamhausor another similar resource, protecting your good name will also be extremely difficult. With regard to mailing lists, it is very important not to try to use the “easy ways” and to treat your work responsibly, because in the long run you need your letters to continue to reach the recipients.
The same is true for push notifications and SMS. You can buy SMS-traffic from "gray" intermediaries who will bombard everyone in a row with messages. This will work for a while, but sooner or later such activity will have to be stopped. I saw a lot of companies making such mistakes, thinking that this would allow them to grow. If you cannot deliver letters, SMS, or push notifications to the recipient, you cannot succeed. As a result, you simply throw spam or notifications to your most experienced users, and make it difficult for them to unsubscribe from all this. As a result, they will begin to block you, and you will not be able to send them anything. And it will be extremely difficult to change the situation in the opposite direction.
Therefore, when considering mailing lists, SMS-messages or notifications, you should first of all think about whether the user will receive them. In addition, the question remains with the values of clickthrough rates or opening (letters). What interesting things can you add to your message so that the user opens your letter or clicks on a link or button?
People everywhere are trying to create marketing campaigns that, in my opinion, look like spam. There is nothing interesting in these letters. Do not create such letters, because in this case, each of your users will receive text with the same content: the one who signed up for you yesterday and the one who uses your service for many years - do they need the same messages? Not.
The most effective mailing option is notifications.What do you send to the user in this case? What should I be notified about? Answering these questions, we very often make erroneous decisions. As a Facebook user, I don’t want to receive emails about every like I receive, because I get a lot of likes from a lot of friends. But if I were a new Facebook user, getting the very first like would be “a moment of magic” for me. When we turned to the notification format, our CTR for mail, SMS and push notifications increased significantly, but we used this format only for those users who rarely visited the site - for them, such messages do not look like spam.
Decision making on this issue has given us invaluable experience. When thinking about a newsletter, you must first understand what notifications to send to users. Next, you need to think about how to create an “event” marketing campaign. One of the best e-mail campaigns in eBay in terms of CTR in my memory was the campaign created after the first transaction between residents of different countries was conducted through our website. This was awesome because the newsletter was extremely timely and informative - exactly what the users needed. Try to get messages to the recipients. Focus on notifications and event marketing campaigns.
I would like to end with my favorite quote from General Patton. Yes, this is an obvious but very true idea:
“A good plan today is better than a perfect plan tomorrow.”
One of the things that Chamat [Palihapitiya] was able to show us, and which Mark [Zuckerberg] continues to show everyone on Facebook, is that it is important to move fast and not be afraid to ruin anything. If you conduct more experiments than others, if you are more greedy for growth, if you are ready to fight for each new user and work late to attract him, conduct an experiment, receive data, if you do it again and again, you will grow faster.
Mark once said that, in his opinion, he won because he wanted it more than others, and I really believe in it. We just worked a lot. It’s not that we are terribly smart, and not that we did something incredible before Facebook. We just worked very, very hard and quickly brought things to the end. I highly recommend you do the same. Growth is a relative concept.
[ Translation of the next lecture by Kevin Hale ]