Customers stalled at the Microsoft tenant step in testing.
A single, globally-deployable marketplace that lets SME & SoHo businesses — small, medium and home-office companies — across Ireland, Portugal, Spain and South Africa discover, configure and buy Microsoft cloud services — with a Microsoft tenant checkout cut from 12+ screens to 4. It grew out of V-Hub, our free SMB knowledge base born in COVID, and after two executive-committee reviews of the work the programme drew two further rounds of investment — €170M, then €350M. I led design across its 2.5 years, from the first idea on.
The Vodafone Business Marketplace storefront — where SME & SoHo customers discover and buy Microsoft cloud.
Vodafone Business sells cloud services — chiefly Microsoft 365 and Azure — primarily to its SME and SoHo customers (small, medium and home-office businesses), while the same platform also serves enterprise customers, across multiple countries. Each market ran its own fragmented, hand-built store. My mandate was to design one marketplace that could scale across all of them without fragmenting the experience or the codebase.
The programme was born in COVID. It started as V-Hub — a free knowledge-base platform teaching small and medium businesses how to digitalise. Lockdown pushed a wave of SMBs online, V-Hub launched successfully, and a simple question followed: if we're giving away the knowledge for free, why not also offer the tools to act on it? That insight became the Cloud Marketplace — and I was on it from the very first idea.
Over two and a half years I led product design end to end: discovery and journey mapping across four markets, the information architecture of a single master-site, the design system that let one build serve every country, and the checkout redesign that collapsed a 12-screen Microsoft tenant wall into four guided steps.
The hardest problem throughout was delivering Microsoft's products cleanly into our own system — which is why I ran a constant cadence of meetings with the product owner, the local markets (OpCos) and Microsoft, plus a daily stand-up with front-end and back-end engineers, the scrum master and QA. The platform has been live and continuously improved ever since.
Lockdown forces small and medium businesses to digitalise fast — and to look for help doing it.
We launch V-Hub to teach SMBs how to digitalise. It lands well and brings a wave of new customers.
If the knowledge is free, the natural next step is the software, packages and bundles to act on it.
A B2B store born from a proven customer need — validated by V-Hub before a line was built.
The programme had to solve two problems at once: a purchase journey so complex it lost customers at the payment moment, and a market-by-market approach that made every new country a rebuild.
Microsoft's security requirements produced a 12+ screen tenant-onboarding flow — sign-in, 2FA, tenant selection, domain checks, consent. Drop-off peaked at the single most critical moment: the point of purchase.
Moving a customer's Microsoft 365 tenant between plans meant fragile, manual coordination between Vodafone, Microsoft and the customer — slow, error-prone, and a magnet for support calls.
Every country had different branding, language and regulatory needs, and no shared design system. Consistency was impossible and each new market launch started close to zero.
Design a scalable, globally-deployable B2B cloud marketplace where customers can discover, configure and purchase Microsoft cloud services across markets — with a checkout simple enough to trust and a single system consistent enough to deploy in weeks, not quarters.
I opened with structured discovery across all four markets — stakeholder workshops, customer interviews and competitive benchmarking — so the redesign was aimed at measured friction, not assumptions.
“The buying stops the moment Microsoft's login appears.”
The tenant wall was the single biggest leak. Everything else was secondary to guiding customers through — or around — those 12 screens. This became the headline problem.
“I'm not sure which plan I actually need.”
Small-business owners — often not IT specialists — hesitated on configuration. This drove guided-selling flows that ask a few plain questions and recommend the right plan, instead of dropping people into a spec sheet.
“Is this Vodafone or Microsoft I'm buying from?”
Co-branding was ambiguous. We set a firm hierarchy — Vodafone always primary, Microsoft product branding always secondary — so trust and accountability were never in doubt.
“Every country's store feels like a different company.”
Fragmentation hurt both customers and delivery. This justified a single master-site with a shared token library and a localisation framework layered on top.
The marketplace was designed SME- and SoHo-first — from a one-person home office to a mid-sized company — the least technical, highest-volume audience; the same platform also serves enterprise buyers. Even at the small end, buying is rarely one person: three overlapping roles judged the same purchase against different concerns — cost, control and complexity.
The architectural spine of the whole programme: a single master-site design and codebase that resolves to each market's branding, language and regulatory rules through configuration — not a fork. Design the store once; deploy it everywhere.
The marketplace turns a fragmented, market-by-market toolset into one coherent product — easy to buy from, consistent to trust, and fast to roll out to the next country.
“From a 12-screen tenant wall to four guided steps — secure B2B commerce that finally feels like shopping.”
A single design system adapts to each market's branding, language and regulation — so new markets launch fast, without design drift or a rebuild.
The Microsoft tenant journey drops from 12+ steps to 4, and guided selling helps customers choose the right plan with confidence — cutting abandonment and support calls.
Vodafone always primary, Microsoft always secondary — a consistent visual hierarchy across product cards, banners and promotions that keeps accountability unambiguous.
Built on Vodafone's global design system, I created marketplace-specific components aligned to Microsoft 365 product standards, and a shared token library across all four markets. Two rules made it scale: every value is a semantic token, and every co-branded surface follows one hierarchy.
Colour is never hard-coded. Semantic tokens — surface, accent, on-accent, state — resolve per market and per theme, which is exactly what lets one build carry four brandings.
Vodafone — voice, container, CTA colour and accountability always sit with Vodafone.
Microsoft — product identity (M365, Azure, Copilot) is present but visually subordinate.
Applied identically across product cards, banners and promotional imagery — so a customer always knows they're buying from Vodafone, powered by Microsoft.
A rigorous semantic token layer is what lets a single master-site resolve to four brandings. Re-skinning a market is a data change, not a redesign.
Marketplace components were aligned to Microsoft 365 product standards, so co-branded plan cards read as authentic Microsoft products inside a Vodafone store.
Language, currency and regulatory rules resolve through configuration — the same reason a new market can be stood up in around six weeks.
The storefront is deliberately shaped like consumer commerce — familiar patterns doing serious business work underneath. These are real screens from the live design. Click any to view the full page.
Search, featured business applications, guided “how it works” and clear co-branded entry points — the discovery layer that gets a buyer moving.
A filterable catalogue of business apps and plans, priced and comparable at a glance.
A product proposition page: what it is, which plan fits, and a guided path into configuration.
Where the co-branding rules earn their keep — Microsoft products presented authentically, always inside the Vodafone frame.
The defining piece of product work. Microsoft's tenant onboarding is genuinely complex and non-negotiable — so instead of removing steps, I re-choreographed them: absorbing the mandatory security into four guided stages, each with one clear job.
Microsoft's tenant security can't be shortened away — 2FA, tenant selection and consent are required. But 12+ raw screens at the payment moment read as risk, and customers walked.
Cluster the mandatory steps into four guided stages, explain what each one is for, carry state forward, and keep Vodafone's frame around Microsoft's auth so trust never breaks.
Continue with the organisation's Microsoft account — one clear action, framed by Vodafone.
2FA and the required tenant permissions, explained in plain language before they happen.
Select the tenant and confirm the configured plan, with costs shown in full — no surprises.
Review and buy on a single Vodafone invoice; the subscription provisions automatically.
The moment customers used to drop off — now a calm, single-decision screen that hands off to Microsoft without ever feeling like they've left Vodafone.
Cognitive load in checkout reduced by roughly 40% across three rounds of usability testing — the tenant wall stopped being the place the sale died.
Two decisions shaped everything else. Both were about holding a line — one on architecture, one on brand — under real pressure to compromise.
Each country builds its own store. Maximum local freedom — but fragmentation, drift and a rebuild every launch.
One design and codebase, localised by configuration. Requires strict token discipline and close work with local teams.
Chosen: the master-site. It demanded rigorous tokenisation and real collaboration with local market teams — and in return, a new market launches in ~6 weeks instead of a fresh build.
Equal weight reads as a joint venture; Microsoft-led reads as “buying from Microsoft, billed by Vodafone.”
Vodafone primary in container, voice and CTA; Microsoft secondary as authentic product identity.
Chosen: one hierarchy, applied everywhere — product cards, banners and promos — so accountability and trust are never in question.
The marketplace wasn't shipped once — it was piloted in Ireland, validated with real customers, then rolled out market by market, improving continuously across 2.5 years of live operation.
Customers stalled at the Microsoft tenant step in testing.
Re-choreographed 12+ screens into 4 explained stages.
~40% less checkout cognitive load; the sale stopped dying at auth.
Buyers hesitated choosing between near-identical plans.
Added guided-selling flows validated across 3 segments.
Confidence to purchase without needing to decode a spec sheet.
A market-by-market rollout risked drift and slow launches.
Established a reusable rollout playbook on the master-site.
New markets launch in ~6 weeks, consistent from day one.
Three rounds of usability testing with Vodafone Business customers, plus continuous performance monitoring on the live platform — a feedback loop that ran for the full project lifecycle, not a one-off validation.
Regular sessions with the product owner, the local markets (OpCos) and Microsoft — aligning commercial goals, market-specific rules and, above all, the Microsoft delivery model that was the programme's hardest problem.
A daily stand-up with front-end and back-end engineers, the scrum master and QA — so every design decision was pressure-tested against build reality the same day, not weeks later at handoff.
Year-end presentations to the Vodafone ExCo turned progress into investment: an additional €170M in funding after year one, and a further €350M after year two.
One marketplace replaced four fragmented stores — reducing the frontend burden of highly secure Microsoft transactions, simplifying tenant porting, and opening cross-sell and up-sell that fragmentation had blocked. And the business backed it, twice.
The work was presented to the Vodafone executive committee at each year-end. Both reviews unlocked a further round of investment to scale the programme — an additional €170M after year one, and €350M after year two — a rare vote of confidence in the product direction.
The heavy frontend of secure Microsoft transactions was tamed and tenant porting simplified — driving easier adoption and unlocking cross-sell and up-sell opportunities.
Core journeys improved continuously through user testing and performance monitoring over 2.5 years live — directly resolving customer pain points and lifting CX metrics.
A single master-site and token library kept four markets consistent and made the next country a configuration, not a rebuild.
The clearest way to judge the work is the shape of the programme before and after — for the customer buying, and for the team shipping.
Lessons from 2.5 years on one long-lived B2B platform — the things I'd take into the next brief.
The win wasn't removing Microsoft's steps; it was sequencing and explaining them. Reframing a constraint beat fighting it.
Strict tokenisation is what turned “four markets” from a cost into a capability. The design system was the product's scalability.
A single, enforced hierarchy removed endless per-asset debates and protected trust across every surface.
The biggest gains came from 2.5 years of measured iteration, not the first launch. CX is a practice, not a milestone.