Together PRD
PRD for together.ng
By Seun Akinwunmi ·
Together.ng — Product Requirements & Implementation Plan
Document type: Product Requirements Document (PRD) / Functional Implementation Plan Product: Together Working domain: together.ng Status: Detailed product specification Scope: Product behavior, workflows, business rules, functional requirements, and acceptance criteria Explicitly excluded for now: Technology choices, system architecture, infrastructure, implementation code, matching algorithms, AI implementation, and detailed data schemas Vision Document: Together Vision
1. Purpose of This Document
This document describes how Together should function as a complete product.
It is intentionally written independently of technology and implementation architecture. A product manager, designer, engineer, QA engineer, or future technical team should be able to read it and understand:
- What users can do
- Why each capability exists
- What happens at each stage of a process
- What information users provide
- What the application does with that information
- What other users see
- What notifications are generated
- How requests progress
- How contributions are completed
- How reputation is established
- How physical-resource lending is handled
- How moderation and trust work
- What happens in exceptional cases
- How each feature can be tested
The objective is to define the behavior of the product before deciding how it will be built.
2. Product Definition
Together is a non-monetary network that connects people who need help becoming more capable with people who can contribute knowledge, skills, resources, access, time, or experience.
The platform does not require reciprocal exchange.
A person can receive help without giving anything back.
A person can contribute without expecting anything in return.
A person may participate entirely as a recipient, entirely as a contributor, or move between both roles.
The core product loop is:
Need → Request → Discovery/Matching → Contribution → Outcome → Trust → Future Contribution
3. Product Goals
3.1 Primary goals
Together should:
- Make it easy and dignified to ask for productive help.
- Help users formulate specific, actionable requests.
- Connect needs with people who may be able to help.
- Allow people to discover worthwhile needs even when they were not specifically matched.
- Make contribution simple and rewarding without turning it into a monetary system.
- Build trust through observable contribution history rather than popularity.
- Enable safe sharing of physical resources.
- Encourage people who receive help to eventually become contributors when they are able.
- Make useful human capability easier to discover.
- Create a sustainable community culture centered around learning, building, creating, and progressing.
3.2 Secondary goals
Together should eventually:
- Connect individuals with institutions.
- Support local and remote assistance.
- Support both digital and physical resources.
- Make accumulated community knowledge increasingly discoverable.
- Use automation to reduce coordination effort.
- Develop a structured map of skills, needs, resources, and contributors.
4. Non-Goals
The initial product should not become:
- A fundraising platform.
- A personal loan platform.
- A jobs marketplace.
- A freelance marketplace.
- A barter platform.
- A marketplace for buying and selling.
- A platform where users earn credits for helping.
- A follower-based social network.
- A popularity competition.
- A general-purpose social media platform.
- A platform where users are expected to repay help directly.
Monetary transactions between users are outside the core Together model.
5. Core Product Principles
5.1 Asking for help is acceptable
A user does not need to demonstrate usefulness before asking for help.
There must be no public indicator that implies a person is less valuable because they have helped nobody.
5.2 Contribution is voluntary
A user may receive help without subsequently contributing.
Together can encourage contribution but should not create debt.
5.3 Help should increase capability
The platform should favor needs involving:
- Learning
- Education
- Skills
- Knowledge
- Tools
- Resources
- Access
- Mentorship
- Projects
- Creation
- Productivity
- Collaboration
5.4 Trust should describe behavior
Trust indicators should answer:
“Has this person historically been reliable?”
They should not answer:
“How popular is this person?”
5.5 Accessibility before complexity
Anyone should be able to participate in low-risk activities without completing extensive verification.
Additional requirements can be introduced when an activity creates additional risk.
5.6 Human dignity
The product language should avoid terms and interaction patterns associated with begging, charity ranking, pity, or social inferiority.
6. User Types
These are participation roles rather than permanent account types.
6.1 Recipient
A person currently asking for help.
Examples:
- Student seeking a textbook.
- Beginner seeking mentorship.
- Builder seeking access to equipment.
- Person seeking advice from an experienced professional.
6.2 Contributor
A person offering help.
Examples:
- Teacher offering tutoring.
- Engineer offering code reviews.
- Person offering an unused book.
- Person lending a tool.
6.3 Supporter
A user who helps indirectly.
Examples:
- Upvoting a worthwhile request.
- Sharing a request.
- Reporting useful resources.
- Confirming that a contribution was helpful.
A user can occupy all three roles.
7. Account Creation and Onboarding
7.1 Registration
A new user creates an account using the supported registration method.
The registration process should collect only information necessary for normal participation.
The user should be introduced to Together's philosophy during onboarding.
The product should explicitly communicate:
You can ask for help even if you currently have nothing to offer.
This is important because the onboarding experience establishes the culture of the platform.
7.2 Basic profile
The user can provide:
- Name/display name
- Profile photo, if desired
- General location, where useful
- Short description
- Areas of interest
- Skills they can help with
- Resources they can potentially share
- Preferred forms of contribution
A user should not be required to complete every section.
7.3 Contribution capabilities
The user can select one or more predefined categories.
Examples:
- Education
- Technology
- Science
- Engineering
- Business
- Arts
- Design
- Writing
- Languages
- Skilled trades
- Books
- Tools
- Equipment
- Career guidance
- Entrepreneurship
- Research
- Local knowledge
The taxonomy should be designed to evolve.
Within a category, the user can specify individual capabilities.
Example:
Technology
- Python
- JavaScript
- Backend development
- Linux
- Databases
- Code review
7.4 Contribution availability
Where relevant, the user can indicate:
- Online
- In person
- Both
- Preferred geographic area
- Willingness to lend resources
- Willingness to mentor
- Willingness to answer questions
- Willingness to collaborate
7.5 No-contribution onboarding
If a user selects no contribution categories, onboarding still completes normally.
The product should reinforce:
“That's okay. You can simply use Together when you need help.”
No negative score, warning, or shame indicator should be displayed.
8. User Profile
A profile should represent a person's participation and trust rather than function as a conventional social-media profile.
8.1 Profile information
A profile may display:
- Name
- Photo
- Short description
- General location
- Skills
- Contribution categories
- Verified skills
- Contributor since date
- People helped
- Successful contributions
- Relevant badges
- Resource-sharing history where appropriate
8.2 What should not be central
The profile should not emphasize:
- Followers
- Following
- Number of likes received
- Wealth
- Social status
- Popularity
- Monetary value of contributions
8.3 Privacy
Users should control what personal information is publicly visible.
Exact addresses should never be publicly exposed for resource exchanges.
Exact contact information should remain private unless users choose to share it through the appropriate interaction.
9. Creating a Request for Help
Creating a request is one of the most important workflows in Together.
9.1 Start request
The user selects:
Ask for help
The application explains that a specific request is more likely to receive useful responses.
9.2 Select category
The user selects the most relevant category.
The user should be able to search categories if the list is large.
The application should allow a request to be classified into the most appropriate category rather than requiring users to understand the full taxonomy.
9.3 Describe the goal
The user explains what they are trying to accomplish.
Example:
“I am trying to learn electronics and build my first simple circuit.”
9.4 Describe the barrier
The user explains what is preventing progress.
Example:
“I understand the theory but don't have anyone experienced who can guide me through practical work.”
9.5 Describe the requested help
The user states what would help.
Example:
“I need someone experienced with basic electronics who can answer questions and guide me through my first project.”
The user should not have to know the exact resource needed. Together should allow people to describe the problem and let potential contributors suggest appropriate forms of help.
9.6 Optional request details
Depending on category, the request may ask for:
- Location
- Remote/online preference
- Time commitment
- Duration
- Deadline
- Skill level
- Intended outcome
- Resource specifications
- Quantity
- Availability
- Whether the user needs to borrow, receive, access, learn, or collaborate
Only relevant questions should be shown.
9.7 Request preview
Before publishing, the user sees the complete request and can edit it.
The application should highlight missing information that may reduce the likelihood of receiving help.
9.8 Publishing
The user publishes the request.
Once published:
- It becomes discoverable according to visibility settings.
- Relevant matching contributors become eligible for notifications.
- It can appear in category feeds.
- It can receive votes.
- Users can share it.
- Potential contributors can respond.
10. Request Quality Guidance
Together should actively improve requests rather than merely reject poorly written ones.
A user who writes:
“I need a laptop.”
should receive guidance such as:
“What are you trying to do with the laptop?”
The goal is not to interrogate the user.
The application should progressively help them provide useful context.
A high-quality request should make it easy for someone to answer:
- What does this person want to accomplish?
- Why are they currently blocked?
- What kind of help would move them forward?
- What constraints matter?
- Where and when is the help needed?
11. Request States
A request should have an explicit lifecycle.
Recommended states:
- Draft
- Published
- Receiving responses
- Help arranged
- In progress
- Completed
- Closed
- Cancelled
- Archived
- Under review, when moderation is required
The user should be able to understand the current state at all times.
12. Request Discovery
Users should be able to discover needs without receiving a direct match.
12.1 General featured feed
The featured area displays requests that the community may find worthwhile.
The feed should prioritize useful discovery rather than simply displaying the newest posts.
12.2 Category feeds
Each category has its own request feed.
For example:
- Education
- Technology
- Books
- Engineering
- Arts
12.3 Search
Users can search requests using natural descriptions and structured categories.
Search should support finding both:
- Requests that match a user's capabilities.
- Requests a user is personally interested in.
12.4 Filters
Potential filters include:
- Category
- Location
- Online/in-person
- Type of help
- Newest
- Most supported
- Needs still open
- Resource type
- Skill level
13. Voting
Any eligible user can upvote a request they believe deserves attention.
An upvote means:
“I think this need is worthwhile and should receive visibility.”
It does not mean:
- The voter can satisfy the request.
- The voter owes help.
- The voter has donated anything.
- The request is automatically valid.
A user should be able to remove their vote.
The application should prevent repeated voting intended to artificially inflate visibility.
14. Sharing Requests
Users should be able to share requests externally.
The shared representation should contain enough context for someone to understand the need and return to Together.
Sharing is another discovery mechanism.
A person who cannot help may know someone who can.
15. Matching
Together should maintain a relationship between requests and potentially relevant contributors.
Matching should consider information such as:
- User-selected skills
- Contribution categories
- Specific capabilities
- Resource types
- Location
- Remote/in-person preferences
- Availability
- Request requirements
- Relevant past activity
- User notification preferences
A match does not automatically mean the contributor has agreed to help.
It means:
This request appears relevant to what this person said they may be able to contribute.
The user should remain in control of whether to respond.
16. Contributor Notifications
When a new request appears to match a user's stated capabilities, the user may receive a notification.
Notifications should explain why the request is relevant.
Example:
“Someone is looking for help with beginner Python. You indicated that you can help with Python.”
Users should be able to control:
- Categories they receive notifications for
- Frequency
- Local versus remote requests
- Resource-lending notifications
- Mentorship notifications
- Other contribution types
The goal is to avoid notification fatigue.
17. Responding to a Request
A potential contributor selects:
I can help
They should be able to explain how they can help.
Example:
“I can help you with the first three weeks of the course. I'm available online on Saturday afternoons.”
The contributor should not have to commit to the entire request.
They may offer a narrower form of assistance.
18. Multiple Contributors
A request can receive multiple offers.
The requester should be able to:
- Review offers
- Ask questions
- Accept one or more contributors
- Decline offers
- Keep the request open
- Close the request once sufficient help has been arranged
The system should not assume the first response is automatically the best response.
19. Communication
Once a contributor and requester decide to proceed, Together should provide a communication channel.
Communication should support:
- Clarifying the request
- Agreeing on what will be provided
- Coordinating timing
- Sharing relevant information
- Confirming completion
Users should not be forced to publicly expose personal contact details.
20. Completing a Contribution
After the help takes place, the contribution should be marked as completed.
Either participant can initiate completion.
The other participant can confirm.
A completed contribution should record enough information to establish that the interaction occurred without unnecessarily exposing private conversation details.
21. Outcome Confirmation
The recipient should be asked:
Did this help you move forward?
Possible responses:
- Yes, significantly
- Yes, somewhat
- Not yet
- No
The recipient can optionally explain what happened.
This feedback should be more important than simple likes.
The objective is to determine whether Together is actually helping people progress.
22. Contributor Feedback
The recipient can provide structured feedback about the contribution.
Potential dimensions:
- Helpful
- Reliable
- Clear
- Respectful
- Completed as agreed
This information contributes to trust signals.
It should not become a public star-rating competition.
23. Reputation
Reputation is accumulated from meaningful participation.
Potential reputation signals:
- Contributor since
- People helped
- Successful contributions
- Completed contributions
- Verified skills
- Resource lending history
- Relevant badges
- Community confirmations
A reputation indicator should never represent a monetary balance.
There should be no concept of:
“You have 50 points and therefore can request 50 units of help.”
24. Verified Skills
A skill can become verified through meaningful evidence.
Possible evidence sources include:
- Demonstrated work
- Completed contributions
- Community confirmations
- Relevant credentials where appropriate
- Successful mentorship
- Other trustworthy evidence
Verification should communicate:
“There is evidence that this person can help with this skill.”
It should not imply professional licensing unless the relevant credential has actually been verified.
25. Badges
Badges may recognize contribution patterns.
Examples:
- Early Contributor
- Helpful Mentor
- Skill Contributor
- Book Sharer
- Resource Lender
- Consistent Contributor
- Community Supporter
Badges should be descriptive and celebratory.
They should not provide economic value or unlock unfair social status.
Badges should not be required for participation.
26. Contribution Without Reciprocity
After a user receives help, Together may encourage them to contribute when they are able.
For example:
“Glad you got the help you needed. Is there anything you know or have that someone else might find useful?”
This should be an invitation, not an obligation.
There should be no countdown, debt, or expectation that the person must give back.
27. Resource Sharing
Together supports sharing of physical and digital resources.
Resource types can include:
- Books
- Tools
- Computers
- Cameras
- Musical instruments
- Equipment
- Educational materials
- Other useful items
The contributor specifies whether the resource is:
- Given away
- Lent temporarily
- Available for supervised use
- Available for local use only
- Available remotely/digitally where applicable
28. Lending Workflow
Physical lending requires additional safeguards.
Step 1 — Resource listing
The owner describes:
- Resource
- Condition
- Intended use
- Location
- Availability
- Lending duration
- Restrictions
- Relevant accessories
Step 2 — Request
A user asks to borrow it and explains the intended use.
Step 3 — Trust assessment
Depending on resource value and risk, Together may require additional verification or an established history of successful participation.
Step 4 — Agreement
Both parties confirm:
- What is being lent
- Expected condition
- Lending period
- Return expectations
- Any special handling requirements
Step 5 — Condition record
The owner records the item's condition using descriptions and, where appropriate, photographs.
The borrower acknowledges the condition.
Step 6 — Handoff
Both parties confirm that the resource was handed over.
Step 7 — Return
The borrower returns the resource.
The return condition is recorded.
Step 8 — Confirmation
The owner confirms whether the resource was returned satisfactorily.
Step 9 — Contribution history
A successful lending interaction becomes part of the relevant trust history.
29. High-Value Resource Rules
High-value resources should not be treated like books.
The platform may require:
- Identity verification
- Established account history
- Successful previous contributions
- Additional agreement
- Additional safety checks
- Higher trust requirements
New users should not automatically gain access to high-value resources.
The exact thresholds can be determined during implementation based on real-world usage and risk.
30. Resource Damage or Loss
If an item is returned damaged, lost, or not returned:
- The owner reports the issue.
- The borrower is notified.
- Both parties can provide relevant information.
- The existing condition and lending records are reviewed.
- The issue can enter a dispute process.
- Together may restrict one or both accounts while investigating.
- Serious violations can result in suspension or permanent removal.
- Where appropriate and legally necessary, relevant information can be provided to authorities.
Together should not automatically decide every dispute in favor of the owner or borrower.
The objective is fair handling based on available evidence.
31. Repeat Abuse and New Accounts
Reputation alone cannot prevent someone from abandoning an account and creating another.
Together should therefore treat trust as more than a public profile metric.
High-risk activities can require stronger account verification.
New accounts should have limited access to activities where misuse could cause substantial loss.
The platform should detect and restrict suspicious patterns where possible.
The product should prioritize:
Preventing a new account from immediately gaining access to high-risk resources
rather than attempting to make it impossible for anyone to ever create multiple accounts.
32. Insurance
Insurance is not part of the basic Together experience.
It may eventually be offered through external partnerships for high-value physical resources.
Potential future options:
- Community lending without additional protection
- Protected lending with optional insurance
Together should not assume responsibility for insuring all resources itself.
This feature should only be introduced if real usage demonstrates a need.
33. Institutions
Organizations should eventually be able to participate.
Potential institutions include:
- Universities
- Schools
- Libraries
- Companies
- Workshops
- Laboratories
- Makerspaces
- Professional associations
- Nonprofits
- Community organizations
They may contribute:
- Equipment
- Training
- Books
- Mentors
- Facilities
- Expertise
- Access
Institutional accounts should have appropriate verification and administrative capabilities.
34. Moderation
Together needs moderation because not every request or interaction will be appropriate.
Moderation should address:
- Fraud
- Harassment
- Abuse
- Illegal activity
- Misrepresentation
- Spam
- Commercial solicitation
- Dangerous activities
- Manipulative requests
- Inappropriate content
- Attempts to circumvent platform rules
Moderation should distinguish between:
A person asking for help
and:
A person abusing the platform.
Need alone should never be treated as evidence of wrongdoing.
35. Reporting
Users should be able to report:
- Requests
- Profiles
- Contributions
- Messages
- Resource listings
- Other interactions
A report should allow the user to select a reason and optionally provide context.
The reporter should not be publicly identified to the reported user unless there is a legitimate reason.
36. Safety Rules
The platform should have clear rules covering:
- Physical meetups
- Resource handoffs
- Personal information
- Dangerous activities
- Professional advice
- Minors
- Illegal requests
- Financial requests
- Medical or other high-risk situations
The platform should not imply professional guarantees where contributors are simply community members.
37. Notifications
Notifications should exist for meaningful events.
Potential notifications include:
Requester
- A potential contributor responded.
- Someone asked a clarifying question.
- A contribution was accepted.
- A contribution is ready to begin.
- A contribution was marked complete.
- A contributor needs confirmation.
- A request is receiving significant community interest.
Contributor
- A matching request was published.
- Someone responded to their contribution offer.
- Their offer was accepted.
- A lending interaction requires confirmation.
- A request they follow has changed.
Community
- A followed request needs attention.
- A contribution has been completed where relevant.
Users should be able to control notification frequency and categories.
38. Following Needs Without Building a Follower Culture
Users may want to follow a request or topic so they can return later.
This should be treated as following a need, not following a person.
The product should avoid turning this into a social follower graph.
Examples:
Follow this request
Follow Technology needs
Follow Book requests
These relationships exist for discovery and participation, not social status.
39. Request Closure
A requester can close a request when:
- Help was received.
- The need no longer exists.
- They solved the problem independently.
- They found help elsewhere.
- The request was created by mistake.
The closure reason can be recorded.
If help was received, the requester should be encouraged to confirm the outcome.
Closed requests remain useful as historical contribution records unless removed for privacy, safety, or moderation reasons.
40. Unsuccessful Requests
Not every request will receive help.
A request can remain open for an appropriate period.
The application may periodically ask:
“Do you still need help with this?”
The user can:
- Keep it open
- Update it
- Close it
- Mark it solved elsewhere
The platform should not imply that failure to receive help means the request was invalid or the requester was undeserving.
41. Request Editing
A requester should be able to edit an open request.
Significant changes should be visible to people who have already responded when necessary.
If the fundamental need changes, the application may suggest creating a new request instead.
42. Duplicate and Related Requests
The same need may be posted by multiple people.
Together should make it possible to identify related requests.
Related requests may eventually help users discover:
- Previous solutions
- People who helped others with similar needs
- Resources
- Repeated community needs
This creates a path from individual requests toward reusable community knowledge.
43. Community Knowledge
When a particular need repeatedly appears, Together should eventually be able to surface accumulated knowledge around it.
For example:
“How do I learn electronics?”
could lead to:
- Existing requests
- People who have helped
- Recommended resources
- Local contributors
- Institutions
- Previous successful solutions
The platform can gradually transform repeated human interactions into discoverable knowledge.
44. AI-Assisted Product Experiences
AI implementation is intentionally outside the current technical scope, but the product should reserve functional opportunities for future automation.
Potential experiences include:
- Improving vague requests
- Asking relevant clarifying questions
- Categorizing requests
- Identifying likely skills needed
- Finding related requests
- Explaining why a contributor is a potential match
- Summarizing long requests
- Helping contributors understand how they could help
- Finding relevant community resources
These capabilities should assist coordination rather than replace human contribution.
45. Search and Discovery of Contributors
A requester should eventually be able to discover people who may help.
A contributor profile should show relevant capability and trust information.
The requester should not need to know the person's name.
They should be able to search by:
- Skill
- Category
- Location
- Contribution type
- Resource
- Availability
The system should prioritize relevant capability rather than popularity.
46. Local Versus Remote Help
Every relevant request should distinguish between forms of help.
Examples:
Remote
- Mentorship
- Code review
- Tutoring
- Research guidance
Local
- Borrowing a tool
- Using equipment
- Hands-on instruction
- Physical resource exchange
Either
- Some forms of knowledge sharing
- Books
- Some mentorship
- Some collaboration
Location should only be required when it materially affects the request.
47. User Privacy
Together should collect and expose information according to the minimum needed for successful coordination.
Public profiles should not reveal:
- Exact residential addresses
- Private phone numbers
- Private email addresses
- Sensitive identity information
Physical exchanges may require users to exchange additional information privately after choosing to proceed.
48. Blocking
Users should be able to block another user.
Blocking should prevent unwanted interaction according to the applicable product rules.
A blocked user should not be able to use ordinary platform features to repeatedly contact the blocker.
Blocking should not remove legitimate safety-reporting capabilities.
49. Trust and Safety Escalation
Trust issues should have levels.
Low-level issue
Example:
- Poor communication
- Missed appointment
Potential response:
- Feedback
- Warning
- Reduced trust signal
Moderate issue
Example:
- Repeated failure to return resources
Potential response:
- Lending restriction
- Temporary suspension
- Additional verification
Serious issue
Example:
- Theft
- Fraud
- Threats
- Serious harassment
Potential response:
- Immediate restriction
- Investigation
- Permanent removal where appropriate
- Legal escalation where required
50. Admin Functions
Administrators should be able to:
- Review reports
- Review accounts
- Suspend accounts
- Restore accounts
- Restrict specific capabilities
- Review disputed contributions
- Review high-risk resource activity
- Manage categories
- Manage badges
- Manage moderation rules
- Feature requests
- Remove harmful content
- Review platform metrics
- Investigate abuse patterns
Administrative actions should be auditable.
51. Category Management
Categories should not be permanently fixed.
Administrators should be able to:
- Add categories
- Rename categories
- Merge categories
- Retire categories
- Add subcategories
- Add skills
- Mark obsolete skills
- Associate related categories
Changes should avoid destroying historical records.
52. Request Quality and Moderation Are Different
A request can be poorly written without being inappropriate.
For example:
“Need a computer to learn coding.”
This is a low-quality request but not necessarily a moderation problem.
The application should first attempt to help the user improve it.
Moderation should be reserved for requests that violate platform rules.
This distinction is important to maintaining a welcoming culture.
53. Abuse of Voting
Voting should not become a mechanism for manipulation.
The product should detect or restrict:
- Repeated voting
- Artificial voting campaigns
- Multiple-account manipulation
- Coordinated abuse
Voting should affect discovery but should not determine whether someone deserves help.
54. Success Metrics
Primary metric
Successful contributions
A successful contribution is an interaction where:
- A meaningful need is identified.
- Someone offers help.
- The help is arranged.
- The contribution occurs.
- The recipient confirms that it helped them move forward.
Supporting metrics
Track:
- Number of requests
- Percentage receiving at least one response
- Time to first useful response
- Percentage of requests resulting in successful contributions
- Number of contributors
- Number of recipients
- Repeat contributors
- Repeat recipients
- Number of people helped
- Resource-lending completion rate
- Resource-loss/damage rate
- Percentage of recipients who later contribute
- Category activity
- Local versus remote contribution activity
- Request quality
- Moderation incidents
55. Core User Journeys
Journey A — New user needs help
- User registers.
- User sees that having nothing to contribute is acceptable.
- User selects Ask for Help.
- User selects a category.
- User describes their goal.
- User explains their barrier.
- Together guides them toward a specific request.
- User provides relevant constraints.
- User publishes the request.
- Potential contributors receive relevant notifications.
- Other users discover and upvote the request.
- Contributors respond.
- User evaluates responses.
- User accepts help.
- Communication begins.
- Help takes place.
- Contribution is marked complete.
- User confirms whether it helped.
- Contributor's successful contribution history is updated.
- User is later invited—but not required—to contribute something themselves.
Journey B — User wants to help generally
- User opens featured needs.
- User browses requests.
- User filters by category if desired.
- User finds a request they can help with.
- User opens it.
- User selects I Can Help.
- User describes their proposed contribution.
- Requester receives the response.
- Requester accepts or declines.
- Communication occurs.
- Help is delivered.
- Both parties confirm completion.
- Recipient confirms outcome.
- Contributor receives appropriate trust/reputation credit.
Journey C — User sees a need they cannot solve
- User discovers a request.
- User cannot personally provide the requested help.
- User upvotes the request.
- Optionally, user shares it.
- The request receives additional visibility.
- Another person discovers it and contributes.
Journey D — User lends a book
- User lists a book as available to lend.
- Another user finds the resource.
- Borrower explains intended use.
- Owner reviews the request.
- Both agree to the lending period.
- Condition is documented.
- Handoff is confirmed.
- Borrower uses the book.
- Book is returned.
- Condition is confirmed.
- Lending interaction is recorded as successful.
Journey E — High-value resource lending
- Owner lists a valuable resource.
- Borrower requests it.
- Together determines that stronger trust requirements apply.
- Borrower completes required verification.
- Borrower's eligibility is established.
- Owner reviews borrower information.
- Both agree to terms.
- Condition is documented.
- Handoff is confirmed.
- Resource is returned.
- Condition is checked.
- Any issue is either closed successfully or escalated into dispute handling.
- Successful completion contributes to future trust.
56. Acceptance Criteria for the MVP
The MVP should not be considered functionally complete until a user can perform the complete core loop.
Account
- A user can register.
- A user can create a profile.
- A user can specify categories and capabilities they can help with.
- A user can leave contribution capabilities empty.
Requests
- A user can create a request.
- A user can select a category.
- A user can describe their goal and barrier.
- The application guides users toward specific requests.
- A user can publish, edit, and close a request.
- A request has a clear lifecycle state.
Discovery
- Users can browse featured needs.
- Users can browse category feeds.
- Users can search requests.
- Users can upvote requests.
- Users can share requests.
Matching
- Contributors can specify capabilities.
- Relevant requests can be presented to potential contributors.
- Contributors can receive notifications for relevant requests.
- Contributors can respond to requests.
Contributions
- A requester can review contributor offers.
- A requester can accept an offer.
- Users can communicate after an offer is accepted.
- A contribution can be marked complete.
- The recipient can confirm whether it helped.
Reputation
- Successful contributions are recorded.
- Profiles display contribution history.
- People helped can be represented.
- Contributor-since information is displayed.
- Verified skills can eventually be represented.
- No monetary reputation currency exists.
Safety
- Users can report content or accounts.
- Users can block other users.
- Administrators can review reports.
- Administrators can restrict accounts.
- Physical-resource interactions have additional safeguards.
57. Detailed Testing Strategy
The product should be tested as complete user journeys rather than only isolated screens.
57.1 Registration tests
Verify:
- Registration succeeds with valid information.
- Invalid registration information is handled correctly.
- Users can skip contribution capabilities.
- Users who offer no help are treated normally.
- User profiles are created correctly.
57.2 Request tests
Verify:
- A request cannot be published without the minimum required information.
- The request guide helps users provide useful context.
- Relevant optional fields appear for relevant categories.
- Location is not unnecessarily required.
- Published requests become discoverable.
- Editing works.
- Closing works.
- Cancelled requests behave correctly.
57.3 Discovery tests
Verify:
- Featured requests appear.
- Category requests appear in the correct category.
- Search returns relevant requests.
- Filters work.
- Upvoting works.
- Removing an upvote works.
- Users cannot artificially vote multiple times.
- Closed requests are handled appropriately.
57.4 Matching tests
Verify:
- A contributor receives relevant requests.
- An irrelevant contributor does not receive excessive notifications.
- Location requirements are respected.
- Online-only and in-person preferences are respected.
- Notification preferences are respected.
- Contributors can respond.
- Requesters can review responses.
57.5 Contribution tests
Verify:
- A requester can accept a contributor.
- Communication becomes available at the appropriate stage.
- A contribution can be completed.
- Both parties can confirm completion.
- The recipient can report the outcome.
- Successful contribution history updates correctly.
57.6 Reputation tests
Verify:
- Successful contributions are recorded.
- People-helped counts update appropriately.
- Failed or cancelled interactions do not incorrectly count as successful.
- Reputation cannot be purchased.
- Reputation cannot be converted into monetary value.
- Users cannot artificially inflate reputation.
57.7 Lending tests
Verify:
- A resource can be listed.
- A borrower can request it.
- Lending terms can be established.
- Condition can be recorded.
- Handoff can be confirmed.
- Return can be confirmed.
- Damage/loss can be reported.
- Disputes can be escalated.
- High-value resources trigger appropriate safeguards.
57.8 Safety tests
Verify:
- Users can report content.
- Reports reach appropriate administrative workflows.
- Users can block others.
- Suspended users lose the appropriate capabilities.
- Serious violations trigger appropriate escalation.
- A banned user cannot trivially regain high-risk privileges through a new account.
58. Edge Cases
The implementation should explicitly account for:
- A request receiving no responses.
- A request receiving too many responses.
- Multiple contributors helping one requester.
- A requester changing their mind.
- A contributor withdrawing.
- A contributor failing to appear.
- A resource becoming unavailable.
- A resource being damaged.
- A resource being lost.
- A user deleting their account during an active contribution.
- A request being reported.
- A contributor being reported during an active contribution.
- Duplicate requests.
- Duplicate accounts.
- Spam requests.
- Malicious voting.
- A request being solved outside Together.
- A request becoming obsolete.
- A user being unable to verify identity.
- A high-value resource being requested by a new account.
- A physical exchange being geographically impossible.
- A contributor offering only part of the requested help.
- A request requiring a professional qualification.
- A user attempting to use Together for monetary fundraising.
Each case should have a defined user-facing outcome and administrative behavior.
59. Product Language
Language is part of the product's identity.
Prefer:
- Ask for help
- Need
- Contribute
- Help someone
- Share
- Lend
- Learn
- Build
- Grow
- Progress
- People helped
- Contribution
- Community
- Supported
Avoid language that implies:
- Charity hierarchy
- Debt
- Financial exchange
- Begging
- Social status
Examples of language to avoid:
- Earn points
- Spend credits
- Repay
- Donor level
- Beggar
- Charity score
- Poor users
- Rich users
60. MVP Scope
The first version should focus on validating the central behavior.
Include
- Registration
- Profiles
- Skill/category selection
- Request creation
- Request guidance
- Request publishing
- Featured requests
- Category feeds
- Voting
- Search
- Contributor matching
- Notifications
- Contributor responses
- Communication
- Contribution completion
- Outcome confirmation
- Basic reputation
- Basic reporting
- Blocking
- Basic administration
Carefully limited
- Physical-resource lending
- High-value resources
- Identity verification
- Complex disputes
These should be implemented conservatively because they create significant trust and operational complexity.
Defer
- Insurance
- Institutional accounts
- Advanced AI coordination
- Sophisticated knowledge extraction
- Complex recommendation systems
- Advanced verification
- Large-scale resource logistics
61. Phased Product Development
Phase 1 — Core community loop
Validate:
Ask → Match → Help → Confirm
Features:
- Registration
- Profiles
- Skills
- Requests
- Request guidance
- Discovery
- Voting
- Basic matching
- Notifications
- Responses
- Communication
- Completion
- Basic reputation
Phase 2 — Better trust and resource sharing
Validate:
People can safely share tangible resources.
Features:
- Resource listings
- Lending workflow
- Condition records
- Trust requirements
- Identity verification
- Dispute workflows
- Lending history
Phase 3 — Network expansion
Expand from individuals to:
- Institutions
- Libraries
- Workshops
- Universities
- Companies
- Community organizations
Phase 4 — Intelligent coordination
Introduce automation to:
- Improve requests
- Understand needs
- Match people
- Surface resources
- Connect related needs
- Build reusable knowledge
62. The Most Important Product Loop to Validate
Before building the full platform, the most important question is:
Will strangers actually help strangers with productive needs when money is deliberately removed from the interaction?
The MVP should therefore optimize for learning this.
A successful early community should demonstrate examples such as:
Person A asks for help learning something.
Person B sees the request.
Person B offers help.
Person A progresses.
The interaction is confirmed.
Then:
Person A eventually becomes Person C's helper.
If that loop happens naturally, Together has evidence that its core social model works.
63. Product North Star
Together should ultimately make this statement true:
If there is something useful a person needs to learn, build, create, or accomplish, there is a reasonable chance that Together can connect them with someone who can help.
And equally:
If someone has knowledge, skills, resources, experience, or access they are willing to share, Together should make it easier for the people who need it to find them.
The product succeeds when these two capabilities increasingly connect.
64. Final Product Philosophy
Together is not trying to eliminate people's needs.
It is trying to eliminate the unnecessary barriers between people and the people who can help them overcome those barriers.
The platform should make asking easier.
It should make contributing easier.
It should make useful needs easier to discover.
It should make trustworthy contributors easier to identify.
It should make dormant resources easier to put to productive use.
And it should create a culture where receiving help is not a position of weakness and giving help is not a transaction.
The fundamental relationship remains:
I can help you.
You can help someone else.
Together, we can all become more capable.