I build the go-to-market engine for businesses with a small, hard-to-find market.
Market mapping, contact enrichment, and outbound automation for companies whose buyers are not on anyone's list. Built end to end, running today, and handed off so your team can run it.
Engines I've built, end to end
Climate biotech · Athens
Mailing & print equipment · Long Island
Sales finds the deal. Engineering finds the market.
A go-to-market engineer builds the system that tells a sales team who to call, why, and when. Three parts.
Map the market
Find every company that could buy. Read each one's website and prove it from what they say and the machines they run. Keep it in one database that checks itself every night.
Find the people
Reach the owner, the plant manager, the buyer. Cheapest and most trusted source first. Verify every email before it is ever used. Log every attempt and what it cost.
Run the motion
Sequences, website-visit signals, reply triage, and a twice-weekly call sheet, all flowing into the CRM. The person sells. The system keeps them fed.
One pipeline, from raw list to booked call.
Five steps, in the order they run. Each one is a real thing I built and run today.
Source
Pull every candidate company from directories, trade press, job posts, government bids, and associations. For Mailtech that was 13+ sources and about 10,000 companies.
Read and classify
Crawl each website. An AI reader scores it against a written rulebook: what kind of business, what equipment they run, prospect or customer or competitor. Most companies run two or more businesses under one roof, so each door is mapped on its own.
One source of truth
Everything lives in one database. A nightly job snapshots it and runs invariants: rules like "a competitor can never be a prospect" that fail the export if broken. Sheets and lists are printouts of the database, never the other way around.
Enrichment waterfall
To find a person: own data first, then Apollo, then the open web and state registries, then pattern-guess and verify, then public records. Later steps only touch what is still empty. Every attempt is logged with its cost. An empty field beats an invented one.
Motion
Sequences go out. Replies are triaged by an AI step and become CRM tasks with the thread attached. Website visits and hiring signals land on a call sheet twice a week. Customers, competitors, and vendors are gated out of cold email automatically.
When your whole market is 2,000 companies, you cannot buy a list.
Generic outbound works when the market is huge and the buyers are on LinkedIn. Most real businesses are neither. Three things change when the market is small.
Your buyers are invisible to the big data tools
Apollo, ZoomInfo, and the rest all see the same LinkedIn-shaped world. The owner of a 12-person shop is not in it. The open web, state registries, and public records are. I built that layer, and it recovered identities for most of the companies the big tools returned nothing on.
Every account matters, so nothing can leak
You cannot burn a market you can count. About one in four "verified" emails from a major provider bounced in testing, so every address is checked again before use. Customers, competitors, and suppliers are blocked from cold outreach by a gate, not by memory.
A finite market compounds
With one database and every touch logged, each week's work makes next week's list better. The salesperson keeps selling. The system keeps learning who is real, who replied, and who visited the site.
Built for
- Founders selling into industrial, physical, or trade markets with a few hundred to a few thousand real buyers.
- Operators and PE ops partners who inherited a company where the customer list lives in one person's inbox.
- Hardtech and deeptech teams whose buyers are plant managers and engineers, not SaaS admins.
- Teams hiring a GTM engineer who has built this, not just run it.
Measured, not claimed.
Positive reply rate on cold outbound, against a 1 to 2% industry benchmark
Account market map built from scratch, 13+ data sources layered in Clay
Companies read, then confirmed prospects mapped, for a 60-year-old equipment dealer
“Food Ingredients Europe gated the attendee list — no export, no API. So I scraped it by hand and ran personalized outbound ahead of the show. It traced directly to 20+ booth meetings and a first distribution agreement in Singapore and Malaysia.”
Conference growth playEvery script has a dry run. Every write leaves a ledger row. Every schema change ships with its migration and its backup.
Domain knowledge meets systems thinking
- ThenCommercial real estate broker, LA
- NextPhysical commodity trader, Greece
- NowGo-to-market engineering, Axis GTM
I started in commercial real estate in Los Angeles, then traded physical commodities out of Greece, moving bulk fertilizer across three continents. That is where I learned what a real, physical market looks like from the inside.
Go-to-market found me at Solmeyea, a Greek climate biotech selling carbon-negative protein to Europe's largest food and feed producers. I owned the whole motion for ten months and built the pipeline from zero: Clay, Apollo, Smartlead, and the data automation behind it.
Today I build the same engine, deeper, for Mailtech, a Long Island mailing and print equipment dealer with a market you cannot buy a list for. The next step is turning that engine into a product other niche businesses can run.
Have a market that is hard to find? Let's map it.
Book a call. I will look at your market and tell you where the buyers are hiding and what it would take to reach them.