Job shopify-bulk-edit-dry-run-diff-and-rollback · revised 12 October 2026
Make one bulk product change, by AI or by an app, safe to try and safe to undo
A tested procedure and checking script so a bulk change is exported first, dry-run compared, applied in batches and restorable, proved on a copy with six planted bad edits.
You might be seeing
- A past bulk edit changed more fields than intended
- Nobody saved a copy of the products before the last import
No passwords, keys, card details or admin invites needed to start.
What usually happened
A bulk change can multiply a price by ten, blank a vendor field or remove variants across hundreds of products at once. A file that imports without error is not proof that the result is right. Without a before copy and a row-level comparison, restoring means rebuilding from memory.
Who it’s for: Owner or developer of a Shopify store that uses AI tools or apps to change prices, copy, vendors or variants in bulk.
Usually starts when: You are about to run a bulk change and cannot say what will happen to your products if it goes wrong.
The result: You get a tested procedure and a checking script. A bulk change is exported first, compared with the before copy on a dry run, applied in small batches with a sample check and restorable. On a copy of 100 of your products, six planted bad edits are caught before apply and a rollback file restores every changed field.
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 runbook
- The comparison script
- The rollback file generator
- The batch plan
- The rehearsal log
Included
- One bulk-change path, for example a CSV import or an app, for products and variants
- A written runbook: export, dry run, compare, apply in batches, check, roll back
- A row-level comparison script that flags price jumps over a factor you set, blank overwrites, status, vendor and variant changes
- A rollback file generator
- A rehearsal on a copy of 100 of your products with six planted bad edits, three of them yours
Not included
- Applying any change to your live store
- Running your bulk tool for you
- Prices, copy or product decisions
- Theme or checkout changes
- Order, customer or inventory history
- Stores you cannot copy or test
How we know it’s done
Agreed with you before work starts. Each check produces evidence you keep.
On a copy of 100 of your products, the comparison flags all 6 planted bad edits before apply, in each of 5 repeat runs: 3 of ours (price multiplied by ten, vendor blanked, variants removed) and 3 chosen by your named person and kept from us until the test run (6 of 6 shows a true catch rate of at least 60% with 95% confidence; a smoke test). One missed edit in any run fails the test.
Evidence: The planted list, with your three kept sealed until the run, and the five run outputs.
The rollback file restores every changed field on the copy; a comparison of the restored copy with the original shows zero differences.
Evidence: The comparison report.
Each batch in the batch plan has a recorded sample check.
Evidence: The rehearsal log.
The runbook is followed once end to end by your named person on the copy.
Evidence: The log of that run.
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 make a test store or a product copy
- You have an export and an import path (Shopify CSV or an import app)
- You can send a product export with no customer data
- A named person approves each batch before it is applied
We stop and tell you if
- The plan is for us to apply changes to your live store
- No test store or product copy can be made
- The bulk tool cannot be rehearsed without live effects
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 |
|---|---|
| A planted edit type is missed because it was not thought of. | Three of the six planted edits are chosen by your named person and kept from us. The thresholds and checks are written down and extended with your own past mistakes. |
| The rollback restores less than the change touched. | The test compares the restored copy against the original and requires zero differences. |
| A bulk tool behaves differently on the live store. | The runbook says to apply in small batches and sample-check each one. |
A separate automated reviewer checks the work and the test results before hand-over. Your named person approves each batch and applies the change to the live store. 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 bulk path, the thresholds and the batch size
- Receive a 100-product export and set up a copy
- Write the runbook and the comparison script
- Plant our three bad edits and your three sealed ones; run the dry run
- Apply in batches on the copy, then roll back and compare
- A separate automated reviewer checks; hand over the runbook and script
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-running the rehearsal when your tools or catalogue change.
Ongoing work is separately scoped and quoted: no monitoring, response-time guarantee or automatic subscription is included in this job.
Explore an ongoing managed lane, or 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
- Export your products before and after any change and compare them yourself in a spreadsheet.
- Use an import app that keeps its own backups, and still keep your own export.
Questions
Does this work for any app?
It covers one bulk path, such as one CSV import or one app. Others can be added at extra cost.
Will you run the change?
No. You apply it. We only prepare the procedure and the checks.
Is this a repair?
No. The repair job is for after something went wrong. This one is for before.
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.
Is this the same as the AI copy check?
No. The AI copy check traces the claims in written product text. This job checks that a bulk change does not alter fields it should not, and that you can undo it.
Send an enquiry
Send us
- How you make bulk changes today and with which tool
- Roughly how many products and variants
- The price-jump factor you want flagged
- No store login
Later, once you agree
- A product export of 100 products through an agreed secure route
- A test store or copy that you own
- Three bad edits of your own choosing, sealed until the test run
- Your batch size and the person who checks each batch
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 “shopify-bulk-edit-dry-run-diff-and-rollback” as the subject.