Kliovo Shop
7 Fraud Checks on Every Order. Before the Box Leaves Your Shelf.
Anomaly Detection in Kliovo Shop runs 7 automatic fake order detection checks on every order in 0.4 seconds — flagging fake orders, input errors, high-risk COD, and stale shipments before they cost you money. Six auto-triage rules then categorize orders based on customer history.
How Anomaly Detection works
Order arrives
Immediately on creation, the anomaly engine evaluates all 7 checks simultaneously. Decision in 0.4 seconds — before any operator sees the order.
Severity assigned
Each triggered check gets a severity: Critical (blocks dispatch), Warning (review recommended), or Info (FYI). Color-coded badges on the order row.
Auto-triage evaluates history
6 triage rules check customer phone history: undelivered count, pending duplicates, coupon usage, address length, return history, and payment status.
Dashboard surfacing
Critical and Warning orders appear in dedicated queues on your dashboard. Operators review and clear, hold, or escalate — clean orders proceed automatically.
Everything included
7 anomaly checks plus 6 auto-triage rules — all always-on, zero configuration required.
7 Anomaly Checks
Invalid phone number
Format validation plus basic carrier check. Catches typos and obviously fake numbers like 0300-0000000 before they reach dispatch.
COD amount = 0
Order placed as COD but COD amount is zero. Almost always a data entry error — either the order isn't actually COD, or the amount wasn't filled.
Unconfirmed COD for 24+ hours
COD order placed but no WhatsApp confirmation received after 24 hours. High RTO risk if dispatched without customer acknowledgment.
Stale order (7+ days)
Order created 7+ days ago and still in processing status. Usually a forgotten order or one that fell through the ops cracks.
Dispatched without CN
Order status is dispatched but no CN number is stored. The order was status-updated manually without actually booking a shipment — impossible to track.
Missing city
Address field has no recognizable city name. The courier will reject the shipment at booking or delivery. Catch this at intake, not on the truck.
High-value COD (>Rs 15,000) unconfirmed
Configurable threshold. Large COD orders carry higher financial risk if they return — flag them for manual confirmation before dispatch.
6 Auto-Triage Rules
Undelivered Hold
Customer's phone has 3+ undelivered orders in history. Auto-hold: no confirmation message sent until operator reviews.
Double Pending
Customer's phone already has a pending unconfirmed order. Possible duplicate — hold for review before any confirmation goes out.
Coupon Hold
Discount applied exceeds your configured percentage threshold. Prevents coupon abuse on large orders before they're confirmed.
Incomplete Address
Address field is below minimum character count. Tags order for address verification before CN booking can proceed.
LO Return
Customer has previous orders with return_to_shipper status. High-risk historical flag surfaced on every new order from that phone.
Payment Pending
Non-COD order has an advance amount recorded. Means partial payment — hold until full payment is confirmed before dispatch.
Built for teams that...
The brand getting hammered by fake orders
The pain: We were shipping 20–25 fake orders a week — Rs 8,000–15,000 COD, elaborate delivery addresses, phone numbers that were turned off. Each one cost us Rs 850+ in logistics.
The fix: Kliovo's Undelivered Hold rule now catches the patterns — same phone, multiple orders, multiple returns. We've blocked over 300 fake orders in 3 months.
The ops team finding errors before dispatch
The pain: COD=0 errors and missing city flags catch data entry mistakes our WooCommerce forms sometimes allow. Before Kliovo, these would go out, get rejected by the courier, and come back 3 days later.
The fix: Now they're caught on arrival — before CN booking, before dispatch, before any cost is incurred. Saves us courier return charges on 15–20 orders a month.
The brand with high-value COD exposure
The pain: Our average order is Rs 18,000. Every unconfirmed high-value order that ships and returns costs us Rs 1,200+ in logistics — on top of the working capital tied up with the courier.
The fix: The high-value COD flag gives us a manual review checkpoint for those orders specifically — without slowing down the Rs 3,000–6,000 orders that move normally.
The numbers
7
Automatic fraud and quality checks per order
All simultaneous. Decision in 0.4 seconds.
0.4s
Time to evaluate all 7 checks per order
Faster than any operator can open the order detail page.
Rs 850
Average logistics cost saved per blocked return
Courier return fee + warehouse receiving + working capital cost.
Common use cases
Daily fraud queue review
Every morning, open the Critical queue. These are orders that triggered the highest-severity anomaly checks overnight. Review each one: invalid phone gets cancelled immediately, high-value unconfirmed gets a manual call, COD=0 gets corrected. 10–15 minutes clears the queue and prevents a full day of problem shipments.
Tuning the high-value threshold
The Rs 15,000 default threshold might be wrong for your business. A brand selling luxury skincare at Rs 25,000+ might set it to Rs 20,000. A low-margin electronics brand might set it to Rs 8,000. Configurable per tenant — calibrate it to your actual risk tolerance, not an arbitrary default.
Stale order cleanup
Weekly, run the stale order check — all orders 7+ days old in processing status. Typically falls into two categories: orders the customer cancelled verbally but no one updated in the system, and genuine exceptions. Clean them up in bulk using status update bulk action before they distort your aging reports.
Frequently asked questions
Can I customize the anomaly thresholds?
Yes. The high-value COD threshold (default Rs 15,000) is configurable per tenant. The stale order window (default 7 days) and undelivered hold trigger count (default 3 undelivered orders) are also configurable in your store settings.
Do anomaly checks run on manually-created orders too?
Yes. All 7 checks run on every order regardless of source — WooCommerce sync, Shopify webhook, Daraz import, manual creation, or API. There are no exceptions. If an order enters Kliovo Shop, it gets checked.
What is the difference between anomaly detection and auto-triage?
Anomaly detection checks the order's own data: phone format, COD amount, age, CN status. Auto-triage checks the customer's history: previous undelivered orders, pending duplicates, return patterns. Both run on every order and surface results together.
Can I turn off specific anomaly checks?
Individual checks can be suppressed for specific scenarios — for example, you might disable the COD=0 check if you have a product that is legitimately zero COD for VIP customers. This is configured at tenant level in your anomaly settings.
Where do flagged orders appear?
Critical orders appear with a red badge in the main order list and in a dedicated Critical queue card on the dashboard. Warning orders appear with an amber badge. Info flags appear as subtle tags on the order row — visible without being disruptive.
How does the double-pending check work?
When an order arrives, the engine checks if any other order from the same phone number is currently in an unconfirmed or pending status. If yes, it holds both — one of them is likely a duplicate or an impulsive second order the customer did not intend to place.