# The Art of Survey(s):Lessons and Results
Author: James He (@hesatprylabs)

The following documentation will be used to track lessons learned and techniques to better survey development in the future.

**note - I am not a data scientist, I am not a statistician, and I did not get formal training on creating surveys; feel free to contact me for feedback and suggestions**

[TOC]

## Survey(s)
Ethereum Staker Onboarding Experience Survey

## Survey Creation
- use Google forms or survey monkey
- follow some guidelines https://www.surveymonkey.com/mp/survey-guidelines/
- Understand the type of form controls, when to use radio buttons vs check boxes vs drop downs etc...
- Have a consistent number range for all your items ( i.e. 1-10 or 1-5)
- It's important to add a disclaimer
- disclaimers can provide background and help frame the survey. Describing things like what is being surveyed, who are the people being surveyed, what will happen with the results, will the results be published, who will the results be shared with, and what it would mean to participate in the survey.
- preface with any bias in the interpretation of the results
- the survey is anonymous and there fore can be subject to misinformation
- Avoid answers that overlap.
- avoid questions like this
- 
- month 3 could represent in 2 answers
- make sure you know your measurements! research the notations for clearer questions
- i.e. gb is ambiguous between "GB" and "GiB" (hard drive storage question)

- Keep the survey no longer than 10 minutes, mostly multiple choice, less than 20 questions ideally; short and to the point.
- make sure you have the right multiple choice answers when you give the "other option"!
- otherwise you'll end up with something like this 
- NUC was by far the most popular choice but i didn't include it, the result was different spellings of the same thing.

## Survey Distribution
- COORDINATION! my biggest lesson learned from this experience is that one of our most valuable resources is community participation, and if everyone starts creating surveys there might be less participation when it's needed. Butta and I rushed the release of the survey without realizing that there was going to be a very similar survey published by the Ethereum foundation.
- coordinate with appropriate channels to distribute the survey to the target audience without too much duplication
- have some vagueness to give yourself a buffer room releasing data, there might be more things to consider before publishing.
- please look forward to another survey coming from the Ethereum Foundation Research group soon.
- short answer questions are hard to summarize and publish

## Survey Results

### Ethereum Staker Onboarding Experience Survey

**Please take the following into consideration when interpreting results!**
- There is inherent selection bias as it was only distributed across the 4 client teams and the eth staker discord.
- Surveys may include bad actors or dishonest data.
- Due to the anonymous nature of the survey users could have done it multiple times or non target audience could have completed it.
- The following results have anecdotes from me (an observer without training on data interpretations/ fairly new to the ethereum foundation)
- If you have any suggestions please don't hesitate to reach out on one of the channels discord (JamesHe#5018 or Butta#8883)
- There was a lot of support to make this survey possible thank you to those that proof read, participated, and contributed.

Responses: over 200!

- 
The majority of surveyors have been in the community for a long time !

- 
Most rate themselves as fairly technical, is the technical bar to entry too high?

- 
A lot at stake, 208 mainnet stakers with a minimum of 1 key each that's a minimum of 6656 eth at stake or a minimum of around 20 million USD at today's market value. Some stakers own dozens of keys so that number is exponentially bigger. The current Ethereum Deposit Contract holds over 23 billion usd worth of Eth at time of writing.

- 
Many stakers have started since the beginning of release.

- 
Most were somewhat anxious staking for the first time. The price to stake at time of writing is around 96k USD, not a small amount...

- 
Most had an ok experience with onboarding but we can see there is a lot of room for improvement. Given the technical proficiency and seniority ( Eth community perspective ) of the surveyors there should have been less poor experiences.

- 
Launch Pad experience was pretty good! A few improvements was suggested here https://github.com/ethereum/staking-launchpad/issues/426

- 
This graph shows that most stakers use Linux, as that is the most supported across the major clients, however please keep in mind this graph does NOT represent the number of keys staked per operating system which may skew the results. You may also notice that almost 9% of users are on Windows which is only supported natively for the beacon node in production by the Prysm client. More support on this front will also likely increase client diversity (let me know if a correction should be made on this statement)

- 
There were many lessons learned around these questions. The options were poor and ambiguous, the measurements for size overlapped and were improper. What I learned as someone newer to the community was that NUCs are by far the most popular way of running nodes with home servers and desktop overlap..

- 
Client guides are still the most used way of onboarding however there is a lot of that the client teams can improve on knowing the popularity of the community guides.

- 

- 
points about client diversity have been discussed outside the survey

- 
The Ethereum foundation's website plays an important role in terms of client selection and according to some developers on the team are actively working to randomize the order in which clients are displayed and other ideas to not promote one client over another directly or indirectly.

- 
currently discord is the number 1 place for support. I assume not many members from the survey group acts as an institution but the few support areas are all social media platforms that are typically banned at institutions. This would be an area worth visiting for better accessibility.

The results of the survey help as supporting guidance on where we can better focus our priorities to make the biggest improvements and help us see what areas that might not have enough support. This data can also help us better define our Personas when mapping out the user experience, another way of identifying areas that need improvement.

Thank you again for your participation.