Available · 2 slots, Q4 2026
Product marketing functions, built 0→1 · EU and GCC
I'll build your product marketing operating system.
Automate the execution.
Then hand you the keys.
Six stages, installed in ninety days, with the research and refresh cycles running on AI, so your team maintains it without hiring a department.
The engagement ends. The function doesn't.
Built the product marketing function at
Sound familiar?
- People read your website but can't repeat back what you do
- You're about to launch and the plan is a date, not a motion
- Sales writes its own material because nothing official exists
- Every rep explains the product differently
- You sell to the user, their manager and the person who signs, and one message serves none of them
- You're entering a new market on instinct rather than a defined buyer
- Launches get negotiated from scratch every time
- You've hired a "strategic" product marketer and you're measuring them on how many assets they ship
Three or more? You don't have a marketing problem. You have a missing function, and hiring one more person into the gap won't create one. One or two, and it's usually a single launch or market that needs the work rather than the whole function.
The offer.
The function build
90 days to install the PMM Operating System
For companies with no product marketing function, or one that's become a content desk. I've done this three times from zero.
Days 1–30 · Understand the business
Product, roadmap, vision. Customer calls. Win/loss notes and support tickets. An audit of every customer touchpoint. The goal isn't answers. It's finding where the story breaks down.
01 Research · 02 Competition
Days 31–60 · Find the commercial gaps
Why customers actually buy. Where deals stall. Which objections repeat. By day 60 we know whether the real constraint is positioning, messaging, pricing, launches or enablement.
03 Segment & position
Days 61–90 · Build systems, not assets
Build systems, not assets. The six frameworks, installed and populated with your market rather than a template, with the automation wired up behind four of them. Then the handover: owners, refresh schedules, a trained team, and the jobs that keep it all current after I've gone.
04 Message · 05 Launch · 06 Measure
The ninety days and the six stages are the same thing.
AED SHOWN FOR REFERENCE. INVOICED IN EUR.
Payable across the three months. About twelve hours of your team's time in total. If you're pre-Series A and this is out of reach, GTM Foundations at €9,500 is the same method applied to one launch or one market.
Not sure you need this? Take the three-minute assessment →
GTM Foundations
3 weeks, €9,500 fixed
For a launch, a new market, or a first proper GTM motion. ICP and beachhead, competitive position, messaging per buyer, the channel and funnel model, and a launch plan with tiers. Stages 01 to 04 of the system, compressed.
About eight hours of your time.
The homepage read
Free, three a month
I read your homepage the way a stranger does and write back what I think you do, where the message breaks, and why it drifted. 48 hours.
No function → build. A launch or a new market → foundations. Just curious → read.
What you own at the end
The PMM Operating System.
Six stages. Six outputs. Four of them keep running after I've gone.
Consultants hand over documents. Documents rot. What I leave behind is a sequence that decides things, so the next launch doesn't restart the argument and the next hire inherits a function instead of a folder. Each stage produces one artifact, and four of the six keep themselves current.
01
Research
Market, customer, and the alternative you're really losing to.
Customer interviews, sales call review, win/loss analysis, review mining and the market map. Includes the null alternative: the spreadsheet, the status quo, doing nothing. Usually the competitor that wins most often, and the one nobody researches.
Output
The evidence base
● Automated02
Competition
Fifty questions, six competitors, every cell sourced.
A structured grid across market, product, positioning, marketing, sales and company. Every cell carries a URL and a date, or is explicitly marked unknown. No silent blanks.
Output
The Fifty
● AutomatedThe Fifty
| Category | You | Comp 1 | Comp 2 | Comp 3 | Comp 4 | Comp 5 |
|---|---|---|---|---|---|---|
| Market | ||||||
| Product | ||||||
| Positioning | ||||||
| Marketing | ||||||
| Sales | ||||||
| Company |
■ Sourced, with a URL and a date □ Explicitly marked unknown
Fifty questions. No silent blanks.
03
Segment & position
Who you're for, who you're not for, and what you're better than.
ICP definition, beachhead selection, personas split by user, decision-maker and blocker. Positioning argued against real alternatives rather than a template filled in.
Output
ICP and positioning statement
○ Judgment04
Message
What to say, to whom, with what proof behind it.
A message per persona per funnel stage, each line tied to a proof point. Claims that cannot be supported get marked as unproven rather than quietly shipped.
Output
The Messaging Matrix + Product Oracle
● AutomatedThe Messaging Matrix
One message per persona per funnel stage, each line tied to a proof point.
Product Oracle
Page one
What it is
Who it's for
How it works
Limitations
Page two
Packaging
Pricing
Personas
Anti-persona
The field most companies refuse to write.
05
Launch & enable
How big is this launch, and what does sales actually need?
A tiered launch process, the activity checklist each tier triggers, sales playbook, battlecards, and objection handling built from objections that actually get raised.
Output
Launch Tiers
◐ HalfLaunch Tiers
Every release gets a tier. The tier decides the work.
06
Measure & maintain
Prove it worked, and stop it going stale.
OKRs measuring research coverage, sales confidence and funnel movement rather than assets shipped. Plus the asset register that tracks when each thing expires and who owns it.
Output
Trust Metrics + The Half-Life Map
● AutomatedThe Half-Life Map
Most teams track who made a thing. This tracks when it stops being true.
Trust Metrics
| Objective | Measure | Target | Now |
|---|---|---|---|
| Sales can explain the product | Rep confidence survey | 8.0 / 10 | 8.4 |
| Research coverage is current | Segments researched this quarter | 4 of 4 | 3 of 4 |
| The story is consistent | Assets passing message audit | 95% | 91% |
| Launches move numbers | Releases with a measured outcome | 100% | 100% |
| Competitive answers are current | Cells sourced in the last 90 days | 100% | 100% |
Not assets shipped. Whether the rest of the business believes the function.
45 skills · research, refresh, audit, flag
Execution is the half that got cheap. Stages 01, 02, 04 and 06 largely run themselves. Stage 03 never will, and stage 05 only halfway. That is where the money is.
"Positioning doesn't break. It quietly stops being true."
THE DELIVERABLES
What you're holding on day 90.
Six documents, populated with your market rather than a template. Every one of them is something a person on your team opens and uses, not a strategy deck that gets archived.
Product Oracle
2 PAGES PER PRODUCT
Spec, metrics, limitations, packaging and personas, including the anti-persona. The single source of truth that stops four teams describing the product four ways.
The Half-Life Map
1 SHEET · EVERY ASSET YOU OWN
Every asset mapped to its owner, its destination and its refresh cadence. Positioning doesn't break. It quietly stops being true, so the useful question is when each asset goes stale and where it lives.
Launch Tiers
1 SHEET · 3 CHECKLISTS · 27 ACTIVITIES
Every release scored 1, 2 or 3, each tier triggering a fixed activity checklist, owners and critical path. Without it, every launch gets negotiated from scratch.
The Fifty
1 SHEET · 50 QUESTIONS × 6 COMPANIES
Fifty questions across market, product, positioning, marketing, sales and company, filled for you and up to five competitors. Every cell sourced with a URL and a date, or marked unknown. No silent blanks.
Trust Metrics
1 DASHBOARD · 5 OKRS
OKRs measuring research coverage, sales confidence and cross-functional trust, not assets shipped. Everyone else counts output. Output is the half that got cheap.
The Messaging Matrix
1 SHEET · PER PERSONA, PER STAGE
Per persona, per funnel stage, every line tied to a proof point, so claims can be checked rather than repeated.
ALL SIX ARE YOURS. NO LICENCE, NO PLATFORM, NO DEPENDENCY ON ME.
Six I built. Six I took apart.
Two kinds of work. Functions I built and ran, where I owned the outcome and the team. And briefs set by companies who wanted to see how I'd solve a real problem in their business before they hired anyone.
MarineTraffic → Kpler
Four product lines, four stories, then an acquisition that merged three more into one.
Payhaps
A checkout product with no cost of capital, sold into a category built around one.
WealthKarma
Retirement for GCC expats, in a market with no enrolment moment to hook onto.
Linagora
Twenty years of open-source credibility with no way to transmit it in a first meeting.
FoodVessel
Container-scale food procurement, digitised without breaking the trust it runs on.
Globcoin
A better stablecoin answer that took a paragraph to say.
Sumsub
KYC for European gambling. Three buyers including the regulator, each with its own message.
Kisi
Video intercom across real estate, tech and healthcare. A competitor table with a counterpoints column, so a rep can use it live.
PhantomBuster
Should a LinkedIn tool chase customers who don't sell on LinkedIn? A TAM argument, and an honest answer.
WalletConnect
One SDK, three audiences, and a competitor set that includes Firebase.
Proton
Growth strategy for a twelfth product. One feature gate doing acquisition, conversion and cross-sell at once.
ChargePoint
Turning commissioned research into a segment nobody was addressing.
How I work.
Systems, not assets. You keep the PMM Operating System. If your team can't run it after I leave, it isn't finished.
Judgment is the product. AI can research, draft and recommend. It can't own a trade-off.
Evidence over assertion. Every claim carries a source, or it's marked unproven.
Say who it's not for. Positioning that appeals to everyone appeals to nobody.
About.
I'm Harsha Kotthapalli. I've built product marketing functions from zero three times, across fintech, SaaS and marketplaces.
I started as an engineer. Four years at Nexen in Calgary running multi-million-dollar operational portfolios in a heavily regulated environment, where a decision without a written rationale wasn't a decision. That's where the systems habit comes from, and it's why I care more about what survives my departure than what I ship while I'm there.
An MBA at HEC Paris moved me into product marketing. Globcoin first, as founding product marketer for a blockchain financial product across five international markets. Then Acting CMO at Linagora in Paris, positioning open-source collaboration software as Europe's alternative to American and Chinese platforms.
Then a global maritime data business: MarineTraffic, and Kpler after the acquisition. I built the function from nothing, led four product marketers across four product lines, and took it through three acquisitions. Technical products, technical buyers: analysts, engineers, operations leads. The kind of product where the differentiation is real but hard to say out loud.
In 2024 I went through Antler in Dubai and co-founded WealthKarma, a retirement platform for GCC expats, with someone I'd met during the MBA. That year is why I take the "can your team actually run this" question seriously. On the other side of the table, a strategy you can't execute is worse than no strategy. It costs the same three weeks and leaves you with a document instead of a decision.
Since then: Payhaps, building the marketing function from zero for a UAE payments startup, and FoodVessel, taking a B2B marketplace into Saudi Arabia.
Based between Dubai and Barcelona, working across the EU and GCC.
Straight answers.
Which of the three should we start with?
No product marketing function and nobody owning the story: that's the Function Build. A function that exists, but a specific launch or new market to get right: that's GTM Foundations. Just want to know whether your homepage lands: take the Homepage Read, it's free. The three-minute assessment settles it faster than a call will.
We're pre-revenue or very early. Is it too soon?
Probably not, but it depends on one thing: whether you have anyone to talk to. I need access to five people who have bought, nearly bought, or walked away. Design partners count. If nobody has seen the product yet, it's too early for any of this, and I'll tell you that rather than sell you a plan built on guesses.
Where are you based, and how do you work?
Between Dubai and Barcelona, working across the EU and GCC. Everything runs remotely apart from workshops, which I'll do in person in either region if it's worth the trip. Invoiced in EUR.
What happens when you leave?
That's the part I design for. You get the frameworks, the automation and a team trained to run both. If the function collapses when I go, I've failed at the actual job.
Is this just AI doing the work?
AI does the execution: research, drafting, refreshing, auditing. I make the decisions, and the decisions are what you're paying for. The automation matters because it removes the excuse for frameworks going stale, which is what happens to every consultant's deliverables about four months after they leave.
Do we need this, or do we just need more leads?
If people convert once they understand you, you need leads. If they don't convert because they don't understand you, more leads just scales the confusion. The assessment usually settles which one you have.
Six of your twelve case studies say "briefed". Were those paid?
No. Each was a brief set by the company itself during a senior hiring process, using a real problem in their business. I didn't pick the questions, set the constraints, or choose the easy ones. Two ended in an offer. The reasoning in them is the same reasoning you'd be buying, which is why they're on the site rather than in a folder.
Can you also run our ads, SEO or content?
No. Go-to-market strategy, positioning, competitive intelligence, launch and enablement. I'll hand your content team a plan they can execute and recommend good people for the rest.
Will you talk to our customers?
Yes, up to five interviews if you can make the introductions. It's the single highest-value hour in any engagement, and their words tend to end up in your messaging more or less verbatim.
What if we're not a fit?
I'll say so before you pay, and tell you what I'd do instead.
Tell me what's not landing.
The launch that moved nothing, the competitor sales can't answer, the website nobody understands. Two sentences is enough. I'll reply within two days with an honest read on whether I can help.