Part 3 of 4: One Source of Truth, Four Audiences
How a national airline reached four distinct customer segments across five markets with a single deterministic data foundation and predictive scoring on every audience.
By Lauren Walsh, Senior Sales, Agencies and Platforms
Part 2 of our series showed what purchase precision and predictive scoring do to a single funnel. Part 3 shows what they do to four of them at once.
The Problem: Four Audiences That Demand Different Truths
A leading U.S.-based airline didn’t need a bigger audience. It needed better ones — each with a different message, a different objective, and a different definition of success — calibrated across five priority markets. Off-the-shelf demographic and behavioral segments don’t bend that way. Custom Consumer Purchase Audiences do.
The airline’s campaign targeted four customer segments, each requiring its own narrative and each mapping cleanly to a high-impact custom Consumer Purchase Audience archetype:
- At-Risk loyal customers showing early signs of lapsing: a Lapsed Shopper Re-Engagement play.
- Conquestable consumers currently flying with competitors: a Competitive Conquesting play.
- Core high-value loyalists to retain and deepen: a High-Value Acquisition / retention play.
- Dreamers prospective travelers who hadn’t yet converted: a propensity-led acquisition play.
Reaching all four with the same level of precision, using one campaign, across the same five markets, is a problem most data sources can’t solve. They can build one good audience. They can’t reliably build four.
The Solution: Audiences Built on Frequency, Spend, Geo and Scored for Future Purchase Likelihood
Affinity built a custom Consumer Purchase Audience for each of the four segments. Every audience was constructed from first principles using the airline’s own definitions of each customer, drawing on the full set of our granular dimensions: recency, frequency, spend thresholds at the brand and category level, demographic attributes, and geographic overlays.
We then layered Propensity Scores onto each segment, prioritizing the consumers most likely to act on the offer in front of them. Geo-targeting tuned every audience to its five markets. The result was four sharp, distinct audiences, all powered by the same deterministic foundation and predictive engine.
One source of truth. Four audiences. Five markets. Every segment outperformed.
The Result: 80% VCR and 16% CTR (Above Benchmark Across the Board)
- 80% Video Completion Rate compared to an industry benchmark of 70%.
- 16% Click-Through Rate across all four audience segments — compared to an industry benchmark of 11%.
Every audience performed above benchmark. Across four definitions and five markets, the campaign delivered consistently — a result that reflects the quality of the underlying data and the precision of predictive scoring.
Precision at Scale
Purchase frequency is a precision dimension. Building audiences around how often consumers actually transact, rather than how they self-identify or what they browse, produces more accurate, higher-performing targeting.
Custom audiences scale to complexity. When a brief calls for multiple distinct audiences, off-the-shelf segments force compromise. Custom segments flex to each definition without losing fidelity, then predictive modeling ranks every audience for likelihood to convert.
Geo plus purchase plus prediction is a force multiplier. Pairing verified purchase behavior with location intelligence and propensity scoring makes regional and multi-market campaigns dramatically more efficient.
Run This Play: The Multi-Segment Map
- When to run it. You’re managing distinct customer states across regional priorities which is common in airlines, hospitality, multi-location retail, and QSR. One-size-fits-all targeting can’t deliver the relevance each segment requires.
- What you need. Clear customer-state definitions (At-Risk, High-Potential, Core, Dreamers, or your own); CPA dimensions tuned per segment (recency, frequency, spend thresholds, demographic and geographic overlays); Propensity Scores layered on every audience; differentiated creative per segment.
- What it delivered. 80% video completion rate and 16% click-through rate with every audience above benchmark across four customer segments and five priority markets.
Up next — Part 4: Attention That Converts.
Read Part 1: A Playbook for Purchase-Led Targeting here.
Read Part 2: When Proof Beats Proxy here.
