[PYTHON] Summarize the main points of growth hacks for web services and the points of analysis

** Contents covered in this article **

The word "growth hack" has been popular for the past few years I haven't received a business card printed as "Growth Hacker" so far ...

Anyway, I read this book the other day

The easiest growth hack textbook http://www.amazon.co.jp/dp/B01BY7HMYO/

This is a book that summarizes the know-how of growth hacks by Mr. Vasily, who runs "iQON". In a nutshell, it was a very nice book with simple and important factors tied together. Here's a summary of what I thought was particularly important in this book.

If you are interested, I recommend you to purchase the original text and read it. The Kindle version costs 1680 yen, so considering the contents, I think it's quite cheap.

• By the way, I am an individual who has nothing to do with Vasily.

Also, this is not a formal commentary on the book. There are some parts that do not completely match the contents of the book because my own translation is included in the wording etc., but please consider it as a memorandum and impression after reading personally.

Constitution

The contents of this book will proceed in the following structure. (Some books are listed in a different order)

--Introduction to Growth Hack --PDCA cycle of KGI and KPI, and growth hack --Definition of framework for growth --Details in line with the above framework ―― 1.Activation = Get used first ―― 2.Retention = Have them continue to use ―― 3.Refferal = Have people use it enough to introduce it to others ―― 4.Revenue = Have them use as much as they charge ―― 5.Acquisition = Have more people use it

It was very easy to understand that actual case studies, usage examples in Vasily, concrete methods, etc. were scattered all over the place, and I felt that it was becoming more usable in practice.

** Introduction to Growth Hack **

What is a growth hack? Those who know it are what they know.

What I thought was important

Incorporate a growth mechanism into the product itself

I felt that was one important keyword. Case studies were also mainly introduced that could understand the above words. The example that came out is

• Hotmail
• AirBnB
• Facebook --Dropbox etc.

** KGI and KPI, and PDCA cycle of growth hack **

Growth hacks cannot be done without clear goals and indicators. Therefore, for Mr. Growth Hack, the design of KGI and KPI becomes a vital issue.

What I thought was important

KGI is the goal, and KPI exists as its decomposition factor KPI must meet the following three conditions

1.Actionable：
It is meaningless unless it is an index that leads to action.
☓ Total service PV → It is difficult to raise and the range of things to do is too wide
◯ Posting rate of users on the first day of registration → Specific measures to increase it are likely to be considered

2.Understandable
Of course, you have to be easy to understand

3.Comparable
The comparison must be meaningful
☓ Number of posts by users on the first day of registration → As the number of users increases, the comparison between last month and this month is not meaningful.
◯ Posting rate of users on the first day of registration → If you make a ratio, comparison with the past will be meaningful

Also, as a point to be careful, there are times when KPI has increased but KGI has not increased.

Example: Increased completion rate of tutorials for KPI users → KGI=Repeat rate does not change

Such. This is a bad effect because the KPI is very broken down. Make sure both are up properly.

Analytical perspective

** Keywords **: Funnel analysis, pattern analysis, user interviews

Funnel analysis:
Determine at what action step the user is leaving

Pattern analysis:
Find differences in KPIs for each group when divided into groups
Alternatively, analyze the characteristic differences between the groups when the KPIs are divided into different groups.

User interview:
Whereas the above two are quantitative analyzes, this is a qualitative analysis.
Words that Crowdworks and Mechanical Turk should be used for conducting questionnaires etc.

** Method-Sharing KPI / KGI ** In Vasily

-Analyze with Google BigQuery
-Visualize with Tableau
-Post to Slack

It seems that they are sharing in the form of. This is good. I want to imitate.

** Framework for Growth **

Click here for commonly used frameworks in the growth hack world

AARRR(It seems to be called Ah)
- Acquitision :Acquire a lot of people
- Activation  :Have them use
- Retention   :Have them continue to use
- Refferal    :Have them use it enough to introduce it to others
- Revenue     :Have them use as much as they charge

I think it's true, but Vasily says that this is a little cool. The reason is (I have a little personal translation)

Not in line with the actual flow of growth measures Also, the perspective on the user experience is weak

What does this mean?

What is important for the product is `" user problem "and" solution " `` You should think about everything from there

It is thought that this is the idea of Mr. Kanayama, Vasily CEO. In AARRR, `" user's problem "and" solution "` are most strongly related to ```Activation'', so that is the process of growth hack. It is correct to bring it first.

In other words, not AARRR

ARRRA(It seems to be Ala)
- Activation  :Get used first
- Retention   :Have them continue to use
- Refferal    :Have them use it enough to introduce it to others
- Revenue     :Have them use as much as they charge
- Acquitision :Have more people use it

It seems that one of the important proposals of this book is to catch the growth hack in the flow.

Activation This is a very deep place, so be sure to read the books listed above.

Retention

What I thought was important

Super important to increase retention (revisit). If the retention is low, no matter how hard you try, ** a bucket with holes. ** **

KPIs such as ** "◯ day revisit rate" ** that determine whether or not to revisit differ depending on the nature of the service. Caution.

Analytical perspective

** Keywords **: Cohort analysis, retention curve

Cohort analysis:
Look at retention (revisit rate, etc.) after ◯ days by county in chronological order

Retention curve:
Analyze where there is a particular decline with a curve that visualizes the results of cohort analysis in chronological order

Refferal

What I thought was important

Referrals are one of the most difficult growth hacks Also not required for some services

There are also ** several types ** of referrals

1.Natural diffusion:A mechanism that spreads naturally when using a service (such as spotify)
2.Artificial diffusion:Incentives such as points
3.Parasite:Aim to spread from similar services(Youtube embedding)
4.Word of mouth:Mechanism to post on SNS etc. with comments

Analytical perspective

** Measure NPS ** For NPS (Net Promoter Score), click here (http://marketingis.jp/wiki/NPS) NPS is a measurement method used all over the world and has abundant case studies. You can also use SurveyMonkey or embed it in your advertising banner, We recommend Wootric, a tool that provides an SDK that allows you to embed surveys in your company's service.

** Analyze viral coefficient ** The viral coefficient is defined as the number of users called in per existing user. If this number exceeds 1, the service will grow exponentially.

Viral coefficient=Number of referral actions per person x Registration rate from referral links

However, in reality, the viral coefficient rarely exceeds 1. In Vasily's sense, 0.5 is enough

In addition, the user actually takes the referral action, and the invited user registers. It is necessary to consider the lead time.

Revenue

What I thought was important

I'll do my best for ARPU. Aim for LTV> CPA.

Both are quite natural.

After that, hack measures are different for each business model, so identify your own model. As a classification,

1.Billing model (game, etc.)
2.Advertising model (media, etc.)
3.Commerce model (such as EC)
4.Fee model (sharing economy, etc.)

In particular, 3 and 4, which are of personal interest, introduced the following hack directions.

3.Commerce model
>Increase product browsing PV → Display recommended products
>Increase the cart input rate → Post information such as free shipping and point campaigns on the cart screen
>Increase the purchase completion rate → Diversification of payment methods, etc.

4.Fee model
>Make a king user
>Bring in anchor users (influential users)

Acquisition

What I thought was important

Be aware of ** creative freshness **. If you keep hitting the same creative ad for a certain period of time, the effect will gradually decrease. In this case, replacing the creative often returns the effect, so have a guideline for the period to switch in advance. (Past achievements, etc.)

Have ** efficient / sustainable improvement measures **. For example, iQON's TV CM was divided into two periods, but it was to decide the creative of the second period based on the effect of the first period.

In addition, in online advertising etc., we improved by the following measures

1. Advertise with creatives on four different themes (appeal)
2. Measure the effect (example: appealing the scene to be used> appealing the voice of the user> appealing the number of users ... etc.)
3. Prepare multiple creatives in "Scene appeal" and set up a campaign for the time

** Understand the potential of TVCM and online collaboration **. During the CM period, it is effective to make the contents of the application linked with TVCM. TVCM can also boost the effectiveness of online advertising.

Analytical perspective

Be careful of ** index design **. For example, the cost many-body effect of advertising is measured not by the number of installations but by the number of continuous users after 7 days.

** It is important to have a mechanism to measure advertising effectiveness **. Even if the effect cannot be measured 100% accurately due to offline advertising, etc., a certain amount of guesswork is sufficient, so create a measurement index. → This makes it possible to carry out A / B tests.

There is no point in sticking to perfect measurements. It's better than not knowing anything.

reference

The easiest growth hack textbook http://www.amazon.co.jp/dp/B01BY7HMYO/

Vasily's Growth Hack Blog http://growthhack.vasily.jp/