Churn prediction.

Every night, Daima gives each active subscriber a 0–100 risk score based on how they have behaved. The subscribers who look most likely to cancel are listed for free on your Analytics page; the full Churn Prediction page is part of Pro. Reach out to the at-risk subscribers before they hit cancel.

Last updated · November 7, 2026

The likely-to-cancel list (free)

On any plan, open Analytics in the Daima admin. You will find the list in two places:

• On the Overview tab, the Needs you card shows “N subscribers look likely to cancel” with a See who link.
• On the Subscribers & churn tab, the Likely to cancel card shows the five subscribers most at risk, with See all for the full list (up to the 200 highest scores).

Each row shows the customer, why Daima thinks so (for example “2 consecutive skips”), their next order date and what they pay per month. Click Open to go to that customer in Shopify. The list includes every subscriber scoring 40 or higher and refreshes every night. It has no plan restriction. See Reading the Analytics page for the rest of Analytics.

How the score is calculated

Daima combines the signals below into a single score from 0 to 100. Higher means a higher risk of cancelling. The signals are:

• Skips in a row. Subscribers who skip two or more deliveries in a row are far more likely to cancel within 60 days. A skip the customer undid doesn't count.
• A late order. How long it has been since their last order compared with how often they are meant to receive one.
• Few orders on an older subscription. Someone who has been subscribed for a while but has received only one or two orders.
• Failed payments. Payments that failed on the subscription in the last 90 days.
• A billing date that has passed without a successful order.
• A brand-new subscriber. The first few weeks carry a little extra risk.
• A paused subscription. A paused subscription is never scored below 40, so it always shows up as at least a watch-list customer.

Whichever signal contributes most is shown as the “why we think so” reason on each row. A score under 20 is labelled “Healthy.”

Risk tiers

• High Risk — score 70 or above. Likely to cancel soon unless something changes.
• Medium — score 40 to 69. A watch list: worth a check-in, but not urgent.
• Low — score below 40. Healthy subscribers. No outreach needed.

The free list on Analytics covers High and Medium. The Low tier is visible on the Pro Churn Prediction page.

Nightly recalculation

Scores are recalculated automatically every night as part of Daima's overnight Analytics run, for every active and paused subscriber on every plan. A skip, a failed payment or a pause shows up in the next morning's scores. Customers whose subscription has ended drop off the list.

The Churn Prediction page (Pro)

On the Pro plan, Churn Prediction in the admin nav opens the full page. Free stores see an “Unlock Churn Prediction with Daima Pro” screen with an upgrade button instead. The page has:

• Four summary tiles: High risk (score 70+), Medium risk (score 40–69), MRR at risk (the monthly value of your high-risk subscribers) and Avg risk score.
• A Subscriber Risk table of every scored subscriber with their product, risk score, risk level, main signal, monthly value, number of orders and how many days they have been subscribed. Use the filter to show All subscribers, High risk (70+), Medium risk (40-69) or Low risk (<40).
• A Refresh Scores button in the top corner. It re-scores your subscribers right away from the information gathered during the last overnight run, which is useful after you change something. If your first overnight run hasn't finished yet, it tells you to check back tomorrow.

Act on risk automatically with Shopify Flow (Pro)

When a subscriber's score reaches high risk (70 or above), Daima fires a Churn risk detected trigger in Shopify Flow. Build a Flow that starts from it — for example sending a win-back email or creating a discount — and it runs without you. The trigger carries the customer, the risk score, the strongest signal, the product and the monthly value. Each subscriber triggers it at most once every 30 days, so a score that stays high doesn't send a new email every day.

Acting on the list

Use Open on any row to go to the customer in Shopify, then reach out yourself — email them, or offer something. Some practical ideas:

• Don't ignore Medium. The High list is small but late. The Medium list is bigger and earlier — that's where retention effort works hardest.
• Make it personal. A message that mentions their actual product or last order beats a generic “we miss you.”
• Check the cause. A customer flagged for failed payments may just need to update their card — the expiring-cards list on Analytics can send them Shopify's update link in one click.
• Pause beats cancel. Customers who start to cancel in the portal are offered a skip, a pause or a swap first. The Saved before they left card on Analytics shows how often that works.
• Test your offers. If you try discounts or reward levels, A/B test the widget side and watch the cancel numbers on Analytics.

Honest limitations

The score is a set of simple rules based on past behaviour — it doesn't know why a subscriber is at risk. False alarms happen (someone who skipped twice because they were on holiday). Misses happen too (a customer whose card quietly expires with no other warning sign — that one is covered by the expiring-cards list instead).

Use the score to prioritize your outreach, not to make automated decisions about people. It is a sort key, not a verdict.