George Jurgens
Growth leader who turns customer friction, product behaviour and commercial signals into repeatable growth systems.
Zurich, Switzerland · Product-led growth in sales-led SaaS · AI-assisted execution
How I work
Selected work
A few of the growth systems built along the way, grouped by what they were for. Click into any of them for the full story.
Building the demand-capture engine, not just the campaigns
Rather than running one-off campaigns, built a repeatable operating model for product-led growth inside a sales-led organisation: capture signal, route signal, enable action, measure outcome. Tested the mechanics with in-product nudges, tooltips, banners and in-app messages surfacing new features and upsell moments at the right time to the right users, then built the automated routing that turned a raised hand into a ticket for the right Account Manager, CSM or Support owner, with no manual triage. Once it worked, trained GTM, Product, Engineering, CSMs and AMs across the organisation to run and extend it themselves, through a reusable Pendo guide template library, shared documentation, and a Growth RACI model that gave Product and GTM clear, non-overlapping ownership.
Turning an awkward conversation into revenue
Customers routinely exceeded their licence count, and the standard motion was an account manager chasing an overdue payment, an awkward conversation nobody enjoyed. Reframed it as a problem the customer could resolve themselves, in-product, at the moment of overage. The pilot made ~$50k ARR in 25 days, then grew 100% year over year, and the time from a flagged overage to new licences purchased dropped to often under 24 hours.
The collections workflow nobody wanted to own
Overdue invoices were a manual, low-priority chase. Tested the approach in Postman first, then built the live workflow in Zapier once the API pattern was proven, later evolving the same approach into a Google Chat app used for a different workflow entirely.
Retiring a legacy plan without losing the customers on it
A long tail of legacy free-plan tenants carried operational risk and technical debt, and no longer fitted the go-to-market strategy. Owned the sunset end to end: a 60-day notice period served in-product through Pendo banners to around 580 in-app users, and by automated HubSpot sequences to roughly 2,000 email accounts, phased against a hard sunset date and a final data purge 30 days later. Inactive accounts were soft-blocked ahead of the rest.
The part that mattered commercially was the retention path. Of the ~1,600 tenants, around 120 had recent activity and 17 were daily active, and the comms flow routed those to a commercial sponsor on the sales side instead of quietly switching them off. Sales, Customer Success, Support and the partner team were each briefed for the specific inbound they would get, down to distinguishing genuine concern from users who had created a free account by accident while trying to log in.
Testing what actually drives an upgrade
A free tier almost nobody upgraded from: 12 upgrade requests in 18 months, while most customers used it regularly. So the packaging question went to an experiment. 405 customers got full access to the paid tier, split two ways: a 30-day trial that expires, against one they could keep indefinitely.
The 30-day group was quicker to create a first form. The unlimited group created 146% more of them, and kept going. Urgency won the first action. Removing the deadline won the habit.
Two decisions followed: a licence price adjustment, modelled at a $345k expansion opportunity, and a reframe of forms usage as a retention lever, once the data showed customers on comms alone churned far more.
Owned the wind-down a year later too: countdown messaging in-product, and a usage-based cut of the tenant base so account managers could work the heaviest users first.
Closing a blind spot GDPR made worse
CRM contacts skewed heavily toward original buyers. For HR users specifically, only around 40% had a usable email address, and the rest weren't reachable through direct outreach because of GDPR constraints. Built a consent-based, self-serve persona-capture flow in Pendo, wired into Planhat and the CRM through Zapier and API automation, so admin-level users could identify their own department on first use instead of relying on acquired contact data.
Making upsell opportunity self-serve
Add-on licences and core licences drifted apart quietly. A customer could be badly under-licensed on an add-on relative to how many people they had on the platform, and nobody would notice without manual review. Built a Looker Studio dashboard sourced from Salesforce comparing total licences sold against add-on licences sold per tenant, so account managers and CSMs could spot the gap themselves rather than waiting to be told.
One review cycle flagged around $53k in incremental ARR opportunity across four add-ons. Refined with sales, CS leadership and RevOps over subsequent weeks to correct edge cases, such as global admin licences being miscounted as under-sold. This dashboard is one of the mechanisms behind the wider expansion opportunity identified across the business.
Rebuilding a fragile workaround with AI, in days
A fragile Zapier automation quietly handled every licence-overage exception, and it was getting harder to maintain. Rebuilt it using Codex, talking directly to the underlying APIs: CSMs and AMs can now pause or resume a customer's overage banner themselves through a single Google Chat command, with automatic 30-day re-activation if the account is still over threshold. The build itself took days; the real bottleneck was Google Cloud and Workspace access and governance approval, which took weeks longer than the build.
Turning a bug report into validated demand
Built and scaled an internal icon library tool inside Pendo that reached 28,000 total views and 3,728 first-time users, with strong repeat engagement. Rebuilt it from scratch in a single day using Codex: drag-and-drop upload, search and filters, colour customisation and instant download, embedded in-product through a Pendo guide, shipped solo with no engineers involved. When a customer reported a bug in it, the response wasn't just "fix it and move on": customers asked for custom icon library upload and brand-colour-restricted palettes instead, Customer Success flagged the demand within days of the feature being disabled, and a subset of customers chose to keep it re-enabled despite the known bug because the value outweighed the inconvenience.
Becoming Apple's own example of a well-built listing
As the lead relationship owner across Apple, Google, Amazon, Roku, Samsung and LG, secured App of the Day placements in key Tier 1 countries and launched a kids' coding app that Apple featured regularly across the entire App Store ecosystem globally, later used internally at Apple as an example of a well-constructed listing. Owned a marketing budget of up to £1.2m a year across paid and non-paid channels, and built the reporting infrastructure behind it.
Fixing a referral programme built to be gamed
An existing peer-to-peer referral programme had no real way to validate a referral, which left it open to abuse. Brought in third-party fraud-prevention tooling to properly track and validate referrals, then launched a wider Refer-a-Friend programme and a scaled influencer strategy on top of a channel worth trusting. The targeting itself had its own constraint: the business monetised by selling aggregated, anonymised spending data to hedge funds and businesses for market analysis, so acquisition needed statistically representative coverage across UK regions, not just the easiest audiences to reach.
Personal projects
Tapkit
An iPhone app that lets a complete novice, a cafe owner, a hairdresser, an Airbnb host, set up NFC stickers that open a review page, a menu, a phone number or WhatsApp in under a minute, using ready-made recipes.
A solo, evenings-and-weekends project since August 2026. George is the product owner, designer and decision-maker; AI coding agents write the code. The more interesting build is the system around directing them: a living source-of-truth document so new AI sessions share the same context, decisions logged as short PRDs, specialised AI reviewer agents, automated guardrails against known repeat mistakes, and a cost and time tool for deciding when handing something to AI is actually worth it.
How people describe working with me
Writing your own strengths section is a reliable way to produce fiction, so I didn’t write this one. Four years of manager reviews, peer kudos and anniversary posts went into an LLM with one instruction: find what people actually repeat, and drop anything that only got said once. These six came back.
Recognition
What managers and colleagues have said about the work. Everyone is by role rather than by name: they wrote these for a colleague, not for a website.
A co-founder and Chief Innovation Officer called him an innovation machine: versatile, fast to master new tools, and inclined to solve problems in ways that got other people doing the same.
A Head of Product Experience and Growth described the work as often behind the scenes but high impact, singling out the in-product licensing and invoice flows for their direct effect on revenue.
“As always with your little sidequests, loving the Outcome. Inspirational and fun having you on the team!”
“This added so much value to our customers! Thank YOU for making things happen instead of thinking it back and forth and hence too much.”
“Frankly, not surprised at all that you managed to bring significant value to both customers and the company with small but very powerful changes.”
“George didn't just execute my request; he consulted on user targeting and messaging, and then went above and beyond by embedding a functional form, a feature that exceeded my initial expectations. We went live yesterday and are already seeing conversions.”
“I really appreciate how quickly you looked into a problem we were having with the overage bot, and found a solution. You always make the time to support us on the A&E side.”