Discarding Changes

When to discard changes

• A prompt or query set was too narrow and skewed your baseline.
• You tested new positioning or pricing language that you don't want influencing ongoing crawls.
• A competitor sprint temporarily distorted share of voice and you want to reset comparisons after cleaning data.

How to revert safely

• Duplicate the collection first so you have a snapshot of the test before discarding it.
• Remove test prompts from active watchlists and mark them as archived so they stay out of reporting.
• If you changed brand terms, restore the previous canonical list before running the next crawl.

What happens to history

• Removing prompts stops future runs; historical runs stay available unless you request deletion.
• Share of voice and citation accuracy recalculates on the next crawl using the active prompts and brand terms.
• Exports created before a discard stay intact for auditing or internal post-mortems.

Best practices

• Make one change at a time so you can attribute shifts clearly.
• Annotate the timeline with context (campaigns, launches, migrations) before discarding, so future reviews stay clear.
• If you are unsure, pause a collection instead of deleting it. You can always resume and compare before/after.