Case studies

Visit promptcowboy.ai →

Overview

Prompt Cowboy went from 2,000 weekly users to 55,000 in two months with zero marketing spend. The founders had a product power users loved, customer love so strong that an M&A consultant said it got him "80% of the way there in 10 seconds instead of two to three hours." But they couldn't articulate the why behind the love, couldn't separate free from Pro, and couldn't tell the story to investors. I came in as the marketing strategist to diagnose what was actually working, build the positioning, and shape the narrative that turned customer love into a fundable thesis.

What I Did

• Diagnosed the real product insight from 1.5M+ prompts. Ran a Top 100 analysis of the highest-iterated prompts. Found that power users weren't just writing better prompts, they were engineering context. Framed the insight as "Context Engineering" and pushed it into the product strategy as the through-line for everything Pro
• Built the Free vs Pro tier positioning. Free for users who want to write better prompts. Pro for users who want to build intelligent AI systems powered by their own context. Clear value-based distinction, compelling upgrade reason, mapped to real user behaviour
• Ran the customer research that informed product direction. Interviewed power users across real estate, M&A, strategy consulting, agency, head of product roles. Surfaced the insight that 34% of repeat prompts wanted to become reusable assistants. That insight became the foundation for the Sidekicks feature direction
• Wrote the fundraise narrative. Built the 15-minute investor presentation script the founders used to pitch the round. Anchored on the 2K to 55K traction hook, threaded customer love throughout, mapped the strategic vision from prompt engineering to captured expertise to platform of expert frameworks
• Authored the Pro tier go-to-market memo. July 2025 strategy memo defining how Pro should be positioned, what features should anchor it, and how a parallel consulting motion ("Context Engineering Audit") could generate both qualified leads and product feedback
• Surfaced competitive moat language. Worked with the founders to articulate the iteration-pattern data as Prompt Cowboy's actual moat. Not the model, not the UI, the proprietary dataset of what makes prompts work across millions of iterations. Language the founders carried into investor conversations

Customer Evidence

• Real estate operator: "You are the GOAT, it changed my workflow with AI overnight"
• Strategy consultant: "Like having a personal AI coach guiding you to better results every time"
• M&A professional: "80% of the way there in 10 seconds rather than spending two to three hours"
• Risk management professional: Compliance reports went from a week to 2-3 hours. "I can use it almost as is"
• Agency owner: Featured Prompt Cowboy in his presentation to 500+ retail executives

Strategic Insight That Drove the Work

Power users spent 20 minutes per prompt, wrote detailed context, used frameworks. Most people wrote 42 words and wondered why AI sucked. The gap wasn't the technology, it was that most people don't know how to be good AI directors. Prompt Cowboy's job wasn't to make AI smarter, it was to make humans feel smart when using it. That insight reframed everything from positioning to onboarding to the Pro tier roadmap.

Results During the Engagement

Where the Strategy Compounded

The positioning carried. Today Prompt Cowboy has 350,000+ users, 2,000 paying customers, and roughly $260K ARR on a live Pro tier positioned around "for professionals who refuse to compromise on AI quality." That's the direct compounding of the Context Engineering thesis, separating users who want better prompts from users who want a system that thinks with them. The frame I built in 2025 is the frame the product is sold on today.

Delivered

Evidence

Overview

I led the launch of Promoted Listings Advanced, a new advertising product targeting larger eBay sellers and brands. I launched four new ad products from zero across APAC, US and EU, attracting 2,500+ active brands and helping grow eBay's ads business from $600M toward $1B (it has since doubled to ~$2B).

Process

• Research: Conducted 20+ in-depth seller interviews and analysed competitor offerings to understand the market landscape
• Positioning: Developed "Get more sales" messaging, resulting in an 87% increase in click-through rates
• Product Collaboration: Worked with the product team to prioritise features based on seller feedback
• Beta Testing: Recruited key brands like Dyson and Nespresso, incorporating their feedback into the final product
• Sales Enablement: Created an education platform, webinars, and sales materials for the GTM team
• Global Rollout: Supported teams worldwide in adapting the launch strategy to local markets across APAC, US, and EU

Results

Overview

I led a collaborative webinar with ChannelAdvisor to drive adoption of our Promoted Listings product through their new API integration. Despite having a robust API integration, adoption and revenue among sellers was unexpectedly low.

Key Challenges

• Low awareness of the API integration among 2,000+ sellers on ChannelAdvisor's platform
• Complexity and early stages of the Promoted Listings Advanced product
• Engaging a diverse audience across various industries
• Balancing technical details with practical insights

Development Process

• Collaborated with ChannelAdvisor to understand their platform and user base
• Worked with product managers to ensure accurate feature representation
• Spoke to sales team to ensure content addressed common seller objections
• Analysed API usage data with SQL queries and seller pain points to inform content
• Worked with global legal and Investor Relations to get script and content approved
• Set clear goals for API adoption, budget increases, and revenue growth

Results

Strengthened partnership leading to more joint initiatives including a newsletter.

Evidence

Listen to the Intelligent Growth podcast →

Overview

Most marketers use AI to write the same posts faster. I started Intelligent Growth to show a different version: AI as an operating system, not a writing tool. Every episode is a workflow demo, not a thinkpiece. The newsletter teaches the systems behind them. Everything runs on beehiiv as the single home for the audience.

What I Did

• Picked the wedge before picking the format. "AI as operating system" defined the show's reason for existing and the type of guest worth booking. Every episode shows a real workflow with the troubleshooting kept on camera
• Chose beehiiv as the single platform. Newsletter, podcast, and lead magnet forms all live there, so the audience meets every format from one record. Wired the beehiiv API into my publishing pipeline so automations run directly against it
• Designed the show format around live demos. Episode 1: Charlotte Norman, ex-Head of PMM at Canva, building Claude-powered brand voice systems. Episode 2: ElevenLabs Head of Growth, in production. The format sells itself to guests because they get to show real work, not deliver a pitch
• Built the AI content engine end to end. Five Claude Code skills do specific jobs: idea mining, angle extraction, drafting, a pre-publish voice gate, and publishing. Pipeline: a real work session becomes a newsletter, three LinkedIn posts, and a reel in roughly two hours
• Wrote my own newsletter HTML email builder. So I can publish plain text first and layer in HTML when needed, without leaving the system or fighting an email tool

Why beehiiv

One platform for newsletter, podcast, and lead-magnet forms, with API access so the Claude Code pipeline publishes straight into it. No stitched stack of separate tools.

Challenges and Solutions

• "Why another AI marketing newsletter?" The category is crowded with tool reviews. Anchored every episode on a workflow demo, not a thinkpiece. The show is named for what it teaches, not for the host
• One person, three formats. Built the workflow so the same raw material (a Claude Code session, a podcast moment, a real client launch) becomes the newsletter, LinkedIn posts, and YouTube clips through a shared content engine
• Booking high-signal guests with no audience. Led with the format, not the reach. Guests get to record a workflow they're proud of and that the audience hasn't seen before. Charlotte Norman said yes via cold outreach. ElevenLabs Head of Growth followed

Results

• Episode 1 with Charlotte Norman live on beehiiv
• Episode 2 with ElevenLabs Head of Growth in production
• AI content engine live: roughly two hour turnaround from raw work to published post across newsletter, LinkedIn, and a reel
• beehiiv platform fully wired: newsletter, podcast feed, lead magnet, and API automation all in one record

Delivered

Evidence

The course trailer. Built with Build Club, made with Manus, free to start.

Overview

Most marketers use AI to brainstorm. I wanted to show what happens when you give a fully autonomous agent a real marketing job and walk away. Manus runs computer use, persistent logins, and parallel agents in the background. So I built four workflows I actually use as a product marketer, recorded them as a course, and shipped it free on Build Club with the Manus team as co-marketing partner.

What I Did

• Picked the four use cases that matter most to a PMM. Competitive messaging, positioning, influencer discovery, pricing intelligence. Each is a workflow I use in real work, not a synthetic demo
• Built every workflow on real product data. Module 2 mines Otter and Fireflies reviews to extract differentiating language (the "feels like malware" framing came from G2 comments, not a workshop). Module 3 maps competitive territories beyond the homepage. Module 4 finds 15 creators already in your problem space ranked by relevance, not follower count. Module 5 reverse-engineers SaaS pricing structures across 10 companies
• Wrote the prompts so the domain expertise is baked into the brief. Every module includes the exact prompt and a breakdown of why each line matters. Marketers swap in their own competitors, ICPs, and platforms and the same workflow runs unchanged
• Got the Manus team on the collaboration. Co-branded launch on Build Club, in their distribution, with a certification track on top of the free course
• Filmed and shipped end to end myself. On-camera host, RØDE setup, editing in Descript, YouTube plus Build Club distribution. Six modules, roughly 15 minutes of total runtime
• Designed the certification track on Build Club so completing the course produces a portfolio piece the marketer can use back at work

Challenges and Solutions

• "How do you teach an agentic workflow without an hour-long screencast?" Kept each module under four minutes. Showed the prompt, showed the output, narrated the why. The viewer can copy the prompt and ship the workflow themselves in their next work session
• "How do you write a prompt a generalist marketer can adapt?" Built every prompt around the structure of the job rather than the specifics of the brand. The script narrates the spec, not the prompt verbatim, so the audience learns the underlying pattern
• "How do you make a free course feel like a real product, not a video dump?" Built the certification track. Marketers who complete the course submit their own Manus-driven workflow for review. Turns the course into productised accountability rather than passive content

Results

• Six modules published on YouTube and Build Club, roughly 15 minutes of total runtime
• Co-marketing with Manus including newsletter feature and Build Club platform placement
• Hosted on Build Club's certification platform with an open enrolment course path
• First Intelligent Growth free product shipped
• Became the credibility opener for AI-marketing speaking and podcast bookings

Delivered

Evidence

Open the full deck (PDF) →

Overview

A strategic GTM deck I built for Gamma: how to take it from 40,000 teams of 3 users to 1,000 teams of 10+ users, across four personas. The deck covers positioning, persona reframing, pricing research methodology, and the GTM motion. The Bridge Sponsor is the framing I built for it.

What I Did

• Ran original primary research before touching the personas. Interviewed Kevin Bogdanov (ex-Indeed and Salesforce sales enablement leader) on how enablement teams actually buy AI content tools. Validated the pattern through conversations with sales enablement and commercial leaders at Mastercard, UpGuard, Deel, and Rithum. Cross-referenced with Gartner buying-committee research and McKinsey adoption data
• Pushed past the four personas in the brief. The real enterprise buyer wasn't the VP Gamma had named, it was the enablement leader who feels the pain, translates product into content, runs pilots, and proves ROI before the VP signs the PO. Reframed the persona model around that insight
• Built a two-persona dual GTM motion. Marcus (sales rep) lands individual users through PLG, Kevin (Enablement Director) converts scattered usage into team adoption. Same product, two stories, no mixed messaging between marketing and sales
• Named the buying-committee dynamic. "Target Kevin to reach James, that's the enterprise motion." Gartner says the average B2B buying committee has 14 people. James (VP) signs the PO. Kevin (enablement) runs the pilot, trains the field, proves the value. Transformation is weaponising enablement
• Designed two parallel pricing-research tracks. Van Westendorp price sensitivity research with two competing hypotheses: is complexity killing conversion (simplify 6 plans to 4), or is credit confusion the blocker (transparency calculator showing ROI upfront). Test both, measure what works, no more guessing
• Anchored positioning on a category claim. "Gamma is the AI creation layer for sales enablement." Used April Dunford's framework on how market category shapes buyer perception. Seismic stores and distributes, Gong analyses, Gamma creates. Reduces surface area by integrating with what exists rather than competing for new platform real estate

The Strategic Reframe

The brief named the VP as the decision-maker. The research said otherwise. Enterprise tools get hired by the enablement leader and approved by the VP after the leader builds the case. Targeting James directly bypasses the person who feels the pain and runs the pilot. Targeting Kevin reaches both.

The Dual Motion

Marcus (PLG entry) and Kevin (enterprise conversion) are two heads of the same monster. Same product, different stories. Marcus message: "Turn client research into winning decks in 30 minutes." Kevin message: "Turn 4-week content cycles into 1-week launches." Marketing and sales stay aligned because the messaging is engineered for the layer they own.

Delivered

Evidence

Visit nextdocs.io →

My Role & Approach

Co-founder driving positioning, growth, and enterprise sales.

• Repositioning That Drove the Growth: Repositioned the product from "another AI writing tool" to one tool that replaces the ChatGPT, Google Docs, and DocuSign juggle. That single frame carried the 0 to 15,000+ user run in the first 4 months
• Market Discovery: Identified the enterprise pain point through direct consulting with Malaysian enterprises struggling with new ESG reporting mandates
• High-Friction Sales Environment: Navigated multi-stakeholder approval processes across finance teams (data ownership concerns), legal departments (AI liability questions), and sustainability officers (regulatory accuracy requirements)
• Partnership Strategy: Built consultancy partnership model to access corporate decision-makers
• Enterprise Contract Negotiation: Secured $30K development contract with Alliance Bank through 6-month sales cycle involving multiple stakeholder committees, legal reviews, and proof-of-concept demonstrations

Results

NextDocs has since grown to 50,000+ users and 1,000+ paying customers across 200 countries.

Evidence

Overview

Tasked with figuring out positioning and driving growth for Rephonic, a podcast analytics platform. Complex product with numerous features, diverse target audience, unclear messaging, limited budget, and a 5% trial-to-paid conversion rate.

Development Process

• Research & Positioning: Ran user interviews across the existing customer base to understand workflows, pain points, and value perception. Identified that technical, feature-focused messaging wasn't resonating
• Messaging Pivot: Simplified the core message to: "Find the right podcasts to be featured on." A/B tested messaging, achieving a 37% lift in landing page CTR
• Activation (PLG Tactics): Shifted from limited freemium to free trial. Implemented 7-email onboarding sequence with 45% open rates. Created in-app tooltips reducing time-to-first-value from 12 days to 18 hours
• Retention & Engagement: Launched weekly email digest with personalised podcast recommendations (increased WAU 40%). Built knowledge base to reduce support load

Results

Improved 3-month retention from 60% to 75%. Grew from 500 to 2,000 active users.

Overview

I developed a modular deck to help our GTM team communicate our advertising products to diverse sellers. This was crucial after our sales team was reduced from 18 to 6 members, including 4 new hires unfamiliar with our products.

Development Process

• Collaborated with product team on weekly syncs to track 15+ feature updates per quarter
• Worked with marketing to craft messaging, A/B testing key phrases with 200+ sellers before finalising positioning
• Shadowed 25+ sales calls to identify top 5 objection patterns and co-create response strategies
• Partnered with analytics to develop key stats and performance benchmarks
• Engaged customer success to identify and develop industry-specific case studies

Results

• Sparked more in-depth conversations with sellers at events like RetailFest
• Enabled sales team to quickly prepare for client meetings and presentations
• Supported new team members in understanding and communicating product value
• Served as the foundation for a global sales enablement program

(Slides unavailable due to eBay's privacy restrictions)

The eBay Ads Academy I ran (the AU homepage): a self-paced, free, on-demand learning platform for sellers, with full learning paths for Promoted Listings Standard and Advanced.

Sample marketing videos

Two marketing videos I produced for eBay Ads, made with the eBay content team and localised for the Australian market. (The Academy itself also had an education-focused how-to series.)

Overview

Promoted Listings Advanced was a powerful product, but a complicated one. If a seller did not know how to optimise their keywords or set their budgets, they got poor results and concluded the product was the problem. The post-launch reality was that the initial adoption spike drops off if people cannot get value. Education was the activation lever, so I launched the eBay Ads Academy: a self-serve platform that gave sellers a path to running campaigns like an expert.

What I Did

• Built the Promoted Listings Advanced learning path. Created the educational module on Promoted Listings Advanced, step by step, covering keyword optimisation, budgets and bidding, reading reports, and seller success stories by vertical
• Made it self-serve. Webinars, step-by-step seller guides, and sales enablement materials, so sellers and the sales team could learn the product without a rep walking them through it every time
• Ran the Australian Academy and produced the content. Localised the global Ads Academy content for the AU market and produced a three-part how-to video series on setting up a Promoted Listings Advanced campaign (manual, quick setup, and suggested), working with the e-learning team
• Designed it into the launch, not after it. The Academy existed to close the gap between buying the product and getting value from it, so adoption and activation were built in rather than bolted on
• Took load off sales. A self-serve path to expertise meant reps stopped repeating the same product education and could focus on selling

Challenges and Solutions

• Complexity was driving churn. Sellers who could not optimise blamed the product, not their setup. The Academy turned that complexity into a learnable skill, so trained sellers got results and stayed
• Education is usually an afterthought. I built it as part of the launch so activation was designed in, the same lesson I now apply to any complex product launch

Results

• Course completers saw around 30% better ad performance
• Around 26% lift in product adoption after launch
• Reduced churn, because trained sellers got results and stuck with the product
• Lower support and sales load, with a self-serve path replacing repeated one-to-one product education

Delivered

Evidence

Overview

I created and managed a monthly APAC Ads Feedback Newsletter to bridge the gap between APAC sellers and global product teams, ensuring valuable insights were effectively communicated and acted upon.

Impact

• Directly influenced the product roadmap, leading to 20+ new features including Quick Setup for automating campaign creation and a Campaign Details Page Dashboard
• Received comments from product and senior leadership including the VP Global Advertising at eBay
• Increased APAC's involvement in global product meetings and decision-making
• Improved cross-regional collaboration between product marketing and product development

Overview

At Ogilvy for KFC Australia, the mobile app had 20% retention. Users downloaded for discounts then deleted immediately. Zero marketing budget to fix it. I discovered frontline staff creating unofficial menu items internally (like "Beese Churger," "Nug-a-Lot"). The creative team proposed hiding these behind an 11-second logo hold tied to "11 herbs and spices."

What I Did

• Coordinated the campaign execution and wrote the discovery seeding strategy
• Wrote cryptic hints buried in app T&Cs: "Secret Menu available for those who know where to look"
• Created the breadcrumb trail across channels: edited 6-month-old YouTube comments with vague clues, coordinated hiding hints in email footers and split-second digital billboard frames
• Worked with product team on the unlock flow and interface copy
• Briefed photographers on menu item styling to look exclusive and shareworthy
• Designed the measurement approach tracking social mentions and app session patterns

Challenges and Solutions

• Staff weren't app advocates: Made them co-creators by featuring their menu hacks. Staff became natural promoters showing customers their creations
• Zero budget for paid discovery: Treasure hunt mechanics made early users feel like insiders who wanted to teach friends, creating organic tutorial content

Results

Users created 1,000+ organic tutorials teaching each other how to unlock it.