Ultra-premium whisky has almost no organic discovery online. Nobody searches for a bottle that costs more than a flight to Cape Town. If a brand like Johnnie Walker Blue Label wants more sales during the festive season, paid media has to find those buyers. It cannot just influence those who are already circling.
That is not a whisky problem. It is what happens to any brand whose real market is a niche of the population: premium skincare, boutique travel, high-end electronics, anything gifted rather than grabbed on impulse. The addressable audience is small on a good day, and most standard targeting was built for audiences that are not.
The fix is not a bigger budget or a wider net. It is a first-party data strategy:
swapping data that guesses who might buy for data that already knows who does.
Interest is a guess. Purchase is a fact.
Most targeting still runs on implied interest: someone likes a whisky page, someone looks like people who bought whisky, someone sits in a demographic bracket associated with whisky drinking. None of that is a purchase. It is a proxy for one, and proxies are exactly where wasted spend enters a campaign. Every impression served against a maybe is one impression not served against a yes.
Flow’s Audience Marketplace classifies audiences by how close they sit to that yes. Tier I is verified first-party shopper data, matched 1:1 to real people, built from what they actually bought, not what they browsed or resembled. It is the most precise tier by design. For an ultra-premium product whose buyer pool was never going to be large, it is the only tool that makes sense.
That precision has a track record. First-party data audiences convert at 50%+ higher rates than the modelled alternative. When the buyer pool is small, it decides whether the campaign works at all.
Flow’s Audience Marketplace classifies audiences by how close they sit to that yes. Tier I is verified first-party shopper data, matched 1:1 to real people, built from what they actually bought, not what they browsed or resembled. It is the most precise tier by design.
For an ultra-premium product whose buyer pool was never going to be large, it is the only tool that makes sense.
How Diageo hit the mark for Johnnie Walker Blue
Johnnie Walker Blue Label had the small-market problem in its purest form: festive season, the most competitive gifting window of the year, for a product almost nobody discovers on their own.
Diageo created the campaign using Flow’s audiences that are built from Woolworths’ verified purchase data: shoppers who had already bought wine and spirits, shoppers with a proven history of premium gifting, and high-frequency buyers in adjacent premium categories. Every impression went to someone who had already shown they buy at this level, then deep-linked straight to the product page. Not one impression was spent finding out who they were.
KEY INSIGHT 💡
The campaign did not try to make the audience bigger.
It made every impression inside that audience count for more.
The results held up: Revenue up 39% year-on-year in the campaign month, a 400% revenue spike in launch week, and 92% of all product-page traffic during the festive peak came from the campaign itself.
For a product with almost no organic discovery, that last number is the tell. The paid media was not supporting the storefront. It was the storefront.
Small doesn’t have to mean unreachable
A small addressable market punishes imprecision faster than a large one does. Waste 30% of impressions on a mass audience of ten million, and the campaign still finds enough real buyers to perform. Waste 30% of impressions on an audience of a few thousand qualified spirits buyers, and there may not be enough real buyers left to hit the number at all.
That is the case for Tier I first-party data used in any niche or premium campaign: not that it performs marginally better, but that at this scale, it is the difference between a campaign that works and one that doesn’t.
Browse audiences built from what people actually buy
A first-party data strategy is only as good as the data behind it. Flow’s Audience Marketplace is powered by data that is verified, tiered by precision, and sourced from real retailers whose customers really made a purchase.
Browse Tier I audiences built from actual purchase behaviour, not modelled interest, on the Flow Audience Marketplace
Frequently asked questions
What is Tier I data, and how is it different from a lookalike audience?
Tier I is verified first-party shopper data matched 1:1 to real people, built from confirmed purchases. A lookalike audience is modelled: built from people who resemble known buyers on measurable traits, not people who have bought anything themselves. Tier I trades reach for certainty. Lookalikes trade certainty for reach.
Why does real purchase behaviour outperform interest-based targeting for niche products?
Interest is declared. Purchase is proven. For a product with a genuinely small buyer pool, the gap between someone who says they like whisky and someone who has bought premium whisky twice this year is the whole campaign. First-party audiences built on purchase behaviour convert at 50%+ higher rates than the modelled alternative, and that gap widens as the addressable market shrinks.
How small can an audience be and still be worth targeting?
Smaller than most brands assume, provided the signal is precise. Johnnie Walker Blue Label ran an entire festive campaign against verified wine and spirits shoppers and premium gifters at one retailer, and it drove 92% of all product-page traffic during the season. The constraint was never audience size. It was audience accuracy.
Does this approach only work for alcohol and FMCG?
No. The same logic applies to any product where the real buyer pool is a small, identifiable slice of the population: premium skincare, luxury travel, high-end electronics, considered financial products. Anywhere the product is gifted or considered rather than impulse-bought, purchase history is a stronger signal than declared interest.
