Positioning
Does it plainly say what you are, who you are for and why you are different?
Quiver quickstart
Getting Started with Quiver
Start by making Quiver's product context trustworthy. Do not begin with a blank prompt asking for ten marketing ideas. Build one small, connected developer marketing loop instead.
The minimum useful loop
Start with the system, not a blank prompt.
Quiver's advantage is that product context, customer evidence, campaigns, content, tasks and results stay connected. Your first useful workflow should preserve those relationships from the beginning.
01
Confirm the context
Open Context, read what Quiver drafted from your site and correct every field marked as an assumption.
02
Add customer evidence
Paste one raw customer source—a call transcript, a support thread, a review—into Research, unpolished.
03
Create one real campaign
Create a campaign for an initiative you are actually running, so the work has somewhere to belong.
04
Run a focused session
Start a session in the mode that matches the job, and run Feedback or Strategy before you run Create.
05
Save the work
Save the useful output as an artifact or a content piece and link it to the campaign.
06
Measure what happens
Log the result after the work ships, so the next session starts from what actually worked.
Step 1
Make the context trustworthy.
During setup, give Quiver your website or a short product description. It drafts the positioning, audience, messaging, customer language, proof and hypotheses that future sessions will use.
The initial context is a draft, not unquestionable truth. Open Context and review the fields marked Assumption before generating production work.
Ideal customer profile
Is it specific enough to exclude buyers who are not a fit?
Core problem
Is this the problem customers recognize, or merely the problem you hope they have?
Proof points
Can you support every claim listed here?
Active hypotheses
What do you currently believe but still need to test?
A useful stopping rule
The context is ready when a capable teammate could read it and produce work that sounds like your company—not when every field is exhaustive. Every save creates a version, so you can improve it without erasing its history.
Step 2
Add one piece of real customer evidence.
Generated context tells Quiver what the company currently believes. Customer research tells it what the market is actually saying. Add one meaningful source before asking Quiver to create a campaign or write content.
- A customer call transcript
- Interview notes
- A support conversation
- Survey responses
- A review
- A sales objection
- A relevant community discussion
Paste the raw material into Research. Do not polish it into marketing language first. The useful evidence is often the exact phrase a customer used, the objection they repeated or the part they misunderstood.
The faster route
Import it through MCP.
If you connected an MCP-compatible agent during setup, use it to move existing research into Quiver without manually reformatting every document.
Log this customer call as research. Extract the main themes and Voice of Customer language, then show which active hypotheses it validates or challenges. Do not update the active context without my approval.Read the MCP setup guide
Step 3
Create one real campaign.
Do not invent a practice project. Pick an initiative already on your list: launch a feature, improve trial activation, reposition a page, promote an event or turn customer research into a content series.
Give the campaign one outcome, an owner, a status, relevant dates and the channels you expect to use. A campaign is the scope connecting the sessions, research, artifacts, content, tasks, distribution, performance and decisions behind that initiative.
Example
- Campaign
- Product Hunt follow-up
- Goal
- Help new Quiver users reach their first useful workflow.
- Evidence
- Repeated questions about what to do first.
- Deliverable
- A focused getting-started guide.
- Success signal
- More new users confirm context and add research before beginning a Create session.
Step 4
Run the right first session.
Choose the mode based on the job, not the novelty of the model. Feedback or Strategy is usually a better first session than Create because it improves the evidence or brief first.
You have customer input
Feedback
Start here after adding a transcript, survey, support thread or set of objections.
Analyze the new research entry. Separate direct evidence from interpretation. Identify repeated customer language, product signals and any active hypotheses the evidence validates or challenges. Propose context changes, but do not apply them.
The decision is unclear
Strategy
Use Strategy when you know the outcome you need but not the smallest sensible plan.
Using the active product context and the research attached to this campaign, recommend the smallest campaign that can test whether [hypothesis]. Define the audience, message, channel, success metric and what evidence would change the decision.
The brief is ready
Create
Use Create for production work only after the audience, evidence, outcome and campaign scope are clear.
Draft [the deliverable] for [the audience]. Use the active context and the evidence attached to this campaign. Keep every claim supportable and end with [the desired next action].
Use Analyze after results exist. Use Optimize when a page, email, draft or onboarding flow already exists and needs improvement against the active context.
Step 5
Save the work.
Do not leave a useful answer in chat. Save production work as an artifact or content piece, link it to the campaign and move it through its explicit state:
Draft → Review → Approved → Live → Archived
Artifacts go Live. Content pieces go Published. Everything else in the sequence is the same.
Step 6
Close the loop.
After the work ships, record where it went, log quantitative and qualitative results, synthesize what happened and review any proposed context changes. Quiver preserves the evidence; a person still decides what becomes true.
Common failure modes
What not to do first.
Do not import everything
Start with minimum trustworthy context and one strong source of customer evidence.
Do not ask for generic ideas
Give the session a real outcome, relevant evidence and a campaign scope.
Do not trust every inference
Review assumptions before they appear in downstream work.
Do not create ten campaigns
One campaign completed through measurement will teach you more than a perfect backlog.
Do not start with content volume
Start with a decision or customer problem. Content is an output of the system.
Your first 30 minutes
Complete the first useful loop.
- Connect your model provider.
- Build the initial context from your website or product description.
- Review the fields marked as assumptions.
- Confirm positioning, ICP, core problem, proof points and hypotheses.
- Add one customer call, survey or support thread.
- Create one campaign for a real initiative.
- Run a Feedback or Strategy session before a Create session.
- Save the useful output and link it to the campaign.
- Define the result you will log after the work ships.
If you complete those steps, you have not merely configured another AI tool. You have begun a developer marketing feedback loop that can preserve its context, evidence, work and learning.