One — the transformation

I led the digital transformation of Routleys, a 100+ year-old omnichannel fashion retailer in Tasmania: new ERP and technical systems, custom middleware, ETL across inventory and CRM, AWS Redshift and Lambda — under a business that still had to trade every day, with brand relationships (Hugo Boss, GANT, Loake) to keep on side.

The technical migration was the part that got planned. The part that decided whether it worked was retraining, and the ordering mattered: train first and people learn a system that does not exist yet; migrate first and an untrained team runs unfamiliar infrastructure at the worst possible moment. We ran them in step.

Zero layoffs, 20% revenue growth. Every transformation I have watched fail, failed on adoption rather than on the stack.

Two — the writing

Freelance writing was the bread and butter first: 8fig, monday.com, Cardi.Health, thirdweb and others, 30,000+ publish-ready words a month, 400+ Quora answers of 700–1,200 words — 420,000–620,000 words across the period, all before AI tooling was usable.

Two books under other bylines: a 140,000-word book for a top eCommerce influencer, published by a major UK marketing and business publisher; and a 26,335-word book on investor pitching for a German investor-marketing consultancy — fourteen chapters and an appendix, one designed rather than written. Plus 40,000 words of data-science course material for a London startup founded by ex-Google, Deliveroo and Harvard PhDs.

Between those: Investor Relations & Marketing Lead at Magdalena Reith & LION Private Equity Partners in Munich — pitch decks that helped startups raise from seed to Series A+, and 70–100-page briefs trusted by government and M&A firms on international projects — plus investor letters for M&A and energy-project mezzanine financing, and freelance writing across the Steinbeis network in Germany.

Writing for somebody else is where the habit sharpened. Their voice, their opinions, their credibility on the line, and every claim has to survive a reader who knows the domain.

Three — selling and delivering

Queue is the long one: Head of Content from October 2022, then Head of Content & AI, now Head of Product Marketing & AI at the AI-powered LinkedIn-growth platform (backed by Sarona Ventures). I repositioned the product several times — working with 25 Madison on positioning, owning the website build, the copy and the messaging each cycle — grew demand generation to ~300% ROI, and built the LLM layer (compliance-tuned models, Weaviate + OpenAI vector search).

Before the consulting practice was mine, it was somebody else’s account book. At Queue I managed client services and delivery for monday.com, Guardio and CirclesUp while they were Queue’s clients: scoping, cadence, escalation, retention on accounts at real scale — and I sold too. Most deals were $15K–$30K that I closed myself; one broader-team engagement reached $100K/month, sold to a US veteran loan lender. Alongside Queue, DemandVolt — fractional product marketing and growth for AI SaaS.

Running Flow meant selling the work first. It was a lean 2–3 person team: I ran growth and sales alongside the founder, and we carried all sales and delivery between us. I personally sold and closed $15K–$30K engagements — FormDR, Scale Insights and DialMyCalls among them. Knowing what a deal feels like from both sides is why the delivery plans survive contact with the contract.

Four — the analytics and growth work

Eleven engagements, now named: FormDR, Scale Insights, DialMyCalls and Timelines AI (product and analytics consulting under Flow), HackingHR (head of growth, and interim CTO standing in for the founder), Net Atelier (product and growth lead), QForm, GetLeo, PitchBird and Gainify (the fintech since shut down) on the product side, and Spikerz, Tasq, CYE and Willow on the content and search side.

The pattern is consistent. Instrumentation before reporting. An audit before a strategy, because every strategy document contains claims nobody has checked. And when the evidence says nothing moved, report that, because a roadmap built on noise is worse than no roadmap.

Three examples of the audit finding the reporting was wrong: property coverage at 0% on a platform about to be reported on, 349,244 rows of billing data nobody had queried, and two systems disagreeing about the same customer base by 2,868 users.

And when the work ships, the impact gets proven the same way: funnel and cohort analysis with statistical validation — before/after compared properly, null results written up as null. That is the difference between a number and a finding, and it is why the claims survive a sceptical reader.

Five — building my own things

Jollywanderer first, in 2022: a no-code travel platform built in a week that filtered countries by COVID entry rules — used through the restrictions era. Then, later, the trade-site generator producing 278 pages across 18 trades and 10 cities from structured data, where a model fills a schema and a template builds the page. A research stack that runs search, AI-visibility and enrichment for under two dollars, with every provider call logged to the cent. And the content system in front of it: a knowledge base, a claims registry, a 197-rule lint layer and a set of publishing gates.

Building your own things tests which parts of your process are real. Every method I use on client work was proven or discarded on my own products first, mostly discarded.

What I am looking for

A senior analytics, growth or product-ops role where the state of the measurement is itself a problem to solve, rather than an assumed foundation. The engagements I do best start with somebody saying some version of “nobody trusts the dashboard.”

Open to remote, and to consulting engagements alongside a role. Based in Tbilisi, working across European and US time zones.

Contact

Working remotely since 2021 — 🇦🇺 🇦🇲 🇬🇪 🇧🇬 🇮🇱 🇹🇷 🇷🇴 🇲🇩 · based in Tbilisi, Georgia 🇬🇪