About this exercise
A 300-row dataset on one sheet, 3 answer cells, about 8 minutes. Graded in your browser: the value and the formula behind it, with feedback that names the fix.
How to solve it
- G2
=UNIQUE(TRIM(C2:C301)) - J2
=ROWS(UNIQUE(TRIM(C2:C301))) - F2
=COUNTIF($C$2:$C$301,TRIM(C2)&"*")
UNIQUE returns every distinct value of a range once and spills the result into the cells below, so one formula in G2 produces the whole customer list, in the order the names first appear. It ignores letter case — "oakridge marine" and "Oakridge Marine" are one customer — but not spaces, which is why TRIM inside the range is what folds "Ridgeway Analytics " into "Ridgeway Analytics": 58 customers instead of 70. ROWS around the same UNIQUE turns the list into a count, and COUNTIF with the pattern TRIM(C2)&"*" answers the per-order question: how many rows belong to this customer, stray spaces included.
Pitfall 1: Remove Duplicates (Data tab) is destructive — it deletes rows, keeps the first occurrence and has no live link to the source; UNIQUE is a formula, so a new order appears in the list the moment it is typed. Both ignore letter case; neither trims spaces. Pitfall 2: #SPILL! appears when anything sits in the cells the array needs (here G3:G59); clear the path or move the anchor. Pitfall 3: COUNTIF compares like Remove Duplicates — case-insensitive, space-sensitive — and needs a range as its first argument, so it cannot TRIM the column: =COUNTIF($C$2:$C$301,C2) misses the trailing-space copies and is wrong on 78 rows here. The wildcard pattern TRIM(C2)&"*" fixes that, but it would over-count if one customer's name were the start of another's ("Summit Labs" and "Summit Labs Group" — none here); the exact version is =SUMPRODUCT(--(TRIM($C$2:$C$301)=TRIM(C2))), which trims the whole column in memory and is slower on long lists. Alternatives: =SORT(UNIQUE(TRIM(C2:C301))) for an alphabetical list; =UNIQUE(TRIM(C2:C301),,TRUE) lists only the 8 one-off customers; the pre-365 distinct count =SUMPRODUCT(1/COUNTIF(C2:C301,C2:C301)) cannot trim and returns 70. Version note: UNIQUE, SORT and spilling need Excel 365 or Excel 2021; in Excel 2019 put =TRIM(C2) in a helper column, paste it as values and run Data → Remove Duplicates on a copy. Desktop Excel can refer to the whole spill as G2# (=COUNTA(G2#)); this in-browser grid does not support # yet.
Other ways that work
- =COUNTA(G2:G301) — counts the spilled list for J2 and grows with it
- =SUMPRODUCT(--(TRIM($C$2:$C$301)=TRIM(C2))) — the exact per-order count: compares every trimmed name with this one, no wildcard needed
- =SORT(UNIQUE(TRIM(C2:C301))) — the same names in alphabetical order (a valid list, but this check compares the first-appearance order)
- =SUMPRODUCT(1/COUNTIF(C2:C301,C2:C301)) — pre-365 distinct count; returns 70 because it cannot TRIM
Common mistakes
Writing =UNIQUE(C2:C301) without TRIM: the list looks right at a glance but holds 70 entries instead of 58 — from G16 on, every name is one row off because "Ridgeway Analytics " is listed again.
How we grade
Any formula that produces the right result is accepted, unless the task names a function: then another function earns half credit. We compare values with a small tolerance and ignore case and spaces in text. A correct number typed by hand instead of a formula earns half credit. Multi-cell tasks earn full credit at 100% of cells, half credit from 50%. Where it matters, we also change an input cell on a copy of your sheet and check that your formula follows.
Answers computed with Excel 365 rules and checked in our grading engine on November 16, 2026. Author: GridBelt team.