FuzzyLookup Errors

[ExcelMac720]

Hey Everyone,

I have two list, each of which contain close to 10,000 companies and various information about both. There are duplicates in both List A and List B, however, only one match is needed from List B. I matched the companies from List A (the more important list) to List B, using FuzzyLookup with the parameters of '1 match' and a similarity threshold of '.85'. The connections it made are fairly accurate. However, upon further review there are many obvious matches it did not make. I've tried other similarity thresholds with the remaining companies from List A to the original List B. The results become much less accurate and the obvious matches were still not made. Does anyone have any suggestions on how to fix this? Or have I exasperated all of my options?

Thanks

 

[ravi2628]

Hey Everyone,

I have two list, each of which contain close to 10,000 companies and various information about both. There are duplicates in both List A and List B, however, only one match is needed from List B. I matched the companies from List A (the more important list) to List B, using FuzzyLookup with the parameters of '1 match' and a similarity threshold of '.85'. The connections it made are fairly accurate. However, upon further review there are many obvious matches it did not make. I've tried other similarity thresholds with the remaining companies from List A to the original List B. The results become much less accurate and the obvious matches were still not made. Does anyone have any suggestions on how to fix this? Or have I exasperated all of my options?

Thanks

Can you share a sample data

 

[Alex Blakenburg]

Are you using the Add-In or Power Query ?
Unfortunately the process is still a best first pass and it will need additional validation or scrutiny.
The Add-In has the advantage of being able to output the probability match as a column, so you can just set a lower probability and then eyeball which is the best match.

Power Query, I think they suggest setting a lower match figure and then selecting that you only want one result and it will return the match with the higher result.
It also lets you build a helper table to include in the Fuzzy Lookup so you can put your identified matches in that helper table, to help whittle down the unmatched.

If both lists have something like and ABN (australian business registration no) or country equivalent or address or email address you can see if that works better in a fuzzy match