Part 1
Description
A message is sent repeatedly, but noise garbles it, so every copy is a bit different. The input has one copy per line, all eight letters long. Reading down each column of letters, the most frequent letter is the right one. Part 1 asks for the corrected message.
Solution
I pasted the 546 lines into A5:A550. The first step is to split every line into its eight letters, one per column. SEQUENCE(,8) makes the numbers 1 to 8 across a row, and MID takes that character of every message:
C5: =LET(a,A5:A550,MID(a,SEQUENCE(,8),1)) ← fills C5:J550
Now each column holds all the copies of one letter of the message. For one column, COUNTIF(C5:C550,C5:C550) tells, for every cell, how many times its letter appears in the column. Comparing that with the largest count gives TRUE for the most common letter. MATCH finds the first TRUE and INDEX returns the letter:
=LET(c,C5:C550,n,COUNTIF(c,c),INDEX(c,MATCH(MAX(n),n,0)))
This is an array formula (in older versions of Excel, enter it with Ctrl+Shift+Enter). I filled it to the right, up to J4. The message is the eight letters joined together:
B2: =CONCAT(C4:J4)
I tried the puzzle’s example (16 short lines) and it gives the message from the puzzle text.
Part 2
Description
The repetition code turns out to work differently: the original message is made of the least common letter in each column.
Solution
That is the same formula with MIN in place of MAX. I put it in L4 and filled it to the right, up to S4. Each cell still looks at its own column (L4 looks at column C, M4 at column D, and so on):
L4: =INDEX(C5:C550,MATCH(TRUE,COUNTIF(C5:C550,C5:C550)=MIN(COUNTIF(C5:C550,C5:C550)),0))
D2: =CONCAT(L4:S4)
A caveat. If two letters tie for the most (or least) common, these formulas quietly pick the first one in the list. In my input every column has a clear winner, so it doesn’t matter, but the margin is small, so it is worth checking that your own columns have no ties.

