Job store-risky-order-hold-rules-backtested · revised 12 October 2026
Hold risky orders for your review, using your own chargeback history to test the rules
Tag-and-hold rules for risky orders, written from 8 months of your orders and chargebacks and tested on the latest 4, with false alarms reported. Rules hold for your review and never cancel.
You might be seeing
- Chargebacks arrive after the goods have shipped
- You cannot tell which orders to check by hand
No passwords, keys, card details or admin invites needed to start.
What usually happened
A rule that holds risky orders also holds good ones, such as customers on a VPN or using a freight forwarder. Without testing on your own history nobody knows how many good orders a rule delays. Automatic cancellation of orders on a score may also be a significant decision about a person.
Who it’s for: Owner of a store that has lost money to chargebacks and wants a short list of orders to look at before shipping.
Usually starts when: You have had a run of chargebacks and are afraid of both shipping risky orders and holding good ones.
The result: You get tag-and-hold rules written from the first 8 months of your orders and chargebacks and tested on the latest 4 months, which we do not see until the test run. For each rule you see the chargeback orders it would have caught and the good orders it would have held. A workflow on a test store holds only planted risky orders and cancels nothing.
Check whether this job fits
Answer these without sending files or logins. Nothing is submitted unless you choose to contact us.
What you get
- The rules as plain steps
- The backtest per rule
- The false-alarm list
- The test-store workflow settings
- Steps to switch it off
Included
- Up to five rules on order total, shipping and billing country match, shipping method, order timing and repeat order patterns
- 12 months of orders with chargeback outcomes, without names or addresses
- A backtest per rule on the latest 4 of the 12 months, which we do not see until the test run: chargebacks caught, good orders held
- A workflow on a test store that tags and holds, and never cancels
- The rules as plain steps you can read
Not included
- Cancelling or refusing any order
- Accusing a customer of fraud
- Chargeback disputes (see the dispute evidence job)
- Payment or fraud-provider setup
- Personal data of customers
- Switching anything on in your live store
How we know it’s done
Agreed with you before work starts. Each check produces evidence you keep.
For each rule the report states chargeback orders caught and good orders held on the latest 4 of the 12 months, which we do not see until the test run (rules are written from the first 8 months), and says if fewer than 5 chargebacks fall in those 4 months.
Evidence: The backtest table for the sealed 4 months.
The rules are reproducible: running them twice on the same export gives the same lists.
Evidence: Both outputs compared.
On a test store, the workflow holds all 10 planted risky orders and holds none of 20 planted normal orders (a smoke test).
Evidence: The test orders and their results.
No order is cancelled by the workflow in any test run.
Evidence: The test-store order log.
Sign-off. You re-run the acceptance checks with the steps we hand over, read the results and sign off in writing. Payment follows sign-off.
If it fails. If an agreed check fails, you do not pay for this fixed scope. We hand over what we found and agree whether to stop or re-quote; no surprise work.
When it fits, and when we stop
It fits when
- You can export 12 months of orders with chargeback outcomes and no names or addresses
- There are at least 10 chargebacks in that history, at least 5 of them in the latest 4 months
- You own a test store
- A named person reviews every held order
We stop and tell you if
- There are fewer than 10 chargebacks, because rules cannot be tested on so few
- You want orders cancelled automatically
- The export cannot leave out names and addresses
- Fewer than 5 chargebacks fall in the latest 4 months, so the report says the rules could not be tested on them
What could go wrong
We change a test store or test account only. Nothing is switched on in your live store or helpdesk by us. Your named person applies the result and can switch it off again with the steps we hand over.
Scroll the table sideways to read it all.
| Risk | How we handle it |
|---|---|
| Good customers are held and annoyed. | The backtest reports good orders held per rule, and nothing is cancelled. |
| A rule reads as an accusation. | Rules tag orders as "check" only. No customer-facing message is drafted. |
| Order data contains personal data. | The export leaves out names and addresses before it is sent. |
A separate automated reviewer checks the work and the test results before hand-over. Your named person reviews every held order and turns the workflow on or off. No human supervisor from us is included unless your agreement names one, and at launch the work is largely automated.
How we deliver
We arrange the work and independent review, then show you the result against the agreed checks. You keep authority over your systems.
- Agree the rules to test and the review process
- Receive the first 8 months of anonymised orders and chargeback outcomes
- Write each rule on those 8 months; test it on the sealed 4 months
- Build the hold workflow on your test store with planted orders
- A separate automated reviewer checks the report; check nothing is cancelled
- Hand over the rules, backtest and switch-off steps
This is a one-off job, not emergency cover or a subscription. We confirm eligibility, price, start window and delivery date in writing before you accept, and work starts only after the agreed inputs are in place. An enquiry creates no charge or booking.
Need to keep it working?
Discuss re-testing the rules every six months.
Ongoing work is separately scoped and quoted: no monitoring, response-time guarantee or automatic subscription is included in this job.
Mention the responsibility you need in your enquiry.
What you can check
This is a new service. We have not delivered this job for a client yet.
Other ways to get this done
- Use the fraud analysis in your Shopify admin and review high-risk orders yourself.
- Hold orders over a value limit for a manual check.
Questions
Will it stop fraud?
No. It holds orders for you to review, and the backtest shows how many good orders each rule would delay.
Why not cancel automatically?
A wrong cancellation hurts a real customer, and a decision with significant effect on a person made with no human involvement can be restricted by law.
Does it replace my fraud provider?
No. It adds simple rules from your own history.
Does a pass mean it will be right next month?
No. A pass describes items like the test sample on the day of the test. It is not a promise about future items, and it does not replace your own checking.
Send an enquiry
Send us
- How many orders a month and how many chargebacks in the last year
- What data an order export holds
- Which payment provider you use
- No customer data
Later, once you agree
- The anonymised export of the first 8 months now; the latest 4 months with chargeback outcomes stay sealed until the test run
- A test store that you own
- The name of the person who reviews held orders
You keep your store, your helpdesk and every login. We work on the files you send through a secure route agreed after you say yes, and hand back files and a report. We never ask for store or helpdesk logins, and nothing goes live until your named person applies it.
Email fallback: open your mail app
If website submission is unavailable, review and send the fallback email yourself. An email fallback is not a website receipt. Or write to hello@syntheticindustry.ai with “store-risky-order-hold-rules-backtested” as the subject.