You can create a Yes or No on Excel without writing a single spreadsheet formula by opening the Yes or No Generator in a browser tab, typing an optional question under 1,000 UTF-16 code units, and pressing the generate button. The tool produces a casual 50/50 answer by drawing one Uint32 value from the browser's Crypto.getRandomValues API and mapping even values to Yes and odd values to No. Because the entire workflow runs locally, your question, the resulting answer, and the running history never leave the current tab and are not uploaded to any account, network service, or remote storage. The generator also refuses to silently fall back to Math.random when Web Crypto is missing or blocked, which keeps the displayed methodology consistent with the actual random source. For readers who searched "create yes or no on Excel" because they want a quick binary decision while working on a spreadsheet rather than a permanent spreadsheet feature, this approach delivers an answer without configuring data validation, building dropdowns, or maintaining random-value formulas inside the workbook. The tool also keeps a newest-first history of the most recent 20 draws, so you can compare results across repeated prompts without losing context.
Why Skip Excel for a Casual Yes or No
The phrase "create yes or no on Excel" can mean two different things, and conflating them is the most common reason people end up over-engineering a simple decision. One meaning is building a Yes/No field that lives inside the spreadsheet: a dropdown column, a checkbox list, or a formula-driven cell. The other is producing a Yes/No answer while you are working in Excel, where the spreadsheet is just the surrounding context. Excel-native approaches such as CHOOSE(RANDBETWEEN(1,2)), data-validation lists, and IF(RAND()<0.5,"Yes","No") each have trade-offs that matter when the goal is a one-off prompt. They require editing the workbook, they regenerate on every recalculation, and they rely on Excel's built-in RNG instead of a cryptographically strong source.
For a casual decision like choosing a lunch spot, picking which column to clean up first, or deciding whether to take a coffee break, those setups are heavier than the question warrants. A browser-based Yes or No Generator is faster because it needs no workbook edits, no named ranges, and no copy-paste of formulas. It also lives in a separate tab, so you can keep your spreadsheet visible while the prompt and history sit beside it. If you want an unbiased 50/50 answer with no spreadsheet footprint, skipping the formula path is reasonable.
How the Yes or No Generator Produces a 50/50 Answer
The generator's randomness comes from a single source: the browser's Crypto.getRandomValues method, which is part of the W3C Web Cryptography specification. When you press Generate, the tool asks the browser to fill a one-element Uint32Array, a typed array that holds one unsigned 32-bit integer. That integer can take any value from 0 through 4,294,967,295, which spans the entire 2^32 range.
For two outcomes, the mapper does something elegant. It computes the largest multiple of 2 that does not exceed 2^32. Because 2^32 divides evenly by 2, that limit equals the full Uint32 range. Every valid Uint32 value is therefore accepted on the first draw, so no value is rejected and no retry is needed during ordinary binary use. The mapper then applies modulo 2: even values become Yes, and odd values become No.
This split is perfectly balanced. Exactly half of the source values, which is 2^31 or 2,147,483,648 distinct values, map to Yes, and the other half map to No. The contract explicitly verifies the rejection path at a three-outcome boundary where division is not even, but for Yes and No the boundary is irrelevant because every draw produces a usable answer.
| Property | Value |
|---|---|
| Random source | Browser Crypto.getRandomValues (Uint32) |
| Source range | 0 to 4,294,967,295 (full 2^32 range) |
| Outcomes | 2 (Yes, No) |
| Acceptance limit | 2^32 (no rejection needed) |
| Source values mapped to Yes | 2^31 (every even Uint32) |
| Source values mapped to No | 2^31 (every odd Uint32) |
| Retry behavior during ordinary use | None — every draw is accepted |
| Fallback if Web Crypto fails | None — explicit error, no Math.random |
Generate Yes or No in Three Steps
- Open the Yes or No Generator in a browser tab. The interface shows an optional question field, a Generate button, and an empty history panel.
- Type your low-stakes question into the field if you want a label attached to the answer. Keep it under 1,000 UTF-16 code units, or leave the field empty to skip the label entirely.
- Click Generate. The current answer appears immediately. If Web Crypto is unavailable, an explicit error message shows and no answer is recorded. To compare multiple draws, click Generate again; each result becomes the newest entry in the history list.
Reading the Newest-First History
The history panel holds at most 20 entries. The newest answer sits at the top, so you can scan recent draws without scrolling. After the twenty-first generation, the oldest entry is removed and a cumulative message tells you exactly how many earlier answers have been discarded, for example "1 earlier answer discarded" after 21 total generations.
The history is descriptive only. It does not influence probability, it does not rebalance after a streak, and it does not prevent repeats. Independent 50/50 results can naturally form runs of Yes or No, and the tool surfaces that fact rather than hiding it. If you see five Yeses in a row, the next draw still has a 50/50 split because there is no streak compensation.
To clear the history, use the clear-history control. Resetting the history also wipes the cumulative discarded count and the currently displayed answer, so the next Generate starts from a clean panel.
What the Question Field Does and Doesn't Do
The question field is purely a local label. It does not change probability, weight alternatives, or feed into the random mapping. The mapping always uses one fresh Uint32 from Web Crypto, regardless of what the question says. Two identical questions can therefore produce different answers, and two different questions can produce the same answer.
The field accepts up to 1,000 UTF-16 code units. The boundary is inclusive: a question at exactly 1,000 code units is accepted, and one code unit beyond fails without truncation or generation. Empty text is valid because the question is optional.
Editing a question after an answer is displayed clears the current answer and any prior error message. This prevents the previous answer from being mislabeled as belonging to the new wording. Editing does not clear the history, so older entries remain for comparison, but the prominent current answer is removed.
The text is stored exactly as you typed it. Trimming is used only to decide whether the history row shows the label "No question" or your actual wording. Internal whitespace, capitalization, and punctuation are preserved verbatim in the stored entry.
Excel-Native Alternatives When You Need a Spreadsheet Feature
Sometimes you actually need the Yes or No inside the spreadsheet, for example a dropdown list, a checkbox column, or a formula-driven answer cell. In those cases, the browser tool is the wrong fit and you should use Excel's built-in features instead. Three approaches cover most needs.
Data validation list (dropdown). Select the target cells, open Data, then Data Validation, set "Allow" to List, and type Yes,No in the Source box with a comma separating the values. Click OK. Users now see a dropdown arrow on each cell and can only pick Yes or No, which eliminates typos like "Yess" or "Nah." For step-by-step instructions, see the Create a Yes or No Drop-Down List in Excel guide.
IF + RAND formula. In a target cell, enter =IF(RAND()<0.5,"Yes","No") and press Enter. Excel recalculates the formula every time the workbook changes, producing a fresh 50/50 answer on each recalculation. Use Copy, then Paste Special, then Values to freeze a single answer if you want stability across edits.
CHOOSE + RANDBETWEEN. Enter =CHOOSE(RANDBETWEEN(1,2),"Yes","No"). RANDBETWEEN(1,2) returns either 1 or 2 with equal probability, and CHOOSE converts that to the literal text. This pattern is easier to extend to more than two outcomes later, and the Create a Yes or No Formula in Excel: 3 Approaches guide walks through the variations.
Each of these Excel-native approaches has trade-offs versus the browser tool. Excel's built-in RAND uses the platform's RNG rather than Web Crypto, so the source is not cryptographically strong. Formulas also recalculate on every workbook change, which can flip a Yes to a No when you did not expect it. For a casual one-off decision, the browser-based Yes or No Generator sidesteps both issues.
Limits and When Not to Use This Tool
The Yes or No Generator is a casual prompt, not a decision authority. The contract explicitly forbids using it for medical, legal, financial, safety-critical, emergency, security, consent, employment, eligibility, compliance, or other high-stakes decisions. Web Crypto's quality does not make a question appropriate, validate the available choices, understand consequences, or substitute for professional judgment.
The tool also does not learn preferences, guarantee alternating results, prevent streaks, prove fairness over a small sample, or predict future events. A mathematically balanced answer can still be unsuitable, unlucky, or harmful when applied to something with real-world consequences.
If you need a binary decision inside the spreadsheet itself, for example a Yes/No field that updates automatically or a dropdown that other users can pick from, this tool is not the right fit. In those cases, an Excel-native approach is what you want, and the guides linked above walk through the common patterns.
Sources Behind the Randomness
The generator's randomness is grounded in the W3C Web Cryptography specification, which defines getRandomValues and its in-place filling of integer typed arrays. MDN's Crypto.getRandomValues reference independently documents how browsers implement the method in practice. The mapper's rejection sampling at non-even boundaries is implemented and tested in code rather than claimed as a feature of the underlying browser source.