Outputs from Hunchworks Challenge 1: Connecting People

[Chris van der Walt]

Hi Everyone,

Attached are the interaction flows and sketches from the team at Adaptive Path focusing on how users join Hunchworks and quickly build trusted networks. We are really excited by the depth of thought and great design that the deck reflects. Particularly interesting to me is the focus on identifying other users with complementary skills and the ability to recommend experts who may be of value in verifying a hunch to other users. We are starting work on a prototype this week.

Challenge 2 will focus on how trust is built in the system both between users (reputation) and with the system itself.

Please share your thoughts and stay posted....
Cheers
Chris + Sara

[Chris Blow]

This is really interesting! The presentation is very clear. I have been merely lurking on the list so I am a little out of the conversation. I have a quick question: where can I find some examples of existing "hunches"? Have there been any "practice sessions" for this interaction model, using a blog or Google Spreadsheets software? Could we get some transcripts to use for the next level of design fidelity? I would be really keen on seeing the documentation from an "ideal hunch discussion."  What does a successful Hunch look like over a year or two? I'm genuinely asking — I'm not suggesting that this can't be done!

However, I do sense there is still some hesitation in this model, perhaps some mismatch with real world processes. The wireframes in this document seem very optimistic that this is a viable interaction — but I don't think I'm quite convinced that this can overcome the cold start problem. I worry that the Adaptive Path crew (whom I greatly respect) may be delivering a design product which is intended to lend confidence to the Hunch model, not necessarily to expose how it could be problematic. So I would like to raise the question: how could the Hunch model be problematic? How does it constrain the conversation? And as a newcomer on this project I would like to take a close look at what already works in the wild, so we can just try to reproduce that with the least harm possible. So I would appreciate any help in seeing real-world examples of how this can achieve useful outputs for the stakeholders.

One more short question: how will these designs be initially launched and piloted? How many people will be using the system in its first few weeks of life?

Thanks,
Chris Blow

[Chris van der Walt]

Hi Chris,

Thanks for your inputs and questions.

Regarding the hunch mechanism itself, it is still in development. We will be focusing on designing the model over 3 challenges starting next week:

Verifying Hunches
How is a hunch verified? What kinds of evidence are attached? How do people vote on hunches and evidence? Should there be different models of verification?

Hunch Merging
How do hunches interact with each other? Can they fork? Can they be aggregated? Who controls these mechanisms? What about contravening hunches?

Hunch Evolution
What is the life cycle of a hunch?

I think your suggestion to create an ideal hunch discussion is a great one. Perhaps we could work on mocking one up here in the working group.

I think it is important to remember that Hunchworks is not intended to be a standalone application. It is intended to be part of the Global Pulse toolkit which will include analytical, visualization and reporting tools that will be used by UN agencies, lab staff and local government staff in partnership with Global Pulse. This means that it will be likely that when users join, there will already be networks established that they can join based on their communities of practice or teams. The network building features will add value by expanding those networks and I guess we need to spend some time thinking about that expansion model. It seems to me that the real test for adoption will be based on how effective the toolkit is as a whole and Hunchworks will hopefully add value by enhancing research through the leveraging the networks effects of the social graph.

We are still working out the right beta testing and deployment model. I think it would be useful to work initially with some motivated communitities (e.g. CrisisMappers) on some very specific projects to get some user testing and feedback. If anyone has suggestions for some interesting pilot projects, please do let us know.

Cheers
Chris

--
Chris van der Walt
Strategic Communications Advisor
office phone: +1 646 723 4806
mobile: +1 646 283 0177
email: vande...@un.org
www.unglobalpulse.org

Global Pulse
Executive Office of the Secretary-General, United Nations
370 Lexington Avenue, Suite 1707, New York, NY 10017

[Chris Blow]

Thanks, very interesting. I look forward to learning more.

A few other questions: 

- Do the existing teams (the target Hunch users) have any standardized outputs already, that we can embed as objects of discussion? I assume they all do paperwork ... could these documents be embedded? Presumably these conversations all happen in the context of other documents or data. So I'm curious if there are any that can be used at this stage, as part of prototyping.

- In general, how can we better support existing team infrastructure, make it more relevant to work that is already being done? 

- What are the intended limits of the design use case: should I expect to find Hunches across all disciplines of the UN Agencies/staff/partners? Which hunch topics lend themselves to a pilot project for the immediate term? (Probably safe, pro-active, non-catastrophic context, but what topic, Epidemiology or Logistics or Design etc?)

I can see how the breadth of mental models involved makes this more difficult. It is difficult not to think in terms of "designing for everyone." So in this sense, I agree that a pilot case, and a plan for building user community in the UN, is going to be the most important design work, the community design. I think a prototype experience should be doable with some lightly customized blogging tools, or even just a spreadsheet.

It sounds very ambitious, and hopefully very useful, I look forward to keeping up with the project as the work continues. I will just keep chattering on the list, if it is useful. Sorry if my questions are redundant.

Cheers,

Chris Blow

PS I'm thrilled to see the github page: https://github.com/global-pulse

[PJ Onori]

Hi Chris, thanks so much for all the great thoughts. I am a part of the Adaptive Path team working on the Global Pulse project. This feedback is truly helpful.

Here are my responses to your questions:

-From talking to the Global Pulse team, there is a deep desire to allow a wide range of media, documents or data as the evidence attached to a hunch. It may be difficult to predict the types of evidence attached or the ways they are used to help a hunch along. If the system is open and flexible, a dedicated community will likely begin to form its own set of parameters and standards. Once those structures begin to take form, it will be easier to design and build functionality to support them.

-That's a very good question. The Adaptive Path team is hoping to scrape the surface of that in the upcoming weeks. Once again however, I see this support deriving as a reaction to how the Global Pulse user base starts to use the platform.

-Another very good question. I think it would be quite helpful to make the hunch process as controlled of an environment as possible starting off for a couple reasons. 1) With less variables, the outcomes and process is, at least theoretically, more straightforward to observe and 2) The Global Pulse team can focus on hunches that have a greater chance of a positive outcome, which will give the community a taste of success and possible give it some momentum moving forward.

I tend to lean towards side-stepping problems where I feel there is not enough information to answer confidently. In these cases, I am an advocate of creating a solution open enough for people to begin to form their own solutions organically. The outcomes can then be observed and either mimicked or improved upon. Considering how diverse the user base will be for Global Pulse (both professionally and culturally), this "observe and react" approach seems to be a proper method to avoid spending significant amounts of time designing based on limited understandings. Everyone has their own philosophy and style around this area, but that's how I like to approach it.

Once again, thanks so much for the feedback. Please keep it coming!