Help / Manual
This app helps you review manufacturer name cleanup suggestions. It does not change the original manufacturer table while you review. You decide which suggestions are approved, rejected, or ignored.
Work carefully. A suggestion can look useful but still be wrong for your business data.
A Simple Review Process
- Start on the Dashboard to see how many suggestions are pending.
- Open Groups to review likely duplicate manufacturer names.
- Open a group and compare every row in it.
- Check the original names, usage counts, confidence score, and match reason.
- Edit the canonical name if needed.
- Approve only the rows you are confident about.
- Reject rows or groups that should not be merged.
- Ignore rows that are junk, placeholders, or not real manufacturers.
- Export approved mappings only after the review is complete.
What The Main Words Mean
- Approve
- Mark a suggestion as correct. Approved rows are included in the export.
- Reject
- Mark a suggestion as wrong. Use this when the names should stay separate.
- Ignore
- Mark a row as junk or not useful for cleanup. Examples include UNKNOWN, NO INFO, N/A, punctuation, or notes that are not manufacturer names.
- Canonical name
- The approved clean name you want to use. For example, you may choose Siemens as the canonical name for a misspelling like SEIMENS.
- Confidence score
- A score from 0 to 100 showing how strong the app thinks the match is. Higher scores are usually safer, but you should still check the names.
- UsageCount
- How often that manufacturer name appears in the source data. A high UsageCount usually means the row is important and should be reviewed carefully.
Important Warning About Fuzzy Matches
Fuzzy matching finds names that look similar. It can find spelling mistakes, but it can also suggest matches that should not be merged.
Be careful with names that include subsidiaries, product divisions, joint brands, product ranges, or two manufacturers in one name. These may be related to a brand, but they may not mean the same manufacturer in your data.
Do not approve a whole group just because the confidence score is high. Check each row first.
Examples
| Original name | Possible decision | Why |
|---|---|---|
| SEIMENS | Approve as Siemens | This looks like a simple spelling mistake. |
| SIEMENS VDO | Do not blindly merge into Siemens | It may be a division, related brand, or a different meaning in your data. |
| ALLEN BRADLY | Approve as Allen-Bradley | This looks like a simple spelling mistake. |
Good Habits
- Use Approve selected rows when only some rows in a group are correct.
- Use Reject when the suggestion is wrong but the row is still a real manufacturer.
- Use Ignore only for junk, placeholders, or rows you do not want in cleanup exports.
- Add notes when a decision may need explaining later.
- Review the exported SQL script before anyone runs it.