Decision · 12 plays
Do this now.
Altvary identifies a specific customer, explains why they matter, recommends a product with the offer you’ve set, and ranks it by what the opportunity is worth.
What Altvary looks for
20 ways Altvary finds a reason to act.
Altvary watches for meaningful changes in purchase timing, routines and customer value — then turns the ones worth acting on into decisions, and surfaces the patterns worth knowing.
How Altvary thinks
Know what matters. Act on what deserves attention.
Every play produces one of two things, so a customer attribute never turns into another campaign by accident.
Customer Intelligence · 8 plays
Know this about your customers.
Altvary identifies important customer groups and behavioural patterns that can inform your campaigns, segmentation, merchandising and retention strategy.
From store data to a decision
Five steps from your store to one decision.
- 1ReadConnect Shopify with read-only access to your store data — products, orders, customers, inventory. Altvary only decides about customers whose Shopify record says they subscribed to marketing.
- 2NoticeEvery night, each play checks every customer for one specific pattern, using the fixed thresholds listed below. A play that needs product details it doesn't have stays empty rather than guessing.
- 3Pick oneA customer can match several plays at once. Altvary chooses between them, so each customer gets one decision, not a pile of messages. The same play won't come back for the same customer within 30 days.
- 4RankToday lists decisions by what each opportunity is worth: its expected value, weighted by a 0–100 confidence score. A decision that would lose money after its discount is held back, unless you allow it for that play. Missing cost data never hides one.
- 5ActEach decision says who, why, which product, and the offer you’ve set for that play. Customer Intelligence gives you the customers behind a pattern, to use in your own campaigns, segments and retention workflows.
When plays compete, safety comes first
For the routine and product plays, the first match in this order wins. Nothing further down can override something above it.
- SafetyA recent return citing irritation (a 21-day hold), then New-active introduction
- Account protectionHousehold accounts, then Margin erosion
- Due nowProduct exhaustion, then Product freshness
- Win backRoutine dropout, then Active-ingredient dropout
- GrowRoutine gap
Every play
What each play notices, and what it suggests.
12 are Decisions: do this for this customer. 8 are Customer Intelligence: know this about these customers.
Replenish
Something they use is due again, running out, or past its best.
- R05Decision
Repurchase timing
- Notices
- A repeat customer who is 80–120% of the way through their own usual gap between orders.
- Suggests
- A reorder reminder, timed to that customer's rhythm rather than a fixed calendar.
- Reads
- Orders
- R06Decision
Product exhaustion
- Notices
- A product estimated to run out within 7 days, or that ran out up to 30 days ago — from the size they bought and the category's typical daily use. Only while it is in stock.
- Suggests
- Replenish that exact product.
- Reads
- Orders, product sizes
- R10Decision
Product freshness
- Notices
- An opened product within 14 days of its period-after-opening, or up to 30 days past it.
- Suggests
- A fresh replacement, before the product loses potency.
- Reads
- Orders, period-after-opening
- R16Customer Intelligence
Out-of-stock hold
- Notices
- Customers Product exhaustion would remind — except the product is out of stock.
- Suggests
- A restock priority list, sized by the reorders waiting on it.
- Reads
- Orders, product sizes, inventory
Example 32 customers are due to reorder a product that is currently out of stock.Prioritise restocking before you run replenishment campaigns.
Grow the routine
A good customer with an obvious next product.
- R07Decision
High-LTV entry product
- Notices
- A first-time buyer whose first order was $60 or more (converted to your currency), placed in the last 45 days.
- Suggests
- A reason to make the second purchase.
- Reads
- Orders
- R08Decision
Cross-sell cohort
- Notices
- A repeat customer who ordered in the last 90 days and buys across at least two routine steps.
- Suggests
- A complementary product.
- Reads
- Orders, product categories
- R09Decision
Routine gap
- Notices
- A customer who buys some core routine steps — cleanse, treat, hydrate, protect — but has never bought another. Shared accounts are left out.
- Suggests
- The product that completes their routine.
- Reads
- Orders, product categories
- R25Customer Intelligence
Seasonal formulation shift
- Notices
- A repeated pattern in past seasons: richer textures heading into winter, lighter ones into summer. Seasons follow the northern hemisphere.
- Suggests
- An audience to reach before the season turns.
- Reads
- Orders, product texture
Example 96 customers have switched to a richer or lighter texture with the seasons before.Review this audience when you plan seasonal merchandising.
Win back
Someone, or some habit, has gone quiet.
- R02Decision
Revenue-ranked winback
- Notices
- A customer whose last order was 45–90 days ago and who has spent $80 or more in total (converted to your currency).
- Suggests
- A winback, ranked by the revenue it could recover.
- Reads
- Orders
- R04Decision
VIP score-drop warning
- Notices
- A VIP whose customer score fell 8 or more points in 7 days while they are overdue by their own rhythm — or 15 or more points, whatever their rhythm.
- Suggests
- Reach them while they are still a VIP.
- Reads
- Orders
- R33Decision
Score-drop warning
- Notices
- A returning customer, the tier just below VIP, whose customer score fell 8 or more points in 7 days while they are overdue by their own rhythm — or 15 or more points, whatever their rhythm.
- Suggests
- Reach them before they lapse. It's a score drop, not a prediction that they will leave.
- Reads
- Orders
- R23Decision
Active-ingredient dropout
- Notices
- A customer who bought a key active in two or more separate orders, then nothing containing it for 90 days. Never an active that irritated them.
- Suggests
- Re-engage them on that active.
- Reads
- Orders, product ingredients
- R28Decision
Routine dropout
- Notices
- A customer who used to buy at least two core routine steps and has bought none of them for 90 days.
- Suggests
- Win back the routine, not just one product.
- Reads
- Orders, product categories
- R32Decision
Bundle dropout
- Notices
- A customer who bought a bundle at least twice and has bought none for 90 days.
- Suggests
- The set they used to buy — not their latest single product.
- Reads
- Orders, which products are bundles
Protect the customer
Moments to hold back rather than sell.
- R12Customer Intelligence
New-active introduction
- Notices
- The 21 days after a customer's first purchase of a strong active, such as a retinoid or an acid.
- Suggests
- A hold: for those 21 days no other play makes a decision for this customer. An audience to introduce the product gently instead.
- Reads
- Orders, product ingredients
Example 58 customers started a retinoid or an acid in the last three weeks.Send how-to-use and barrier-care content rather than another active. - R27Customer Intelligence
Reaction risk
- Notices
- Two or more strong-active products bought in the last 60 days, or a recent strong active with no moisturiser ever bought.
- Suggests
- An audience for barrier-support education, rather than another active.
- Reads
- Orders, product ingredients and categories
Example 41 customers show signals associated with potential product-reaction risk.Consider education and barrier-care content rather than another product promotion.
Know the account
Facts about a customer that change how you should treat them.
- R11Customer Intelligence
Margin erosion
- Notices
- A customer whose recent margin is 10 or more points below their own earlier orders, as they lean on discounted or low-margin items.
- Suggests
- An audience to take out of your discount flows.
- Reads
- Orders, product costs
Example 84 customers' margin is sliding as their orders lean on discounts.Use this audience to limit their discount exposure. - R13Customer Intelligence
Household accounts
- Notices
- An account buying for opposite skin concerns — at least two products for younger skin and two for mature skin — most likely two people sharing it.
- Suggests
- An audience to split the messaging for, instead of recommending for one person. Plays below it in the competing order, such as Routine gap and Product exhaustion, skip these accounts.
- Reads
- Orders, each product's skin concern
Example 19 accounts buy for both younger and mature skin.Keep them out of single-profile routine campaigns, or message each concern separately. - R21Customer Intelligence
Skin-type loyalty
- Notices
- Three or more concern-tagged purchases, at least 70% of them for one concern, on an account that isn't shared.
- Suggests
- An audience to speak to in terms of that concern.
- Reads
- Orders, each product's skin concern
Example 418 customers consistently buy products for sensitive skin.Use this audience for targeted product launches and campaigns. - R30Customer Intelligence
Brand advocates
- Notices
- Three or more orders, a first order at least 180 days ago, at least one order detected as a gift, and no reaction history.
- Suggests
- An audience for referral or review invitations.
- Reads
- Orders, gift orders, returns
Example 73 customers show strong signals of brand advocacy.Consider them for reviews, referrals, UGC or community campaigns.
Thresholds are the ones the product uses today, and dollar amounts are converted to your store’s currency. None of these plays promises a result. Decisions and what customers did afterwards are recorded together, with a holdout group left untouched for comparison, so what a play is worth on your store can be measured rather than assumed.
Start with your catalog
See what your own catalog already says.
The Preview reads your public product list. No account, no install, nothing to connect.