Difference between revisions of "Milestone 3 Sky"
(→Top 2 Trust-based Idea)
(→Top 2 Trust-based Idea)
|Line 221:||Line 221:|
==== Top 2 Trust-based Idea ====
==== Top 2 Trust-based Idea ====
[http://crowdresearch.stanford.edu/w/index.php?title=Milestone_3_Sky_TrustIdea_1:_Collective_Statistics_Panel_%28CSP%29 Collective Statistics Panel]
[http://crowdresearch.stanford.edu/w/index.php?title=Milestone_3_Sky_TrustIdea_1:_Collective_Statistics_Panel_%28CSP%29 Collective Statistics Panel
==== Top 2 Power-based Idea ====
==== Top 2 Power-based Idea ====
Revision as of 01:35, 19 March 2015
- 1 Initial Brainstorm
- 1.1 Needs of Workers
- 1.1.1 Trust Based Search
- 1.1.2 Power of Selection before Starting HIT
- 1.1.3 Trust
- 1.1.4 Power
- 1.1.5 Trust
- 1.1.6 Power
- 1.1.7 Power to change by their feedback
- 1.1.8 Power
- 1.1.9 Trust
- 1.1.10 Power
- 1.2 Needs of Requester
- 1.2.1 Trust
- 1.2.2 Power
- 1.2.3 Trust
- 1.2.4 Power
- 1.1 Needs of Workers
- 2 Dive Deeper into Specific Ideas
- 3 Dark Horse idea
Now it's time for your team to brainstorm some ideas based on these needs.
Work with your team to brainstorm as many ideas as you can under two headings: trust and power. Here are some examples of " How might we" questions (a technique which can inspire specific brainstorms) which can drive the generation of your ideas:
- “ How might we” enable workers to trust the requester’s intention to pay?”
- “How might we” enable requesters to trust the results they get back?
Use whatever tools you need - if you can get together in-person, whiteboards and sketchbooks are great tools, while if you're a remote team, services like Google Docs and sketchboard.io should help you with the brainstorming process.
Sketch out enough ideas until you find a set that you’re inspired to explore further. It should be at least 20 ideas total - 10 ideas for trust, and 10 ideas for power. This brainstorm should be wild and broad. Focus not on usability patches, but deeper design innovations.
Needs of Workers
Trust Based Search
Having search based on requesters reputation would increase the trust of worker ---> requester. So we need to give workers this opportunity to search HIT according to their trust level. According to our need finding in milestone 2 workers wants to trust two things. First Requester's intention and second HIT description.
Realtime HIT-Description statistic panel
As we mentioned we need something to fill the gap of trust in workers to HIT description. We can have a panel that have many statistics. This panel must be realtime based on previous HIT attempts. Average completion time, Average Handy score of HIT, Safety score of HIT, Clarity score, Fun score and etc. These panel's element is statistic based on previous HIT so they are trustable. we will explain it in details in nest step.
Having this panel enable workers to search based on trust-based statistics gathered from HITs and other workers that create more trust in workers to believe HIT descriptions.
Realtime Requester-intention statistic panel
As we discussed we need also something to fill the trust gap between workers and requesters. One mechanism is having realtime statistic for requesters like approval rate ( in all of his request ), number of approve/rejection for each HIT (average is not good here), Response rate( In all of his request ), worker's rating average, Number of total request, Total payment , Average hourly payment (Payment /average completion time ).
Having this panel enable workers to search based on trust-based statistics gathered from HITs and other workers that create more trust in workers to believe requester's intention .
Power of Selection before Starting HIT
Items that are listed below are going to balance distribution of power in system among requesters and workers by giving workers to find the best matched HIT to them. Having match maker will also reduce the rejection rate of HIT.
Pining/subscribing requesters, HITs and keywords
n this mechanism workers can pin their favorites keyword , requesters and HIT's for future recommendations. I Its like subscription on specific requester or category of HITs.
Real-Time KeyWord suggestion based on active keywords
Having an efficient tagging system and also tag search system. The key important point in this idea is 'Active Keywords'. Active keywords are existing keywords in current available HIT's. We need to index most valuable keyword and suggest them to workers as they search keyword in the search box.
News Feed based on Workers favorites and History
We have a panel that is always list just created HIT in the time which is related to the user in some way.
Dynamic overall and specific payment statistics panel
Having statistics like Total payment, Average hourly payment , Average time completion can fill this trust gap in future payment by requesters.
Task Pricing Standard
By standardizing the task payment and pricing, based on a comprehensive, quality based per hour price list of the HITs, we ensure the workers of the fair payment.
Mturk should have an ability to show real-time average payment for workers so as long as they are close or above average they feel they are paid in fairness.
A minimum wage schema
There should be a minimal payment for turkers so they can trust they get at least an amount of money out of their work. It might actually affect some turkers in a negative way but I assume it will benefit the majority. This could be implemented globally or a requester should specifically define a minimum amount of payment for each HIT
In Mturk after completing the task, requester was the person who has the power to approve or reject HITs with any reason. By these Ideas below we are going to give some power to worker or restrict requester's power and benefit. This would decrease the rejection rate and consequently workers power.
Power to define rates
Currently workers don't have any means to negotiate payments. We can add a bidding mechanism so workers could come and enter their minimum hourly payment and requester can choose the minimum or they can balance worker quality and payment.
Power to report abuse
Workers should have the ability to report a requester that doesn't pay enough and there should be an ability to check on that. So there should be a god view somewhere so that an authority can check the fairness based on statistics. There could be an automated process that if enough workers report a requester, the requesters account is suspended automatically and be reported for an investigation.
STOPING requesters to use rejected Worker's HIT output efficiently
We need to see workers job as service or product that can be return. Then we need eBay Like payment system, Keep product until both satisfied by the service. We can have two step for delivering the result to he requesters. First step is Filtering stage which they can filter their result based on they criteria and second is exporting the result. In exporting sage we just allow the requester to export output of approved workers. Requester might export all data manually in Filtering stage but its not as easy as before. So requesters for reducing overhead tend to just reject those that are not really useful.
By having a contract-based system, penalizing the requester "Reputation" for unjustified rejection and enabling micro-contracts between worker and requester, we ensure worker is being paid and mediate the power of payment between worker and the requester.
By having a standardized platform for HIT acceptance/rejection, like quality, accuracy, completeness we can resign the power for rejection from the requester and enforce a system-based, standardized platform to accept or reject. Using the method we address the unjustified rejection and emotional involvements as well as unfair payment or payment rejection.
Worker-Requester relationship guideline
Just like any other social situation, there should be some rules that everyone must follow while interacting with each other. This could be implemented as a bot that detects bad language or as a community supervised authority or the owner of the website must impose these guidelines. These rules should be updated based any complaint on bad language and worker abuse.
A rating and review scheme
In order to build trust for workers, there should be a reviewing scheme so that workers could trust requesters for treating them respectfully. There should also be an imposing rule on requesters so that they can't just change account and continue their bad behaviour.
Workers need to have a forum so they can discuss their experiences with requesters and talk with moderates and the owner of the website.
Workers should have the ability to vote for certain requesters as distinguished. Like requesters of the month and requesters of the week, a trusted badge, etc.
Power to change by their feedback
Efficient Up-Vote enabled feedback system to communicate Managers \ requesters
We need to have a feedback system for users. Workers can post their own feedbacks and vote others feedbacks in using system. And make sure that top feedbacks will be considered in next upgrades. These feeds backs are for improving the system and it's functionalities.
Also each HIT can have one page like reddit that is for communication between requester and All of workers. This help workers ask any question or clarification from both requesters and other workers that have done that HIT.
In our opinion there are Four scenario in witch workers have difficulties in exposing their skills. First according to milestone 3 template when they new user without any history of previous job. So in systems like mTurk there are less opportunities cause they are not qualified for lots of HITs. And the second one regardless of workers experience they need to show their previous history in more efficient manner. And third group are those who has high rejection rate for any reason like lack of knowledge in using system or scam requesters. But know they are experienced and we need to give them another chance. And Fourth We have some workers with specific skills that want to participate but with more value.
Tender offer for unqualified , base price for qualified , Auction offer high qualified workers
As you see in Mturk requesters are specifying a qualification level and then filter lots of workers. Here every HIT should have three way to apply the job through three prices. Also here every HIT should have base value which is same as current HIT value in Mturk and also base qualification level which can be like current qualification requirement in MTurk. But there are two more way to get the HIT. First offering lower price than base price from currently unqualified workers. Second are those who think they have specific skills that is valuable for requesters and they are ready to pay more for that.
For base price workers just accept and start to doing that, but Unqualified and also high qualified workers should wait until requester accept their application. workers application should be with a comment to motivate requester to accept it.
Efficient Up-Vote enabled feedback system per HIT (forum per HIT)
We can have forum like page that is for communication between requester and All of workers about issue related to the HIT . This help workers ask any question or clarification from both requesters and other workers that have done that HIT. Its save time for both requester and workers.
HIT comprehensive description
By enabling a description about the goal of the HIT, the procedure the worker should follow and giving the clear directives we can ensure that the user understand the tasks. In one interview one worker said, if we were aware of the goal of the task, we could contribute better and by knowing the objective and the incentive to the HIT the workers feel they understand it better
Mandatory face-to-face meeting
The first milestone in each task should be a mandatory face to face meeting so that workers could resolve any ambiguity in the goals and the description of the task.
HIT Design Standard platform
The standard platform for HIT design can eliminate many confusions a worker might face while doing the HIT and clarifies the HIT by standards, and enforce the worker to provide useful directives and procedure along with the HIT.
Needs of Requester
Online price calculator
Requesters should have access to a transparent online price calculator where they input amount of work required and the quality desired and the calculator shows minimum payment to both requester and worker.
There should be several categories that requesters could put their HIT there. Each category could have an estimated minimum price and average price that requesters could plan accordingly.
There should be contract scheme with several pre-contracts that requesters could negotiate the price with workers so that both could agree upon.
Requesters must have access to a statistical data gathered from the free market or from the website itself that determines the appropriate price for each type of work. For example the requester can see that if he/she wants his work done under a week and he/she has a particular type of work, then it could be easily determined that statistically a certain minimum price should be set.
Requester Reputation Builder
By having a requester reputation builder we enable the previous workers dealing with requester to rank them based on their promptness in payment, HIT design, fairness in payment and responsiveness. What about the new requesters? We set a value for default, the first HIT is the critical point for the requester to build their reputation, if they succeed in the first HIT and get satisfactory rates, they can maintain and even go up, but if they are scammers or "unfair" ones, they will rated downwardly.
An authenticity scheme should be implemented. For example, the owner of the website should provide a mechanism for requesters to provide IDs to the owner and the owner can confirm the authenticity of the requester.
There should be a solid contract between requesters and workers where they can negotiate and come to their terms and there should be a mechanism to compensate for any damage when the contract is breached.
Worker-System Ranking Mechanism
Predefining a set of standards, and qualifications, and terms and agreements, and contracting the requester in the first place, we can give a medium score to the worker to enter the market, and by posting a HIT we enable the Worker or the HIT Micro-Syndicate to vote and rank the requester and change them up. The system and the worker has the power to rate the requester and as mentioned earlier, the wight of the rate coming from the Micro-Syndicate has more importance than the individual weights, to avoid emotional involvement. Also we should enforce requester rating by the worker, maybe provide bonus for giving rank, and prompt them to rate the requester after they got the payment and justifications.
Breach of contract in an event of lack of trust
This seems out of reach but requesters should have a limited power to request a breach of contract if there is a lack of trust between workers and requesters.
Ability to advertise
Requesters should have the power to advertise on the platform. Advertisement is the key to establish trust between workers and requesters.
Dive Deeper into Specific Ideas
For each of the 4 ideas (2 for trust, 2 for power), describe (using diagrams, sketches, storyboards, text, or some combination) the ideas in further detail.
Please create a separate wiki page for each of your ideas, so we can link to them individually. The title of the wiki page should be Milestone 3 followed by your team name and a description of the idea itself (ex: Milestone 3 YourTeamName TrustIdea 1: Automatic Pricing for Tasks based on Average Completion Time). Post a link to each of your trust-related ideas to http://crowdresearch.meteor.com/category/milestone-3-trust-ideas and a link to each of your power-related ideas to http://crowdresearch.meteor.com/category/milestone-3-power-ideas when done.
Top 2 Trust-based Idea
Top 2 Power-based Idea
Dark Horse idea
Now that you've identified some design directions you like, it's time to change tack and toss in a dark horse idea. A dark horse, in horse racing, is a contender who most people don't think will win, but may turn in an unexpectedly strong performance and produce a huge payoff. Dark horse ideas are intended to be something far out there or nearly impossible. In the best case, your dark horse ideas might end up winning the race. However, even in the worst case, they can give us tremendous design insight and prevent design fixation, where the design space shrinks too rapidly.
There are three requirements for dark horse ideas. First, they must be "dark": they must explore a space that is risky, radical, infeasible, and/or in a direction orthogonal to previously explored solutions. They should feel slightly uncomfortable. Second, they must be brainstormed after the more traditional ideas — you can't have a dark path without a traditional "light" path to contrast it against. Third, they must be refined enough that they could be prototyped and objectively tested. That is, it cannot be infeasible: it needs to be something that we could put in front of real people to see whether it would work.
If you have dark horse ideas that came up in your initial brainstorm, you can use that. If you're not satisfied, brainstorm some! Try using Powers of Ten and other techniques to push further and generate even more. After you brainstorm and sketch out dark horse ideas, choose one that you'd like to include among your set of two top candidates from before. Expand on your dark horse idea like you did in the previous section.