MindMap Gallery User growth experiment series
User growth cannot just stop at making plans and setting goals. Only through practical experiments can we truly understand the pain points and shortcomings, so that we can promptly remedy or continuously improve, and get better results in the next experiment. The map lists the entire process from the beginning of the experimental idea to the analysis of the experimental results. I hope it will be helpful to you.
Edited at 2022-06-16 14:05:08Find a streamlined guide created using EdrawMind, showcasing the Lemon 8 registration and login flow chart. This visual tool facilitates an effortless journey for American users to switch from TikTok to Lemon 8, making the transition both intuitive and rapid. Ideal for those looking for a user-centric route to Lemon 8's offerings, our flow chart demystifies the registration procedure and emphasizes crucial steps for a hassle-free login.
これは稲盛和夫に関するマインドマップです。私のこれまでの人生のすべての経験は、ビジネスの明確な目的と意味、強い意志、売上の最大化、業務の最小化、そして運営は強い意志に依存することを主な内容としています。
かんばんボードのデザインはシンプルかつ明確で、計画が一目で明確になります。毎日の進捗状況を簡単に記録し、月末に要約を作成して成長と成果を確認することができます。 実用性が高い:読書、早起き、運動など、さまざまなプランをカバーします。 操作簡単:シンプルなデザイン、便利な記録、いつでも進捗状況を確認できます。 明確な概要: 毎月の概要により、成長を明確に確認できます。 小さい まとめ、今月の振り返り掲示板、今月の習慣掲示板、今月のまとめ掲示板。
Find a streamlined guide created using EdrawMind, showcasing the Lemon 8 registration and login flow chart. This visual tool facilitates an effortless journey for American users to switch from TikTok to Lemon 8, making the transition both intuitive and rapid. Ideal for those looking for a user-centric route to Lemon 8's offerings, our flow chart demystifies the registration procedure and emphasizes crucial steps for a hassle-free login.
これは稲盛和夫に関するマインドマップです。私のこれまでの人生のすべての経験は、ビジネスの明確な目的と意味、強い意志、売上の最大化、業務の最小化、そして運営は強い意志に依存することを主な内容としています。
かんばんボードのデザインはシンプルかつ明確で、計画が一目で明確になります。毎日の進捗状況を簡単に記録し、月末に要約を作成して成長と成果を確認することができます。 実用性が高い:読書、早起き、運動など、さまざまなプランをカバーします。 操作簡単:シンプルなデザイン、便利な記録、いつでも進捗状況を確認できます。 明確な概要: 毎月の概要により、成長を明確に確認できます。 小さい まとめ、今月の振り返り掲示板、今月の習慣掲示板、今月のまとめ掲示板。
User growth experiments
generate experimental ideas
1. Clarify the test objectives
Correct operation: start from user and business problems
Improve single store conversion rate
Copywriting test
design test
Single page test
Improve full-funnel conversion rates
path test
Compare indicators of new and old versions
New features or versions are online
Explore new features
Complex experiments, MVP, functions, algorithms
2. Find insights from data
High quality experimental ideas
show
Experimental hypotheses have a high success rate
Experimental indicators have greatly improved
reason
Experimental hypothesis is not supported by data
Develop high-quality hypotheses
3 types of data support
Quantitative data
Qualitative data
Best Practices
Attract attention
boost motivation
reduce resistance
Consider the scenario
N rounds of data analysis
Analyze data and find problems
form preliminary hypotheses
Further analyze data and improve hypothesis quality
Best Experiment: Lift Model
value proposition
A clear and powerful marketing slogan allows users to accurately perceive the benefits they can gain.
Correlation
Landing pages and conversion pages meet user expectations and are closely related to the value proposition.
clarity
The experience process is clear and smooth, and users clearly know what to do next.
feeling of anxiety
Make subtractions and don’t give users too many choices
distraction
Reduce visual interference and information noise, and only serve one core purpose
sense of urgency
Prompt users to make decisions and lose aversion
3. Form experimental hypotheses
Output template
If [specific changes]
It is expected that [a certain indicator can be improved by X%]
Because [deep reasons-hypotheses supported by data]
Use templates to output clear experimental hypotheses
Prioritization
ICE model
Expected impact (impact) What is the impact of a successful trial?
Probability of success (confidence) What is the probability that the test will be successful?
Ease (ease) How much resources or cost does the experiment require to go online?
Increase influence
Most core product teams only care about core users
Reach more users
Actively expand coverage and focus on non-core users
Start with high-traffic pages and paths and experiment multiple times
Increase ease
Validate experimental hypotheses at nearest cost via MVP
How to invest the minimum resources and prove the experimental hypothesis as quickly as possible
Whether the experiment can provide valid results and insights
Don’t pursue absolutely precise sorting, but increase the frequency and quantity of experiments
experimental design
1. Select experimental indicators
Correct experimental indicators can comprehensively and accurately test the authenticity of the experimental hypothesis, thereby measuring the success or failure of the experiment.
Three types of indicators to measure the success or failure of experiments
Core indicators (1)
Key indicators that determine the success or failure of an experiment
Auxiliary indicators (<10)
Funnel Segmentation Step Conversion Rate
Important downstream indicators
Other key user metrics
Reverse indicators (1-2)
Possible negative effects of the experiment
2. Determine the experimental audience
Who will be included in the experiment?
Run experiments on specific groups of users
Approximately how many samples are needed for the experiment?
statistical significance
The difference in conversion rates between the control group and the experimental group is real and not caused by random error
Factors affecting the number of samples required for an experiment
Original version conversion rate
New version conversion rate
Statistical significance requirements
The lower the traffic or users, the greater the changes in the experiment.
3. Design experimental version
How many versions are designed? What's the difference?
Point 1: The version depends on the number of experimental hypotheses
Point 2: Clarify whether it is an optimization experiment or an exploration experiment
Point 3: The more versions, the greater the total number of samples required
How is traffic distributed among versions?
Uniform flow distribution
Eliminate the influence of all external factors
Development and launch
Specific steps
Development experiments
Develop experimental version
Buried data
Experimental QA and UAT
Check the experimental version and data burial points
Online experiment
Online code
Start experimenting
Experimental indicator burying method 1 Third-party A/B testing tools
Determine experimental indicators
Find the corresponding user behavior
Define requirements: buried points of behavioral events
After the development is completed, the data is returned to the A/B testing software
A/B testing software automatically calculates experimental indicators
Experimental indicator burying method 2 Manually analyze experimental results
Determine experimental indicators
Define the requirements: Which users are included in the recording experiment? What version of the experiment were they seeing?
Development is completed and the data is returned to the database
Manual analysis, experimental indicators
Analyze and apply results
1. Credibility of assessment results
Determine whether the results are statistically significant
2. Analyze experimental results
Did the experiment succeed or fail?
Observation period
short term observation
long term observation
What's the reason behind this?
Segmentation funnel
User research
Result grouping
Follow-up experiments
3. Decide on the next step in the experiment
Experiment completed → analyze results
Commercialized
give up
Keep iterating
Amplify experimental impact
Take advantage of victory
learn by analogy
Adjustment plan