Case 05 — Vuse Flavour

Overhauling Vuse's entire approach to flavour — with AI.

Flavour was the number-one reason people chose a vape1 — and the whole category was communicating it in the same basic, uninspiring way.

Flavour was Vuse's biggest opportunity and its biggest blind spot. The brand had genuinely high-quality flavours — but they weren't differentiated, their breadth and quality were poorly communicated, and there was only limited consumer data to guide how best to range and sell them market by market. So this became a top-to-bottom overhaul of the brand's approach to flavour: a full review and audit feeding new product development, campaigns, channel experiences and a rebuilt portfolio architecture — with a new AI-powered flavour experience at its centre. The question driving it all: how could the brand fix the immediate problems — undifferentiated, poorly-communicated flavour — while also building the consumer understanding and the foundations for sustainable, long-term growth?

RoleGlobal Category Lead — Flavour

ClientBAT · Vuse

ScopeGlobal

Year2021–22

ProgrammeFIELD · FOCUS

ProofPublic · directional

01The problem

Flavour drove the category — but everyone sold it the same basic way.

Flavour was one of the strongest reasons a consumer chose a vape1 2 — yet the whole category communicated it the same way: basic, uninspiring and skewed too young. Vuse's flavours were genuinely high quality, but they weren't differentiated, and the breadth and quality of the portfolio went largely uncommunicated. Underneath sat a data problem: only limited structured consumer flavour-preference data to guide ranging or net-revenue decisions — a long tail of low-volume flavours, different SKU assortments across markets with little quality data behind them, and sub-optimal flavour experiences across channels.

02The idea

Overhaul the whole approach — benchmarked against the best outside the category.

The answer wasn't a single campaign or tool — it was to overhaul the brand's entire approach to flavour. We looked outside vapour, to best-in-class consumer experiences in other categories, and took our benchmark from there, not from the category's low standard. Part of that meant adapting the AI flavour approach pioneered on What's Your Whisky at Diageo — but right-sized for vapour's specific challenges: tighter regulation, a leaner budget, and a more divergent flavour portfolio.

03The AI role

AI did two jobs: elevate the flavour experience, and turn preference into insight.

The AI played two roles. First, it powered a gamified flavour-discovery experience: consumers reacted to engaging flavour stimuli and received a personalised recommendation, turning a basic shelf decision into something genuinely engaging — built on a bespoke flavour-mapping AI engine, developed with a specialist AI flavour house. Second, every interaction generated structured consumer flavour-preference data, mined for actionable insight that fed back into strategy — which flavours to range where, how to communicate them, and where to focus new product development.

04The approach

Audit first; build pre-data; then adapt on what the data revealed.

The work ran in sequence. It began with a full audit of the brand's flavour presence — portfolio, ranging, communication and consumer experience across every channel. Phase one built the overhaul on the best evidence available before any first-party data existed: new product development, campaigns, channel experiences and the first version of the AI experience. Phase two then adapted everything on the real consumer data the experience generated — sharpening ranging, communication and the experience itself, once the brand finally had a structured view of what consumers actually wanted.

05The outcome

A brand that became the world's number one — with optimised flavour as a growth engine.

Over this period, the flavour overhaul fed a brand on a steep climb: Vuse became the number one global vaping brand by value share, leading in most of its top markets.3 Flavour didn't deliver that alone — but it was a primary reason people chose a vape, and getting it right across the products, the communication, the experience and the data underneath was a real contributor. And the brand came away with something it had previously lacked: structured consumer flavour-preference data, at scale, to guide ranging, NPD and net-revenue decisions for years after.

Sources & substantiation

In a choice-based study of adult e-cigarette users, flavour was the single most important driver of product selection (48.1%) — ahead of product messaging (21.0%) and nicotine level (15.3%). Scientific Reports (Nature), 2021. nature.com
Systematic review of consumer preference for e-cigarette attributes confirms flavour as a primary determinant of product choice, alongside nicotine strength and device type. PLOS One, 2018. journals.plos.org
Vuse became the number one global vaping brand by value share. BAT announcement, September 2021. bat.com

Fuller detail on the experience and its purchase-intent results is available on request. Request the full case study →

Case study describes work led by Andy Parton in a senior role at British American Tobacco plc. All brands and trademarks are the property of their respective owners; Parallax Advisory is independent and not affiliated with or endorsed by them. Figures cited are drawn from the public sources noted; where results are commercially confidential they are shown directionally. Outcomes reflect a specific engagement and are not a guarantee of future results.

FAQ

What is AI flavour personalisation in consumer goods?

An AI experience that turns a few consumer signals — visual ratings, stated preferences — into a personalised recommendation across a product portfolio. The consumer gets a tailored set; the brand learns flavour-print data at scale. Same consumer outcome as expert recommendation, delivered through an engine.

How do you decide how much AI a consumer experience needs?

Match the architecture to the problem. Subtle product differences (single-malt whisky) need deep ontology and learned weights. Categorically divergent products (mango vs. mint) need lighter matching. Over-engineering an experience for an obvious-difference category wastes engine cost and slows iteration.

Why overhaul a whole approach to flavour instead of just building a tool?

Because a clever tool on a weak foundation changes little. Flavour is a primary purchase driver, so the fix had to span the whole system — product, range, communication and experience — not just a front-end. The AI experience was one component of a much larger overhaul; on its own it would have been a gadget, not a growth lever.

Why structure the work as audit → pre-data → post-data?

Because you can't personalise what you've never measured. The audit established the real state of the brand's flavour presence; phase one built the overhaul on the best evidence available before any first-party data existed; phase two then adapted ranging, communication and the experience itself once the AI had generated a structured view of genuine consumer preference. Build on the best assumptions you have, then let real data make the second version sharper.

How is consumer flavour data used after the experience?

First-party signal that informs ranging, merchandising and creative — across digital and retail. The flavour-print becomes a planning input, not just a recommendation output. The data compounds; the consumer experience and the commercial intelligence improve in the same loop.

How do you measure whether an AI consumer experience is actually working?

By the commercial signal it drives, not engagement vanity metrics. Here the decisive measure was purchase intent inside the experience versus the site-wide benchmark — a direct read on whether personalised discovery moved people closer to buying. The full benchmarks and per-market activation detail are available on request.

Can I see the detail behind the headline figures?

Some additional detail can be shared under NDA — please request the full case study or start a conversation below.

← Index ← Previous: Vuse Sustainability Next: What's Your Whisky →

What's next

Two ways forward.

The full case

Architecture, the 2.0 iteration learnings, per-market activation — available on request.

Request the full case

A conversation

About building an AI consumer experience that iterates and compounds. FIELD or FOCUS.

Start a conversation