“Before we got Pecan AI, if we had to build a machine learning model, that would have taken us anywhere from 8 to 12 months.”
Manas Desai, Data Engineer
Meet Whistle Express
Whistle Express is a fast-growing express car wash operator with locations across the United States. The company runs on a subscription model, where customers pay monthly for wash plans.
As Whistle expanded through acquisitions and new site launches, retention became mission-critical. In new markets, brand awareness was still developing, competition was intense, and churn risk was higher.
To sustain growth, Whistle needed a scalable way to understand which members were likely to leave – and act before they did.
The challenge
Whistle wanted a predictive churn capability that would allow them to:
- Identify members likely to cancel
- Intervene early with loyalty and marketing offers
- Support retention across newly acquired and competitive regions
- Deliver results without hiring additional data scientists
The constraint: the data team consisted of just three people, responsible for many parallel initiatives. Traditional model development timelines weren’t feasible.
The solution
Why Pecan?
Whistle chose Pecan AI to operationalize machine learning without adding headcount or building complex data infrastructure.
The team connected existing data, defined the churn objective, and moved from idea to usable predictions in weeks instead of months.
“When we usually onboard tools, it takes us time to see value. This wasn’t the case with Pecan AI,” says Desai.
A predictive question marketing could act on
Whistle anchored the project around a clear business moment:
Predict, two weeks after a member recharge, whether the customer will fail to renew in the next month.
This window gave marketing time to deploy targeted incentives, SMS outreach, and regional campaigns before cancellation.
Fast model development with a lean team
Using their existing membership and behavioral data, Whistle trained a model focused on voluntary churn.
Results came quickly:
- Initial models ready in 2-3 weeks
- Production deployment in under two months
- Predictions flowing directly into marketing workflows
“Within a couple of weeks, we were able to start creating initial versions of our predictive churn model, and within two months, we were in full production with the model that marketing now uses to create campaigns.”
How Whistle Express Uses Pecan AI
Pecan’s predictions are now embedded in Whistle’s retention operations.
The team uses them to prioritize outreach, tailor incentives, and plan around seasonal demand patterns. This enables Whistle to:
- Flag members at highest churn risk
- Trigger loyalty offers and SMS engagement
- Focus spend where intervention will matter most
- Prepare for volatility across markets
From predictions to business strategy
Whistle also built an internal dashboard that highlights the key drivers behind future churn risk.
This visibility helps leadership and regional teams understand why members are likely to cancel, shaping decisions about promotions, pricing sensitivity, and market focus.
Precision targeting in paid media
Predictions now guide where advertising dollars go. Whistle uses churn forecasts to concentrate YouTube and regional campaigns in geographies expected to experience higher risk, directing spend where it can deliver the greatest retention impact.
The impact
30% reduction in churn in key markets
Whistle improved retention in new and highly competitive regions.
From months to weeks
What once required 8-12 months now produces actionable predictions in a fraction of the time.
Immediate activation
Marketing integrated predictions into live campaigns within weeks.
Organization-wide adoption
Leadership embraced the initiative and expanded its use across teams.
“Everyone immediately loved it. They loved the capabilities we were able to bring on to the internal team by utilizing a tool like Pecan AI.”
Key Takeaway
Whistle Express shifted churn management from reactive reporting to proactive intervention.
With a three-person team, they deployed a production-grade predictive system in under two months, reduced churn by 30% in new markets, and equipped marketing with a scalable engine for loyalty and retention.
