An idea, a green team, and no playbook to build it from
Senior leadership handed Roll, a 30-day rolling broadband contract, to a young and inexperienced project team to bring to life: a junior product manager, a marketing executive, a sales manager, and me on web and SEO. There was no precedent and no template. I came in to handle the website and grew into the de facto digital and delivery lead, owning the page and persona journeys, the creative, the full conversion tracking, and the reporting that connects an online enquiry to a confirmed sale.
Roll · hub plus three persona journeys · live on kcom.com
A concept from the top, handed to a team still learning the ropes
Roll started as an idea. Senior leadership saw the potential in a 30-day rolling broadband contract, 300Mbps and 900Mbps at £49.99 and £59.99 a month, no upfront connection fee, no long tie-in, and handed it to a project team to make real.
That team was young and, honestly, inexperienced: a junior product manager, a marketing executive, a sales manager, and me, whose remit was strictly web content and SEO. There was no playbook, no senior owner steering day to day, and no precedent inside the business for launching a proposition like this. Just a concept with promise and a group of people who had to work out how to deliver it.
My job on paper was the website. But a green team with no clear lead leaves a lot of gaps, and as they appeared I stepped into them. I grew from the web and SEO seat into the de facto digital and delivery lead, holding together the page, the persona journeys, all the creative, the conversion tracking and the reporting.
Throughout it I framed the work the way I frame all of it: not as tasks, but in terms of sales the proposition can be shown to drive.
A promising idea, and almost nothing built underneath it
A concept with potential is a long way from a product people can buy. When I looked honestly at where Roll actually was, the gaps weren't problems with an existing thing, they were everything that didn't exist yet. Three of them shaped what I built.
Nowhere for it to live
Roll had a name and a price and not much else. There was no page, no journey, no way for a customer to understand it or act on it. The proposition needed an actual home online before anything else could happen.
One idea, several very different buyers
Students, healthcare workers and people moving home all have different reasons to want a no-tie-in contract. The concept didn't say how to speak to any of them, and a single generic page would speak to none with real force.
The sale would happen where we couldn't see it
Enquiries would come in by phone or callback form and be closed by telesales offline. Left alone, GA4 would see the visit and the sales sheet would see the sale, with nothing joining the two, so the team could never prove the channel worked.
The job was never "make a nice page". For a team with no playbook, it was to turn a leadership idea into a measurable acquisition channel from scratch, and as the gaps in delivery showed up, to be the person who closed them.
One proposition, built in deliberate layers
With no existing product to fix, everything had to be built. Roll became a stack of layers that each solve a stage of the journey: giving it a home, speaking to the right buyer, capturing the enquiry, and finally being able to measure all of it. Each layer was scoped and shipped in turn.
The Roll page, designed and built
I designed and built the Roll product page end to end in the CMS: the hero, the proposition, the pricing, and the postcode checker that starts the journey. Iterated on against performance data rather than left as a one-off launch.
A hub plus three persona journeys
A main hub at /roll with three targeted subpages: Students, Healthcare and Movers. Same spine, copy and emphasis tuned per audience, for example static IP and eeros for students, flexibility and KCOM's emergency-services heritage for healthcare workers.
An enquiry journey that knows who it's talking to
A shared enquiry form embedded across all four pages, with a hidden persona field prefilled per page and carried through the thank-you redirect, so every enquiry is attributable to the journey it came from, with no manual tagging.
The campaign creative around it
All supporting creative: segment-based Meta ad carousels for Students, Healthcare and Movers, and a die-cut print leaflet for events and letting agencies, drawn from pre-approved copy so it cleared sign-off cleanly.
Conversion tracking, then reporting
GA4 event tracking via GTM for both enquiry actions, form submissions and call clicks, each carrying persona. On top of that, a performance report designed to stitch the online funnel to the offline telesales sale.
Closing the gap between the click and the sale
The hardest and most valuable part of Roll wasn't the page, it was the measurement. The conversion happens offline: a visitor submits a callback form or taps a phone number, then telesales closes the deal on a spreadsheet GA4 never sees. Most product reports fall apart in exactly that gap.
I built the join self-serve, with no developer dependency. A hidden persona field is prefilled on each page and carried through the form's thank-you redirect in the URL. In GTM I built two GA4 events, form_submit on the thank-you page and call_click on tel: link taps, each reading persona from the URL or, for calls, from a RegEx table on the page path.
It was verified properly: both events confirmed firing in GA4 DebugView with persona resolving correctly across every journey before anything went live.
The whole thing was built to be simple and reliable, not clever and fragile.
Getting it right meant catching the failure modes that quietly break this kind of setup: live paths carrying trailing slashes that defeat a naive lookup, a "clear hidden field values" setting that silently wipes persona, and a redirect that breaks the GTM Preview connection so the event has to be tested by loading the thank-you URL directly. Each of those was documented in a build sheet so the configuration could be rebuilt or handed over without me in the room.
Instrument first, scope tightly, verify, then build the report
Roll really demonstrates a way of working. The thread running through it is a refusal to report on data I can't trust: fix the measurement before building anything on top of it, and label honestly where the data has limits.
Work out what "deliver this" actually means
Leadership gave the team a concept, not a plan. The first real task was turning "make Roll happen" into concrete pieces of work, and spotting that the unglamorous one, measurement, was the piece that would make or break proving it worked.
A journey per buyer, not a single page
Hub plus three persona pages, each tuned to a distinct audience, with the postcode checker as the consistent entry point. The creative was designed to match, so the ad a student saw and the page they landed on told one story.
Tracking before reporting
I built the persona-attributed event tracking before building any report, because a report on incomplete tracking is worse than no report. Enquiries became attributable by persona the moment it went live.
Confirmed in DebugView, then published
Both events checked in GA4 DebugView with persona resolving correctly across all four journeys, then the GTM container published to live with a documented version history so the change is traceable.
Stitch the funnel, label the limits
A performance report designed to join the GA4 funnel to the offline telesales sale, the half neither system can see alone, with measurement breakpoints flagged honestly rather than smoothed over to look cleaner than the data is.
A measurable channel, and a model the business reuses
Roll moved from a leadership idea with nothing underneath it to a live, fully instrumented acquisition channel with persona-level attribution, delivered by a team that had never done anything like it. The clearest proof of the underlying model is the Black Friday landing page I ran on the same campaign-page approach, where the numbers are in.
Roll: from concept to channel
The shift isn't a single headline metric, it's an idea turned into a product the business can actually run, read and grow.
| Dimension | At the start | After |
|---|---|---|
| What existed | A name, a price, an idea | Live hub + 3 persona journeys |
| How customers reach it | Nothing in place | Paid, SEO, events & print |
| Audience targeting | Undefined | Tuned per buyer |
| Enquiry attribution | None | Persona-level, automatic |
| Funnel visibility | Online and offline disconnected | Joined visit → enquiry → sale |
| Delivery ownership | Green team, no clear lead | Single digital owner end to end |
The Black Friday page validated the wider campaign-landing-page model so clearly, including a backlink earned organically when an industry news site picked it up before launch, that it's now embedded as standard practice in the FY27 SEO strategy.
One campaign, every surface it had to live on
Across digital and physical: three Meta carousel slides in different formats, each carrying a different intent (proof point, package, call to action), and a die-cut print leaflet for events and letting-agency drops. All produced from pre-approved copy so it cleared brand sign-off cleanly, all built to the same campaign idea so the ad someone saw and the leaflet they picked up felt like one product.
Print leaflet · die-cut
Meta ads · carouselThe detail that made it owned, not just delivered
The difference between shipping a page and owning a proposition is in the parts that are easy to skip. These are the ones that made Roll genuinely mine.
Persona-aware enquiry capture
A hidden field prefilled per page and carried through the redirect, so every form submission and call click is filed to the right buyer with no manual tagging.
Two events, two intents
form_submit as the key conversion and call_click as a secondary intent signal, kept distinct because a callback request and a phone tap mean genuinely different things.
Failure modes designed out
Trailing-slash paths, silent field-clearing and a Preview-breaking redirect all caught and handled before launch, not discovered as gaps in the data later.
Documented to stand alone
A build sheet and a versioned GTM publish, so the setup can be rebuilt, audited or handed over without relying on me being there to explain it.
From a launch to a repeatable growth engine
Because the measurement now exists, Roll can be optimised on evidence rather than opinion, and the patterns it proved out feed the wider strategy.
Report live
Finish the funnel report joining GA4 enquiries to confirmed telesales sales, per persona.
Optimise
Use persona conversion data to sharpen the weakest journeys and the creative that feeds them.
Migrate cleanly
Carry Roll through the domain migration, bundling any naming or URL changes into that single move.
Templatise
Reuse the campaign-page model, already FY27 standard practice, for the next propositions and town pages.
The toolkit behind the work
A pragmatic stack chosen to keep the build self-serve, the tracking reliable, and the reporting honest, with no developer dependency for the measurement layer.
The lessons that outlast the project
Gaps on a team are where you grow
A green team with no playbook and no day-to-day lead leaves work nobody owns. Stepping into those gaps, rather than staying inside the web and SEO remit I was given, is what turned a task into a proposition I led. That judgement is the kind a job description never captures.
Fix the measurement before you report on it
A report built on broken tracking actively misleads. Instrumenting the funnel first, and verifying it in DebugView, meant every later number stood on something real.
Label the limits honestly
Joining an online funnel to an offline sale is where attribution gets fragile. Flagging the breakpoints openly builds more trust than numbers polished to look more precise than they are.
Owning something means documenting it
A build sheet and a versioned publish turn "the thing only Josh understands" into a system the business owns. Ownership and a single point of failure are not the same thing.