Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Annual OSD survey to better understand and serve our stakeholders #77

Open
studiospring opened this issue Mar 19, 2017 · 18 comments
Open

Comments

@studiospring
Copy link

As mentioned in issue #73, I would like to start an online survey to learn more about the various stakeholders in OSS so that it easier for OSD to achieve its goal of improving the UX of OSS. Design is not, and has difficulty being, a regular part of OSS development. I approach this problem from a design perspective and the first thing to do as a designer, is understand the problem better. Therefore, with the above goal of OSD in mind, I would like to learn more about the main stakeholders involved. These include:

  • OSD members
  • designers not involved in OSD
  • open source developers
  • users of OSS

I would like to gather enough information to be able to:

  • create personas for each stakeholder
  • understand motivations and pain points in the OSD process
  • understand how to make the inclusion of design in to the development process easy and worthwhile for relevant stakeholders

There are other bits and pieces (nice to haves), but the three points above are probably the meat of the survey. I also mentioned in issue #73 an idea for a usability index. Given the difficulty and the fact that the focus of that research is projects, not people, I do not want to include it as part of this survey (but please think about it!).

Ultimately I want this survey to show us how OSD can serve its stakeholders (not just members!) better.

Some other points:

  • it would be great if this were an annual survey with questions that do not change from year to year, to monitor change over time.
  • we also need ideas on the best ways to gather respondents. We want as many respondents as possible who fit our stakeholder criteria and as few as possible who do not.
  • I find that "preliminary research" in the form of open ended discussion among as many of the different stakeholders as possible makes it easier to ask the right questions. Thoughts welcome. Suggestions for an online venue to have this discussion (IRC, Reddit?, IRL) would also be welcome.

I have looked at technical details and my first thought is that a self hosted LimeSurvey site (GitHub repo here) is the best tool for the job. Other suggestions, experiences, etc are most welcome.

The other thing we will need is (free) hosting. Suggestions welcome.

The next thing to do, after the high level decisions have been made, is to set up a dedicated repo. Given that a new repo will be set up, please keep this discussion focussed on high level issues and keep details (like survey questions) for the new repo.

Finally, I do not want to do this all myself (especially since PHP and MySQL are not my strengths), so please leave a "thumbs up", so I can gauge contributor interest.

@jdittrich
Copy link
Member

jdittrich commented Mar 20, 2017

Technology

I have looked at technical details and my first thought is that a self hosted LimeSurvey site
(GitHub repo here) is the best tool for the job.

I heard that Lime Survey needs relatively attentive administration, since there are relatively frequent security problems (hard to assess for me, but the answer came up already several times)

Libreoffice/@HeikoTietze have a Lime Survey installation, they may know more or be even able to "donate" the infrastructure (?)

There is also a by-us tool by @simonv3 (?) for surveys.

Method

I strongly suggest we split the survey according to the groups we want to ask:

  • OSD members
  • designers not involved in OSD
  • open source developers
  • users of OSS

We also should prioritize these to know which group to tackle first. I personally think that OSD would be the easiest and we should do it first. I hope for the most interesting outcome with asking Devs but that would be 2nd since it is more complicated, method wise.

@HeikoTietze
Copy link
Member

Run my first survey with Lime recently. Great flexibility seems to always come with bad usability, and the setup of a test requires to get used to (mostly html/javascript). I don't know how much effort the administration makes, we have an individual for our infrastructure.
And I'm pretty sure that LibreOffice is happy to support OSD with the platform. Not so sure if the admin is happy to add many users, though. Will ask the team this evening.
Another open source survey tool is User Weave by @bbalazs https://github.com/bbalazs/userweave but not actively developed anymore. We conducted all surveys at https://user-prompt.com/ with this tool.

@studiospring
Copy link
Author

@jdittrich, the problem with splitting the survey into stakeholders is that many people are users, developers and designers at the same time. This is why I prefer a more powerful tool that supports what LimeSurvey calls skip logic or branching.

Some other survey tools are described here.

@jdittrich
Copy link
Member

jdittrich commented Mar 20, 2017

@studiospring We can, at the end, combine these again and branch etc. I suggest planning separately: The questions we want to ask and how we ask them might be very different: For developers, I would ask broader, free form questions, more in a qualitative direction; for OSD members we can assume a bit more existing knowledge, so there, more quantitative questions will be asked etc.

Like in design (well, it is research design) we should collect our questions and assumptions first.

@simonv3
Copy link
Member

simonv3 commented Mar 20, 2017

FWIW - if then statements are not off the books for Quick Survey, I just don't have time to add them. If you know of a Meteor developer who wants to take a stab at it, then they should feel more than welcome.

@studiospring
Copy link
Author

@jdittrich yes, definitely group related questions together, but ask them in the one survey.

@HeikoTietze, when you say LimeSurvey has bad usability, is that for respondents or survey builder? If the survey builder suffers bad usability, do you think the extra features justify the added time and frustration?

@HeikoTietze
Copy link
Member

Definitely for the survey builder. Guess the system is hard to top for respondents when everything is setup nicely. Hard to say if extra costs are justified in the end. Guess you may become familiar with the system, some people like to fiddle with all the options (including me), and when there is no other choice the answer is clear.
However, I don't think we win a prize with a maze-like survey. If you are not interested in inappropriate answers just tell the respondent to step over and don't analyze it. When I read your initial post with
"create personas for each stakeholder", "understand motivations and pain points in the OSD process", "understand how to make the inclusion of design ... easy" it sounds rather like a qualitative test. Let's say you ask for age, gender, tech-savvyness, self assessment of visual design, information architecture and UX. My guess is 35 +/- 7, 95% male, 5.1/0.8, 3.4/0.2, 2.7/0.6, 4.1/0.8. How would you derive a persona?
Qualitative means to ask "What was your initial reason to join?" or "How do you describe your skills?". It's your task then to extract the core concepts from the various replies. And perhaps ask quantitative question based on this knowledge if you don't trust the first survey.
Not sure what question all this solve. Let's assume a persona results from this (why not some rather than one), how do you apply her to what scenario?

@studiospring
Copy link
Author

@HeikoTietze I would prefer to make it easier for the respondent by hiding unnecessary questions rather than asking them to check whether it is relevant for them. Hopefully, there will not be too many branches. This will help in improving completion rates. Further, because (I assume) OSD intends to be the go-to authority in open source design, I think we are obliged to set the standard. It also improves our credibility as an organisation.

You raise some great points! All else being equal, the richer and more qualitative the data the better. However, we do have to work within the constraints of our research tools and resources. I have not given much thought to the number crunching part of this research. Perhaps this is another reason to do a pilot survey.

Not sure what question all this solve.

I guess if I summarised the survey in to (almost) one question for each stakeholder/role, it would be:

  • OSD members: how can OSD help you as designers improve the UX of OSS?
  • Designers not involved in OSS: how can we encourage you to get involved in OS?
  • OS developers: how can we make you value UX more and make it part of your development process?
  • Users: how important is UX to you (and a whole bunch of other stuff!)?

These questions all tie back in to improving the UX of OSS, which is the main goal of OSD. I'm sure all these things will be refined and polished as we go along. This is just a starting point!

Perhaps personas are not the best way, but I do want to get inside our stakeholders' heads. Remember personas are more a psychographic profile, less demographic. Maybe I could just put these questions in the survey (more tactfully!), but I think we can tease out a more accurate picture with more subtle questions. I'm happy to use whatever is the best tool, persona or otherwise, to answer these questions.

@grahamperrin
Copy link

… improving the UX of OSS, which is the main goal of OSD. …

I drafted a new issuemake the primary goal of Open Source Design more immediately visible at the home page – but refrained because the existing primary line is subtly different:

  • Bringing great design to Open Source Software

– with things such as user experience and interface design further down the page.

Re: opensourcedesign/opensourcedesign.github.io#17 (closed) and opensourcedesign/opensourcedesign.github.io#43 (merged) a review of the home page may be timely; should it be made an issue now (in readiness for outcomes of the survey), or later?

@studiospring
Copy link
Author

@grahamperrin, good point. Consistency is a good thing. I think "great design" and "UX" are largely compatible with each other, so it seems to me to be (merely) semantics. I prefer "UX" because it is clearer and more professional. That said, I don't mind using "great design" from now on in the survey. Is it worth creating an issue to change the home page? Thoughts?

@jdittrich
Copy link
Member

jdittrich commented Mar 28, 2017

One suggestion for (a) survey(s). I don't have a section for designers, since I was unsure of what we want to know about them (and what of that could be answered well in a survey instead of e.g. talking/focus group/whateverothermethod)

Survey part for Developers

  • Good experiences with design in the past [Free Text]
  • Bad experience with design in the past [Free Text]
  • In my work I like to [all 5point Likert + don't know]
    • Add features
    • Make code easier to understand
    • improve the Interface
    • Write Documentation
  • UX infrastructure: The project I participate in does… [Problem: People may be in multiple projects]
    • Have UX guidelines [Yes/No/Dont know]
    • Have designers [Yes/No/Dont know]
    • Do usability testing [Yes/No/Dont know]

part for OSD Community members

  • Skillset [all 5point Likert + don't know]
    • Research (User Testing/Need Research…)
    • Coding Front End
    • Coding Back End
    • Interaction Design
    • Graphic Design/Look and feel
    • Other: __________________
  • I like about OSD [Free text]
  • I find difficult about OSD [Free text]
  • I spend [number in h] per week at OSD
  • In any FOSS projects aside of OSD? [Yes Number/No]
  • I am (partly) payed for OSD work
  • My employer supports work like the one I do for OSD [doubles with prev. entry?]

@HeikoTietze
Copy link
Member

Some first thoughts:

Not sure if that's relevant for only me but design is to split into visual design (aethetical presentation mainly) and information architecture (people discuss this term and whether or not it has to get separated further; but in contrast to the visual aspect I see here the general workflow, where functions are placed, how controls are utilized etc.). Asking for design puts me in a difficult situation hence.

"In my work I like to" could be "I use to do the following tasks" (Likert from never to always): Documentation, Help (I would split docu and help as the latter is to me what goes into a program; would be interesting if the results show this), Localization, Visual Design (CSS/Styles/Animation), Usability (HIG, Persona etc.), Information architecture (Mockups/Wireframe), Prototyping, Developing, User only (Level could be interesting), and maybe more.

UX infrastructure: l10n (Transifex, Pootle etc), VoIP (SIP, Jitsi, Mumble), Email, Forums, Code hosting (Github, Sourceforge...) 1st level support, Regular meetings (weekly, monthly etc.), IRC/Telegram (or any other chat), Donation (Gittip, Paypal etc), official/individual blogs (Planet aggregation).

@jdittrich
Copy link
Member

Not sure if that's relevant for only me but design is to split into visual design…

Are you referring to the items in the "skillset" part?

"In my work I like to" could be "I use to do the following tasks" (Likert from never to always):

It depends on what we want to know – reported emotional stance or reported frequency. I went for the emotions, based on what I read about community motivations (e.g. adding functions is more fun than refactoring)

UX infrastructure:

Do you suggest asking for the infrastructure and if yes, both Devs and OSD members, or just one of them? (I like the idea)

@HeikoTietze
Copy link
Member

HeikoTietze commented Mar 30, 2017

Are you referring to the items in the "skillset" part?

It starts with "good/bad experiences with design in the past". And my answer would be: depends.

Do you suggest asking for the infrastructure and if yes, both Devs and OSD members, or just one of them? (I like the idea)

That was your idea :-). I just added some more aspects.

@studiospring
Copy link
Author

Just letting everyone know that I am aiming to create a survey repo containing an initial outline of the project some time next week, if not earlier.

Keeping in mind that the intention is to make this an annual (?) survey, what is an appropriate name for the repo?

  • github.com/opensourcedesign/survey
  • github.com/opensourcedesign/annual-survey
  • github.com/opensourcedesign/stakeholder-survey

Can the name be more specific, in case other surveys are created in future?

@jdittrich
Copy link
Member

I am fine with having a repo (Annual-Stakeholder-Survey?), I assume, though, that the planning of the survey will not happen via pull requests&co; however it might a a good container for the research related discussions (on the other hand, we then have a additional repo which is hard to follow for non-core contribs)

@studiospring
Copy link
Author

@jdittrich I was thinking of using pull requests and Issues, simply because there were no alternatives. To make it easier, I intend to create a "rough draft" of questions, etc in a markdown document, which will later be coded up. However, since Discourse looks like it is coming online very soon #79 that may be a better platform (hmmm, I'm sure I've said that somewhere before!). I'm obviously open to suggestions.

@grahamperrin
Copy link

grahamperrin commented Apr 10, 2017

a community of designers and developers

2017-03-14:

… design is …

2017-03-17:

… we (a community of designers) … amidst … developers- network effects can be useful …

2017-03-21:

… semantics. … clearer and more professional. … Thoughts?

Whilst I love clarity and consistency, I don't require things to be defined.

I value professionalism, in a be excellent to each other way.

I should describe myself as:

  1. neither a designer nor a developer – definitely not a member of the OSD community
  2. part of OSD's plan to include passionate people – a member of the community

… results of a survey should help to untangle contradictions such as those. Towards an understanding of stakeholders; of members of this extraordinarily open GitHub organisation.

Sorry. My intention to provide a meaningful answer to the question was overtaken by a sensation of walking through rooms full of these –

43ea858592ada3510eacf8153591b461 blinds

– except they're knotted, and some of them have torn off but still, the remnants are over my shoulders and in my face and now I need to sit here

p7240290 1970s huon pine armchairs

– and complete a survey 😄

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Development

No branches or pull requests

5 participants