Similar text is useful for finding possible matches, but it is not an identity check. Two product descriptions can differ by a spelling error or by a detail that identifies a genuinely different item. Decide what evidence must agree before allowing a proposed match into the final table.

This workflow uses Power Query fuzzy matching in Microsoft 365 Excel for Windows. Build a small reference list with known identities and keep the original imported description alongside any proposed standard name. Avoid replacing the source description while the match is still being assessed.

Make the matching choices visible

Fuzzy matching applies to text columns in a merge. In the Merge dialog, select the relevant columns and enable Use fuzzy matching to perform the merge. Its options include a similarity threshold from zero to one, case handling and a maximum number of matches per input row. The default threshold is 0.80. A transformation table can supply deliberate From-to-To mappings.

Use a sample containing a known spelling variation, a clearly unrelated description and two distinct reference items whose names look alike. Do not assume which candidates will match before running the sample. Inspect the candidates actually returned, with the original description and an independent product code or other confirmed identifier available for comparison.

Accept evidence, not just a higher match count

Lowering the threshold may supply more candidates, but that does not establish that the additional associations are correct. Likewise, limiting the output to one candidate only limits what is returned; it does not make that candidate authoritative.

Record each ambiguous decision separately. If you create a transformation mapping, document why the two labels refer to the same thing. A mapping should express a confirmed equivalence rather than conceal uncertainty about a short or incomplete source description.

Before using the result, account for unmatched inputs and every accepted ambiguous pair. Keep rejected suggestions available for review rather than silently assigning them to the nearest-looking name. Repeat the sample when the reference list changes, because a newly added similar item can make an earlier matching assumption worth revisiting.

Sources: Microsoft: fuzzy merge options; Microsoft: fuzzy matching in Excel.