My CV, with the detail behind it: the roles, the brands I’ve delivered for, how I run delivery, and the parts I don’t claim. Available for fractional and part-time contract delivery leadership.
Kingsteignton, Devon · South West · UK-wide by video · 07955 346 403 · [email protected]
I founded Vision2 in 1999 and have worked in digital delivery for over 25 years, running it alongside agency roles. I started as a front-end developer, ran servers and data feeds, and moved into project and delivery management from there, which is why I can sense-check an estimate rather than just pass it on.
In agency roles I owned whole project and retainer portfolios: budgets in six and seven figures, resourcing, delivery standards and client relationships from product managers up to CFO level. Through Vision2 the work has been platform strategy, integration design and a long run of migrations and replatforms, including clean rebuilds out of codebases that had degraded past saving.
I’ve built and managed the teams that delivered it, across remote and hybrid set-ups, in the UK and internationally.
2025 – now
Scoping, builds and practical automation for growing businesses, and fractional delivery leadership. Using AI tooling on real delivery problems, including the document system described below.
2023 – 2024
Employed at the agency. Owned its full project and retainer portfolio: concurrent Hyvä builds, feature programmes, bug pipelines and retainers across enterprise and SME accounts, with accountability for budgets, risk and reporting. Primary delivery contact for an enterprise watch group running several concurrent Adobe Commerce sites. Rebuilt the agency’s operations in Productive.io and set acceptance-criteria-first standards. Led the development team in reducing security patching delays from months, through weeks and days, to within 24 hours of release, with patches mapped and scheduled and a clean staging environment for testing.
2022 – 2023
Employed at the agency as its delivery lead across a WordPress, WooCommerce and Magento portfolio: relaunches, Magento to WooCommerce migrations and bespoke portals, owning budgets, retainers and resourcing. Rescued a stalled client account, cleared a months-old backlog, rebuilt the agency’s Avaza set-up and introduced automated regression testing.
2009 – 2016
Employed as Project Manager, and in practice the project and delivery manager: projects, design and client delivery for consumer brands and membership organisations.
2008 – 2009
Project Manager on paper, working end to end in practice: design, development, deployments, server management and cPanel hosting, and the stakeholders on each project.
2005 – 2008
Project Manager on paper, working end to end in practice: design, development, deployments, server management and cPanel hosting, and the stakeholders on each project.
2004 – 2005
Officially Project Manager; in practice I owned every project, along with the development and design of the membership platform and the affiliate marketing behind it. Helped grow the main site from around 10,000 users to over 300,000.
2002 – 2004
Design and front-end development for client sites, the start of moving into project delivery.
1999 – 2022
A delivery consultancy and agency partner, run alongside employed roles: platform strategy, integration architecture, Magento eCommerce builds, migrations, hosting on its own servers, and the teams that delivered them.
Vision2 clients
Hosting and support for over ten years
From my years as a project and delivery manager employed at digital agencies, 2022 to 2024. These were the agencies’ clients, not Vision2’s: I led the delivery work, and the agency held the client relationship.
All trade marks are the property of their respective owners. Showing them here doesn’t imply any endorsement or a current relationship.
Testimonials given to the agencies I worked for, about projects I delivered while employed there. They describe the agency team; I led the delivery.
“The project was delivered on time, and the progress was transparent, with excellent communication along the way.”
eCommerce Manager, Peter Tyson
“The attention to detail to achieve the correct solution to each task, rather than just taking the most straightforward option, was refreshing, always with an eye on long-term use and success.”
eCommerce Manager, Peter Tyson
“They kept me informed throughout the project and I’d happily recommend them to anyone.”
Key Organics
All of this was delivered while I was employed at digital agencies, on behalf of their clients. I ran the delivery; the agency held the client relationship.
Citizen, Bulova and Frederique Constant: primary delivery contact across several concurrent Adobe Commerce sites, covering B2C storefronts, B2B, multiple regions and internal staff-facing instances. Engaged from product managers through to CFO level.
Peter Tyson’s eCommerce relaunch onto Hyvä. Rodda’s WooCommerce relaunch, delivered on time and within budget. Vospers Parts’ move from Magento 1 to WooCommerce, integrating DVLA, CRM and national parts database services, with a vehicle lookup re-architected so each selection populates the next without a page reload.
Kawasaki, Noble Isle, Turnstyle Designs, Bed Kingdom, Hideaway Beds, Bakare Beds and Freckl: concurrent projects and retainers across luxury retail, automotive, furniture, medical equipment and sustainable packaging.
The Open Group (standards publishing), Key Organics (pharmaceutical research), China Fleet Country Club (hospitality), Vision Zero South West (public sector) and eManuals (technical documentation), plus independent education, property and lettings.
A custom product configurator for the marine sector. An office and coworking booking portal with digital access-code management. A secure client document portal. A mapping integration for land measurement, feeding dynamic product recommendations. All scoped, delivered and handed over.
Active budgets, resource allocation and delivery across concurrent programmes, retainers and bug pipelines, with enterprise and SME accounts running side by side. Financial tracking and forecasting, risk, and stakeholder reporting from discovery through to production. Cross-functional teams of developers, QA and designers, and mentoring the project managers and coordinators alongside them.
There are results attached to some of this work, and I can talk you through them. They are not on the page because none of them has a documented baseline: I can tell you what a site did after a relaunch, not reliably what it did before.
An after-number without a before is a marketing number, and I’d rather show you how I measure than quote you something I can’t stand behind. On current engagements the baseline gets captured on day one, which is the whole point.
Before delivery management took over, I built an affiliate marketplace network: a master site with several million products and around two hundred specialist subsites beneath it, on Magento, fed from Affiliate Window, Tradedoubler and Amazon.
It worked. I got the subsites ranking in Google Shopping alongside national retailers and took a serious share of their traffic, so I built more of them. As revenue came in it went back into dedicated servers, Apache tuning and PHP upgrades, because the crawler load alone was hammering them.
Then Google changed the algorithm and it fell over overnight.
I include it because it’s the only thing here I built, scaled, monetised and lost entirely on my own, and because the ending taught me most about how much of a business can rest on a dependency you don’t control.
In 2025–26 I built a case-management system to support the solicitors and barristers preparing a multi-year construction dispute. Hundreds of PDFs running to 30, 50 and more pages each, professional reports, thousands of emails in threads that ran for pages, audio recordings and other records, all on different and unconnected reference systems.
The objective was a single view that could tell you, for any one of hundreds of issues, what it was, when it was first reported, by whom, through what channel, to which party, what happened next, and where it stood, robust enough to hold up in a formal legal process. It now brings around 1,800 records from five independent sources into one view, each cited back to its source, with duplicates merged and contradictions surfaced, and the legal team work from it as the single source of truth.
I prototyped it in general-purpose AI tools, which proved the approach and then hit its limits: inconsistent output, no persistent context between sessions, and more manual verification than the time saved was worth. I rebuilt it properly, breaking the work into discrete orchestrated tasks with sub-agents checking against defined criteria rather than processing everything in one pass.
Structuring model work turned out to be the same discipline as structuring sprint work: break it down, define what “done” means for each piece, verify against it, and break it down further where it still fails. See a working demo on a fictional case
I’ve repeatedly inherited project and operations platforms that had been bought, badly implemented and then quietly worked around, Productive.io and Avaza among them.
Rebuilding means starting with how the work actually moves, then setting up project structures, budgeting, time recording and reporting to match, and training the people expected to use it: project managers, coordinators, developers and leadership, each needing a different conversation.
The thing that made it useful wasn’t the reporting. It was that resourcing decisions could finally be made on real numbers: who was available, what they cost, and what level of experience the work genuinely needed.
I’ve established acceptance-criteria-first delivery standards more than once: a structured brief with a short video capture before anything entered development, and criteria written before the work rather than argued after it.
The mechanism mattered more than the policy. It started in the bug pipeline, where the rework was most visible and least defensible. I wrote the first ones myself so it couldn’t be read as work pushed downward. And the video capture is what made it survive: thirty seconds of screen recording where a written description would have taken ten minutes, so complying became cheaper than not.
Done properly, a developer builds it right first time, which is less development time, lower cost, and more of the budget available for the next thing.
Everything I run is built so the client can see where things are without having to ask. If you need a meeting to find out the status, the system has failed.
Scope it, agree acceptance criteria, front-load the thinking. Agile after that is fine, as long as the consequences of each change are made explicit at the time.
A specification tells you what should happen. A specification plus a design gives you something to hold the build against, and QA gets considerably shorter.
Before and after every deployment. Not because something usually goes wrong, but because when it does, the difference between calm and chaotic is whether you already know what good looks like.
A blockage is a solution awaiting a named owner. A blocked column is a place where things go to be nobody’s problem.
The specification that’s sixty-nine pages long and contains no acceptance criteria. The platform bought to fix a visibility problem that was never a tooling problem. The integration that looks like a fortnight and is actually a quarter because nobody has looked at the state of the data.
I’ve written them, priced them, defended them and delivered against them. So when one lands on your desk I can tell you which assumptions are load-bearing and what “we’ll scope that in phase two” is protecting.
I started as a front-end developer, ran servers, handled data and migrations, and moved into delivery from there. A delivery manager without that background has to take the answer on trust.
Portfolio budgets in six and seven figures, resourcing decisions that moved margin, and the judgement about when a project needs a senior specialist and when it doesn’t.
A multi-region data alignment where my own team’s instinct was a custom build. I argued against us: the custom route would have burned the timeline protecting a much larger rebuild, and the rebuild was where the commercial value sat. We took the cheaper, uglier bounded option deliberately, fixed its performance in order of leverage with each pass re-authorised against an explicit cost gate, and it landed on time. What I’d change: set the cost ceiling before starting, not in flight.
A checkout redesign, approved on a friction hypothesis everyone agreed with, built and shipped. Behavioural data showed users navigating around the new flow to reach checkout their own way, and converting. The friction was the primary path itself. So we reinforced the route users had chosen and removed what we’d built.
A subscription product where churn and retention refused to make sense. A chunk of the user base wasn’t real: trials created on zero-balance prepaid cards that passed validation. The fix was a refundable one-pound charge, which forces a genuine authorisation and costs a real customer nothing. Until the gamed accounts are gone, neither churn nor conversion is telling the truth.
Also worked in: Adobe Commerce (including Enterprise and B2B), Magento 1 and 2, Google Analytics, Tag Manager and Search Console, Advanced Custom Fields, Avaza, ManageWP, cPanel, MySQL. All trade marks are the property of their respective owners.
Development first, then websites, infrastructure, agency operations and delivery leadership, often several at once, which is where the useful part comes from.
Most of what I do now is translation: turning what a business needs into something a developer can build against, and turning what the technology is doing into something a board can make a commercial decision on. It’s the reason I can be in a room with leadership, an internal team and a supplier and give all three a version of the same answer.
One long client relationship worth mentioning, since consultancies are usually judged on the first month: I’ve hosted and supported MadMax Adventures for over ten years. It isn’t the biggest thing on this page. It’s the one that says most about how I work.
Based in Devon: in person across the South West, and by video call anywhere else.
Tell me where things stand, what needs to change and any date you are working towards.
You might have a project to scope, systems that aren’t joining up, or a task that keeps taking too much time. You don’t need a finished brief to start a conversation.
A few lines about the situation, the people involved and what a useful result would look like are enough. I’ll come back to you personally with a practical next step.