The most common Excel test is a short multiple-choice screening sent as a link before anyone talks to you: 10 to 30 questions, 10 to 30 minutes, one question on the screen at a time, and a percentage or percentile at the end. The questions are theory and small scenarios, not a live spreadsheet. Here is what the vendors publish about the format, what candidates report, and how to be ready in a few hours.
Vendors describe the content as features, functions and formulas, creating and formatting spreadsheets, and managing data. The question types repeat across vendors: a small spreadsheet screenshot with "which formula produces this result", a conceptual question about a feature, and a few on shortcuts or menus.
SUM, AVERAGE, COUNT and COUNTA on a screenshot range; IF with a simple condition; COUNTIF and SUMIF, the two functions TestGorilla names on its test page along with VLOOKUP.
What VLOOKUP returns for a given table, exact versus approximate match, why a lookup shows #N/A.
What happens to a formula when it is copied down, where the $ goes, how to point at another sheet.
Number and date formats, wrap text, merge versus center across selection, freeze panes, what a red triangle in a cell means (a Mettl sample question).
Multi-level sort, what a filter hides, how a table's total row behaves.
Pick the right chart type, read a chart from a screenshot, switch rows and columns.
Read a PivotTable from a screenshot: what its rows, columns and grand totals show.
iMocha's basic test lists keyboard shortcuts as a topic; Mettl asks for the AutoSum shortcut (Alt + =) and how to freeze a pane.
You cannot look anything up during the test, so the preparation is recall under a clock: name the function from the result, and know what the common functions do without opening Excel.
Take the free Level 1 test once, cold. It uses the same rhythm (a clock on every task, no going back) and its five sub-scores show which of the six areas to work on.
Do the basics exercises: SUM, AVERAGE, MIN, MAX, COUNT, COUNTA and COUNTBLANK on real data, until the difference between COUNT and COUNTA is automatic.
Do the two absolute-reference exercises. "What happens when this formula is copied" is on every screening.
Review the formatting mechanics that screenings ask as theory: number formats, wrap text, merge, freeze panes.
Read the cheat sheet entries for IF, COUNTIF, SUMIF and VLOOKUP and work one exercise on each.
Retake the test the day before. Aim to finish with time to spare: the clock, not the content, is what catches people.
Level 1 is modeled on this family and on the discontinued Indeed test: 15 tasks, 90 seconds each, one task per screen, no going back.
Candidates who know the material still lose questions by re-reading a screenshot. Read the question first, then look at the sheet for exactly what it asks.
COUNT counts numbers only; COUNTA counts anything. SUMIF takes range, criteria, sum range; SUMIFS starts with the sum range. Screenshot questions are built around these swaps.
"This formula is copied from C2 to C10; what does C10 show?" tests whether you notice an unlocked reference to a rate cell.
Multi-select questions score only when the whole set is right. Check every option; a single missed one is a zero.
Screenings also ask about menus and concepts: where freeze panes lives, what a filter does to hidden rows, what #N/A means. Theory you never read is theory you cannot recall.
Usually not. TestGorilla, iMocha, Mettl, Adaface's main test and Criteria's standard test are multiple choice. Some vendors add a hands-on task (Adaface's Workbook test, Criteria's Interactive version with three simulations); the recruiter's invitation names the test.
Usually not. TestGorilla shows one question per screen with no return, and the legacy Indeed test timed each question separately. Treat every question as final.
Vendors do not publish pass marks. TestGorilla reports a percentile against all test takers, Criteria a rating from Beginner to Advanced, Adaface suggests a threshold around 55% that employers can change. Prep sites commonly recommend aiming for 70% or better.
Only if the employer sends a new invitation. Each attempt draws different questions from the bank, so a retake is a new test, not the same one.
The multiple-choice tests work on a phone, but screenshot questions are hard to read on a small screen and some tests disable switching tabs. Use a laptop with a stable connection.
SHL-type, eSkill, Criteria Interactive, Testlify, Vervoe
25–35 minutes 20–30 tasks
A mock Excel window with a task under it. One attempt per task, graded on the end result, some shortcuts switched off.
Чего ждать от этого тестаMicrosoft Office Specialist exams MO-210 and MO-211 (Certiport)
50 minutes 23–37 tasks
Several projects in live Excel, each a workbook with 4–7 tasks. Scored 1–1000, 700 to pass.
Чего ждать от этого тестаWritten by the hiring team; sometimes run live over screen share
20–60 minutes One workbook
Raw data on two or three sheets, a list of questions, your own Excel. Join, aggregate, clean, summarize, explain.
Чего ждать от этого тестаБесплатно, без регистрации и без установки.
Facts checked on November 15, 2026 against the test-library pages of TestGorilla, iMocha, Adaface, Mettl, Testlify and Criteria and against TestGorilla's scoring documentation. Question counts for TestGorilla's Excel tests are the range the company publishes for its multiple-choice tests; the exact count per test is not published.
Excel — товарный знак группы компаний Microsoft. Indeed, TestGorilla, SHL, iMocha, Adaface, Criteria и другие товарные знаки принадлежат их владельцам. GridBelt — независимый проект: он не связан ни с одной из этих компаний, не одобрен и не спонсируется ими.