What if every piece of product marketing research your team needs could be triggered by a single command?
The Problem with Ad Hoc Research
Product marketing teams do not lack for research tools. They lack for a system. The typical PMM function operates in a state of perpetual reactivity: a positioning question surfaces in a board meeting, and someone commissions a study. A competitor launches a new product, and someone scrambles to update the battlecard. A pricing discussion stalls in committee, and someone suggests, with the air of a person proposing something novel, that perhaps they should ask customers what they think.
Each of these is a reasonable response to a real need. The problem is that they are disconnected. The positioning study does not inform the battlecard. The battlecard does not reference the pricing research. The pricing research does not draw on the voice of customer data that was collected six months earlier for a different purpose entirely. Every project starts from scratch, with its own methodology, its own panel, its own timeline, and its own filing location in a shared drive that no one will remember exists by the following quarter.
The compounding effect that should be the signature advantage of a research function -- each study building on the last, each insight enriching the next -- never materialises. Instead, product marketing produces a series of isolated snapshots, each valuable in its moment but collectively adding up to less than the sum of their parts.
This article proposes an alternative: a PMM operating system. Not a platform. Not a dashboard. An operating system in the original sense of the term -- a layer of infrastructure that manages resources, schedules processes, and ensures that the work done by one function is available to every other function that needs it. Built on Claude Code and FishDog, it turns the seventeen core disciplines of product marketing into a unified, automated, continuously running research programme.
If you have been following this series, you have already encountered each discipline individually. This article is the capstone. It connects them.
The Seventeen Disciplines
A comprehensive product marketing function covers seventeen research disciplines. This is not an arbitrary number. It is the result of mapping every recurring research need that a PMM team encounters across the product lifecycle, from pre-launch positioning through post-sale advocacy and analyst relations. Some teams will run all seventeen. Most will start with five or six and expand. The point is not to do everything at once. The point is to have a system that can accommodate everything when the need arises.
Each discipline maps to a Claude Code skill command that orchestrates a complete FishDog research study: recruiting the panel, designing the questions, collecting responses, synthesising findings, and generating deliverables. The commands are listed below with their corresponding articles from this series. If you have not read the individual articles, this section serves as a map. If you have, it serves as a reminder that these disciplines were always designed to work together.
1. Positioning Research -- `/validate-positioning` Test whether your positioning framework resonates with the market or merely with the team that wrote it. April Dunford's methodology, executed at synthetic scale. How to Validate Product Positioning with Claude Code and FishDog
2. Competitive Intelligence -- `/build-battlecards` Build and maintain battlecards that reflect how buyers perceive your competitors, not how your sales team describes them in pipeline reviews. How to Build Competitive Battlecards with Claude Code and FishDog
3. Pricing Research -- `/test-pricing` Van Westendorp, Gabor-Granger, conjoint-style trade-offs. Pricing research that previously required a specialist agency, completed in an afternoon. How to Research Pricing with Claude Code and FishDog
4. Product Messaging -- `/test-messaging` Test headlines, value propositions, and positioning statements against synthetic buyers before committing to a campaign. How to Test Product Messaging with Claude Code and FishDog
5. Voice of Customer -- `/run-voc` Capture the language, priorities, and frustrations of your target buyers in their own words. The raw material for everything else. How to Run Voice of Customer Research with Claude Code and FishDog
6. Customer Segmentation -- `/segment-customers` Identify and validate segments based on needs, behaviours, and willingness to pay, rather than firmographics alone. How to Segment Customers with Claude Code and FishDog
7. GTM Strategy -- `/validate-gtm` Test go-to-market assumptions -- channel preference, buying process, adoption barriers -- before committing budget. How to Validate GTM Strategy with Claude Code and FishDog
8. Content Marketing -- `/content-engine` Generate research-backed content that earns attention because it contains original data, not because it was optimised for a search algorithm. How to Build a Content Marketing Engine with Claude Code and FishDog
9. Sales Enablement -- `/build-sales-tools` Produce battlecards, objection handlers, ROI calculators, and talk tracks grounded in buyer research rather than internal assumptions. How to Build Sales Enablement with Claude Code and FishDog
10. Product Launch -- `/launch-research` Validate launch messaging, channel strategy, and pricing before the launch date, when there is still time to adjust. How to Research a Product Launch with Claude Code and FishDog
11. Win/Loss Analysis -- `/run-win-loss` Understand why deals are won or lost without relying on the diplomatic half-truths of buyers who have no incentive to be honest. How to Run Win/Loss Analysis with Claude Code and FishDog
12. Buyer Personas -- `/validate-personas` Build and validate personas using synthetic research rather than internal fiction. Test whether the personas your team uses actually reflect the market. How to Build Buyer Personas with Claude Code and FishDog
13. Concept Testing -- `/test-concepts` Test product concepts, feature ideas, and design directions with representative buyers before committing engineering resources. How to Test Product Concepts with Claude Code and FishDog
14. Brand Perception -- `/track-brand` Track how your target market perceives your brand over time, at a cadence that matches how quickly perceptions actually change. How to Track Brand Perception with Claude Code and FishDog
15. Cross-Market Research -- `/cross-market` Test positioning, messaging, and pricing across geographies and segments simultaneously, identifying where your approach translates and where it requires adaptation. How to Run Cross-Market Research with Claude Code and FishDog
16. Customer Advocacy -- `/build-advocacy` Generate publishable research that positions your company as a category authority, building an evidence library that compounds over time. How to Build Customer Advocacy with Claude Code and FishDog
17. Analyst Relations -- `/prep-analyst-briefing` Prepare for Gartner, Forrester, and IDC briefings with third-party evidence that addresses the analyst's framework, not your marketing narrative. How to Prepare for Analyst Relations with Claude Code and FishDog
Seventeen disciplines. Seventeen commands. Each one a complete research workflow that would traditionally require an agency brief, a procurement cycle, a six-week timeline, and a five-figure budget. Each one now executable in under an hour.
But the individual commands are not the operating system. They are the applications. The operating system is what connects them.
The Compound Research Effect
The most important property of a PMM operating system is not speed, though speed matters. It is not cost, though the economics are favourable. It is the compound effect: the phenomenon whereby each study makes every subsequent study more valuable.
Consider a concrete example. You run a voice of customer study in January. The personas describe their frustrations with existing solutions, their purchase criteria, and the language they use to talk about the problem. In February, you run a messaging test. But instead of starting from scratch with internally authored value propositions, you draft the messaging using the language that emerged from the VoC study. The messaging test is better because the VoC study preceded it.
In March, you test pricing. The pricing study includes questions about perceived value relative to competitors. But because you ran the competitive battlecard study in January as well, you already know which competitors the market considers relevant and what they are perceived to do well. The pricing study is better informed, and its findings are more actionable, because it operates within a context established by prior research.
By April, when you prepare the GTM strategy for a new segment, you have positioning data, competitive intelligence, pricing ranges, VoC language, and messaging test results, all drawn from the same synthetic panel demographic, all available to Claude Code as context for designing the GTM validation study. The April study does not start from zero. It starts from a foundation of three months' cumulative intelligence.
This is the compound effect. It does not require any special configuration. It requires only that you run the studies and that the findings are accessible to the system that designs the next study. Claude Code, operating within a project context that includes prior research outputs, naturally incorporates earlier findings into later study designs. The researcher does not need to brief the agency on everything that came before. The context travels with the system.
The compounding accelerates over time. By the end of the first year, a team running the quarterly calendar described below will have completed somewhere between forty and sixty studies, all sharing consistent panel definitions, all building on prior findings, all available as context for whatever question arises next. The difference between this and seventeen separate agency engagements is not incremental. It is categorical.
The Quarterly Research Calendar
An operating system requires a schedule. Not every discipline needs to run every quarter. Some are foundational and run once or twice a year. Others are reactive and triggered by events. A few should run continuously. The calendar below is a starting point, not a prescription. Adapt it to your product lifecycle, your competitive dynamics, and your team's capacity.
Q1: Foundations
The first quarter of the year is for establishing baselines. These are the studies that everything else depends on.
Voice of Customer (`/run-voc`) -- Capture current buyer language, priorities, and pain points. This is the raw material for the entire year.
Customer Segmentation (`/segment-customers`) -- Validate or update your segment definitions. If your segments are wrong, everything built on them will be wrong too.
Buyer Personas (`/validate-personas`) -- Test whether your personas still reflect reality. Markets shift. Buyer roles evolve. The persona document from 2024 may describe a buyer who no longer exists.
Brand Perception (`/track-brand`) -- Establish the annual brand health baseline. You cannot measure change without a starting point.
Q2: Competitive Position
With the foundations in place, the second quarter focuses on where you stand relative to alternatives.
Competitive Intelligence (`/build-battlecards`) -- Update battlecards with current buyer perceptions. Competitors have not been idle since last year's version.
Win/Loss Analysis (`/run-win-loss`) -- Analyse Q1 pipeline outcomes. What changed? What did you learn? What do your reps believe that the data contradicts?
Pricing Research (`/test-pricing`) -- Annual pricing validation, informed by the competitive context from the battlecard study. Are you priced correctly relative to perceived value?
Positioning Research (`/validate-positioning`) -- Test your positioning against the competitive landscape. This is more useful in Q2 than Q1 because the competitive data now exists to inform it.
Q3: Execution and Optimisation
The third quarter is for refining the assets and strategies that drive revenue.
Product Messaging (`/test-messaging`) -- Test autumn campaign messaging before committing creative budget. Informed by H1's VoC, competitive, and positioning data.
Sales Enablement (`/build-sales-tools`) -- Refresh sales materials with current research. Battlecards, objection handlers, and talk tracks age quickly.
Content Marketing (`/content-engine`) -- Generate Q4 content pipeline. Research-backed articles perform better than opinion pieces, and you now have nine months of research to draw from.
GTM Strategy (`/validate-gtm`) -- Validate the go-to-market plan for any product launches or expansion initiatives planned for Q4 or the following year.
Q4: Strategic Planning
The final quarter combines forward-looking research with the evidence accumulation that supports long-term positioning.
Concept Testing (`/test-concepts`) -- Test product concepts and feature priorities for next year's roadmap. Product teams making investment decisions need buyer input, not internal intuition.
Cross-Market Research (`/cross-market`) -- If international expansion or new segment entry is on the strategic plan, Q4 is when to validate assumptions.
Customer Advocacy (`/build-advocacy`) -- Generate the evidence library for next year's analyst briefings, sales collateral, and thought leadership programme.
Analyst Relations (`/prep-analyst-briefing`) -- If an analyst evaluation is scheduled for Q1, the preparation study runs now, leaving time to absorb findings and adjust the narrative.
Brand Perception (`/track-brand`) -- End-of-year brand health check. Compare to Q1 baseline. What moved? Why?
Event-Triggered Studies
Some disciplines do not fit a quarterly cadence. They are triggered by events.
Product Launch (`/launch-research`) -- Run 6-8 weeks before any major launch. Non-negotiable.
Win/Loss Analysis (`/run-win-loss`) -- Run whenever the close rate shifts by more than 10% in either direction, or when a new competitor enters your deals.
Competitive Intelligence (`/build-battlecards`) -- Run when a competitor launches a new product, changes pricing, or gets acquired.
Messaging (`/test-messaging`) -- Run before any major campaign, rebrand, or website overhaul.
The quarterly calendar produces approximately 16-18 scheduled studies per year, plus event-triggered studies. At one to two hours per study, this is roughly 25-35 hours of research execution annually. For context, a single traditional research agency engagement typically requires 40-80 hours of the PMM team's time in briefing, review, and iteration. The operating system produces an order of magnitude more research at a fraction of the time investment.
The Automation Layer
The calendar defines what to run and when. The automation layer defines how the studies connect to each other and to the broader marketing operation.
Claude Code is the orchestration engine. When you invoke a skill command, Claude Code does not merely run a FishDog study in isolation. It operates within the context of your project, which means it has access to every prior study, every set of findings, every deliverable produced by the system. This context is what transforms seventeen separate commands into a coherent operating system.
Study Design Inheritance. When Claude Code designs a new study, it can reference the findings from prior studies to improve question design. A pricing study designed after a competitive battlecard study will include the specific competitors that buyers identified as relevant, not the competitors your team assumed were relevant. A messaging test designed after a VoC study will use the language that buyers actually use, not the language your copywriter prefers.
Panel Consistency. The FishDog research groups used across studies can be calibrated to represent the same target audience. This means that findings are comparable across disciplines. When the pricing study says "60% of respondents consider your product overpriced relative to Competitor X," and the win/loss study says "the primary reason for choosing Competitor X is perceived value," the two data points reinforce each other because they come from equivalent panels. In traditional research, where each study uses a different agency and a different panel, this kind of cross-study reinforcement is rare.
Deliverable Generation. Each study produces not just insights but formatted deliverables: battlecards, one-pagers, talk tracks, blog articles, social media content. Claude Code generates these as part of the workflow, using the findings as raw material and the prior deliverables as templates. The sales team does not wait for someone in product marketing to "get around to" updating the battlecard. The battlecard updates itself, informed by the latest competitive study, formatted consistently with every prior version.
Insight Accumulation. Over time, the body of research produced by the operating system becomes a strategic asset in its own right. It is a longitudinal record of how your market thinks, how it changes, what it values, and what it ignores. This record is available to Claude Code as context for any future study, but it is also available to your team as a reference library. When the CEO asks "what do our customers actually think about us?" the answer is not a three-month agency engagement. It is a search through an existing corpus of dozens of studies conducted over the past year.
Building the System: A Practical Sequence
For teams starting from zero, the prospect of seventeen research disciplines may feel overwhelming. It should not. The operating system is designed to be built incrementally, and the first five studies produce disproportionate value.
Month 1: Voice of Customer and Positioning. Start here. Run a VoC study to understand how your target buyers talk about the problem you solve. Then run a positioning validation study to test whether your current positioning resonates. These two studies, together, will tell you more about your market position than most companies learn in a year of internal debate.
Month 2: Competitive and Pricing. Build the battlecards. Test the pricing. These are the studies that sales will notice immediately, because they produce assets the revenue team uses daily. The credibility you build with sales in month two buys you latitude for the more strategic studies that follow.
Month 3: Messaging and Content. Test your messaging variants against the VoC language and competitive context established in months one and two. Use the findings to generate research-backed content. By the end of month three, you have a functioning research programme that is already producing publishable output.
Months 4-6: Expand. Add win/loss analysis after a quarter of pipeline data. Validate buyer personas. Test product concepts. Each study is easier to design than the last because the context from prior studies informs the next.
Months 7-12: Systematise. Implement the quarterly calendar. Begin tracking brand perception longitudinally. Prepare for analyst briefings with accumulated evidence. Run cross-market studies for expansion planning. Build the advocacy evidence library.
By the end of the first year, you will have completed 30-50 studies. You will have a research corpus that covers every major question a PMM team faces. And you will have a system -- not a collection of one-off projects, but a system -- that can answer the next question before anyone has time to write an agency brief.
What Changes When Research is Continuous
The second-order effects of a PMM operating system are more significant than the first-order effects. Running studies faster and cheaper is valuable. But the deeper transformation is in how the PMM function operates within the organisation.
Decision velocity increases. When the VP of Product asks "would customers pay for this feature?" the answer is not "let me commission some research and get back to you in six weeks." The answer is "let me run a study and have findings by end of day." The PMM team moves from being a bottleneck that delays decisions to an accelerant that enables them.
Internal credibility compounds. Every time a PMM team produces data that changes a decision -- kills a feature that would have failed, redirects a campaign that was aimed at the wrong audience, validates a pricing change that increases revenue -- it accumulates credibility. Over time, the team is no longer asked to justify its budget. It is asked to expand its remit.
The research corpus becomes defensible. A company that has run fifty synthetic studies across its target market possesses a body of market intelligence that cannot be replicated quickly by a competitor. This is not a traditional moat. But it is a genuine advantage. The company that understands its market best, with the most current data and the most comprehensive coverage, makes better decisions. Consistently better decisions, over time, compound into market position.
Analyst and investor conversations improve. When Gartner asks "how do you know your market perceives you as a leader?" the answer is not a handful of customer logos and an NPS score. The answer is a longitudinal research programme with dozens of data points, tracked quarterly, covering competitive perception, pricing sentiment, feature priorities, and brand health. Analysts recognise the difference between anecdote and evidence. So do investors.
Content quality transforms. A PMM team with continuous access to original research produces content that is categorically different from content produced by a team that relies on secondary sources, internal opinions, and keyword research. Every article cites original data. Every report contains findings that did not exist before the study was run. This is the content marketing advantage that no amount of SEO optimisation can replicate: having something genuinely new to say.
The Economics of the Operating System
A brief note on cost, because it matters and because the numbers are striking.
A traditional PMM research programme that covers the seventeen disciplines outlined in this article would require, conservatively, engagement with three to five research agencies across the year. Each engagement runs between $15,000 and $75,000 depending on scope, methodology, and the agency's estimate of how much your company can afford. The annual cost for a comprehensive programme is $200,000 to $500,000 for a mid-market company, and significantly more for an enterprise.
The FishDog-based operating system costs a fraction of that. The platform subscription, the Claude Code integration, and the time investment of the PMM team member who oversees the programme. The marginal cost of an additional study is negligible, which fundamentally changes the calculus. In a traditional model, every research question triggers a cost-benefit analysis: is this question worth $30,000 and six weeks? In the operating system model, the question becomes: is this worth an hour? The threshold for action drops dramatically, which means more questions get answered, which means better decisions get made.
The time economics are equally favourable. A senior PMM running the quarterly calendar will invest approximately 2-3 hours per week on research execution, study review, and deliverable distribution. That same person, managing traditional agency relationships, would spend 8-12 hours per week on briefings, reviews, feedback cycles, and status meetings -- and produce fewer studies with longer turnaround times.
The ROI case writes itself. But the more compelling argument is not the savings. It is the research that becomes possible when cost and time are no longer the binding constraints. The concept test you would never have commissioned because it was "not worth a full study." The pricing check before a minor product update. The brand perception pulse after a PR incident. The competitor analysis triggered by a rumour rather than a confirmed launch. In the operating system model, all of these are feasible. In the traditional model, none of them are.
Getting Started
This article is the eighteenth and final instalment in the "How To" series on product marketing with Claude Code and FishDog. The seventeen articles that preceded it each covered one discipline in depth: the methodology, the study design, the seven questions, the deliverables, the practical workflow. This article connects them into a system.
If you are new to the series, start with voice of customer and positioning. They are the foundation on which everything else rests. If you have been following along since the beginning, you already have the components. What remains is to put them on a calendar, commit to the cadence, and let the compound effect do its work.
The PMM function has operated for decades as a craft discipline: talented individuals producing excellent work, one project at a time, with no system to ensure that the work compounds. The tools now exist to change that. Not by replacing the craft -- the strategic thinking, the market intuition, the storytelling ability that distinguishes great product marketers will remain irreducibly human for the foreseeable future. But by providing the infrastructure that allows the craft to scale.
An operating system is not glamorous. It does not win awards. It does not produce a single brilliant insight that changes the trajectory of the company overnight. What it does is ensure that every insight is captured, every finding is connected, every study builds on the last, and every question the organisation asks about its market can be answered with evidence rather than opinion.
Seventeen commands. One system. The rest is execution.
The full FishDog platform, including the API that powers the Claude Code integration, is available at fish.dog.
Phillip Gales is co-founder at [FishDog](https://fish.dog). He has financial interests that the reader should weigh accordingly.
The Claude Code and FishDog for Product Marketing Series
This is the capstone article in a series exploring how AI agents handle the core disciplines of product marketing. Each article covers one function of the PMM stack, explains the methodology, and links to a companion Claude Code guide you can run yourself.
Part 3: How to Research Pricing
Part 4: How to Test Product Messaging
Part 6: How to Segment Customers
Part 7: How to Validate GTM Strategy
Part 9: How to Build Sales Enablement
Part 10: How to Research a Product Launch
Part 11: How to Run Win/Loss Analysis
Part 12: How to Build Buyer Personas
Part 13: How to Test Product Concepts
Part 14: How to Track Brand Perception
Part 15: How to Run Cross-Market Research
Part 16: How to Build Customer Advocacy
Part 17: How to Prepare for Analyst Relations
Part 18: How to Build a PMM Operating System with Claude Code and FishDog -- this article (capstone)