Wrapped
Marketing had never used on-slope behavioural data. A chance comment in a meeting changed that. Two weeks later, every skier and snowboarder had a personalised season recap built from their own data. Six months on, riders who opened their recap are converting to season pass-holders at twice the rate of those who didn't.
On-slope behavioural data had never been used in marketing. Nobody had figured out how to connect it to the tools.
During a call in my first week with my client's marketing team, I asked a colleague how their week was going. They told me about sending NPS surveys to customers who'd been on the mountain the previous week.
"How do you know who was on the mountain last week?"
"I email them. After exporting them from the system."
The data wasn't locked away at all. It was being hand-exported for customer satisfaction surveys, and the only thing missing was someone asking what else we could export, and what it could be used for.
A season's worth of data
The ski season was almost over, but we had detailed on-slope data for thousands of customers: days skied, consecutive days, earliest arrival, biggest day on the mountain.
The idea: a Spotify Wrapped-style season recap, built for each customer on their own data.
The bottleneck has gone from how to build, to what to build.
The response was enthusiastic, then immediately: "If only we could do this..."
A few years ago that hesitation might have been well-founded. But AI had changed the equation. I had enough technical knowledge to describe exactly what needed to be built to execute this, and AI advances meant the bottleneck between that and being able to actually build it had disappeared. I went away and couldn't let it go.
Building It
Two weeks with a lot of moving parts.
- The badge system. Working with the team, we defined criteria for each badge. Century Carver for 100 days on mountain. Powder Hound for riding an epic pow day. Alpine Elder for still chasing it at 75+. Sixteen badges in total, each with a rule, each needing to map cleanly to the data. Those rules went to our designer, who produced SVG assets built for procedural assembly rather than a one-off layout.
- The data. Multiple CSV and Excel exports from different systems, differently structured and differently organised. Python to join and clean it into a single master dataset: one row per customer, badge status across all sixteen. Customer data never went near an AI model throughout. AI only ever saw abstract descriptions of file structures or sanitised examples.
- The images and the email. A script read each row, assembled the right SVG components for that customer's badge combination, rendered a unique image, and uploaded it to AWS S3. Each image got its own URL, which went into the email as a preview, with a click-through to the full version.
Concept to delivery in less than two weeks.
What Happened
Recipients posted their Wrappeds to social media. Others started asking how they could get one. At the same time, parents were contacting guest services asking if their children could get a Wrapped. We'd only generated them for email subscribers, so kids had missed out. For each of those requests the script would be run on that customer's data.
One came back almost straight away. The mum had received her son's Wrapped and replied to say it was missing his snow streak badges, and that he'd skied 100 days that season, not the 98 the system had recorded. The extra two days had scanned on a different pass and hadn't rolled up correctly.
I adjusted the data and ran the script again. There wasn't much deliberation about whether to take a parent's word over the system. What's the alternative, telling an 8-year-old his mum was wrong about how many days he'd spent on the mountain? He got his Century Carver badge, his streak badges, and was, according to his mum, a very happy wee boy.
People not on the email list were DMing to ask how they could get their Wrapped. The data had given people something that felt genuinely theirs, and everyone in the business wanted to keep finding ways to do that.
Wrapped's impact
Non-passholders who opened their Wrapped are converting to 2026 season passes at twice the rate of those who didn't, with additional uplift on retention and other metrics.
Wrapped proved the data could be reached, cleaned, and used. The pipeline is still there: extracted, structured, and ready for personalisation, segmentation, re-engagement, lifecycle triggers, whatever the next campaign demands. The first one did the hard work of proving it. The next ones compound from there.
A question about NPS opened the door. The data infrastructure holds it open.