Know who needs attention before revenue drops
Use purchase history and behavior signals to spot churn risk, seasonal demand, and high-value customer opportunities before they show up in your revenue report. See who to win back, when to prepare for demand, and which actions are worth doing first.
Keeping a customer costs less than winning a new one
Churn shows up when it's already too late. Forecasts flags risk early and tells you where to spend effort now for the biggest return — and when to prep for a seasonal surge.
- Churn risk. Spot customers who are likely to churn.
- Seasonal demand. Predict it before it hits.
- High-value opportunities. Prioritize the customers worth acting on first.
- Straight into action. Turn forecasts into campaigns and reminders.
Churn signal
An early flag from behavior: dropping frequency, longer gaps between purchases.
Seasonal windows
When to prep stock and a campaign to catch the peak, not chase it.
Risk × reward
Balancing what to protect against what to grow — by segment.
Action queue
Priority: who to win back first, who to upsell, who to leave alone.
Trends in the model
Social Recon signals strengthen the forecast — demand shows before sales.
Straight into action
A forecast fires a flow in Quill or a reminder in Bookings.
Guesswork versus a ranked list
Most teams find out a customer churned when they stop showing up. Forecasts moves that moment earlier, so there's still time to do something about it.
The old way
- Planning next quarter off last year's gut feel
- Finding out a customer churned only after they're gone
- No idea a seasonal peak is coming until it's already here
- Scrolling every segment by hand to guess what matters
With Forecasts
- Churn risk flagged while there's still time to act
- Seasonal peaks flagged in advance, with a suggested lead time
- One ranked action queue instead of scanning every segment
- Priority set by value at risk, not by whoever asks loudest
What's coming in the next 30 days
The same engine, one layer closer to your calendar: a short list of what to do this month and how confident the model is in it.
From history to what to do today
Data
Purchases, frequency, gaps from CDP.
Context
Social signals and seasonality.
Map
Risks, peaks, priorities.
Launch
A flow in Quill or a reminder.
Questions before you rely on a forecast
How far ahead can it actually predict?
Enough to act on. Churn risk typically shows up weeks before a customer would otherwise go quiet, and seasonal peaks are flagged with a lead time so you can prep stock or a campaign instead of reacting after demand has already spiked.
What data does it need from us to start working?
Purchase history is the baseline — order dates, frequency, gaps between visits. Connecting Social Recon adds trend context on top, but Forecasts starts producing useful signal from purchase history alone.
Does it make decisions for us, or just recommend?
It recommends. Forecasts ranks who to act on and why, and can trigger a flow in Quill or a reminder in Bookings, but nothing goes out to a customer without a flow you've already approved.
How accurate is this for a small customer base — do we need thousands of customers?
No. The model weights recent behavior per customer rather than needing volume to average across, so it's built to work for a base of a few hundred, not just a few thousand. Confidence is shown alongside every forecast so you can see how much weight to put on it.
What happens to a forecast if our sales pattern changes suddenly — like a new product launch?
The model re-weights as new data comes in, so a sudden shift shows up as lower confidence on the affected segments rather than a stale prediction. It catches up as the new pattern establishes itself.
Be the first to see ahead
Request access — we'll connect Forecasts to your data at launch.