If every budget priority can receive a five-star rating, the answers may not tell you how respondents would divide a limited resource. A budget allocation question asks them to distribute a fixed total across competing categories.
This example gives respondents 100 points to divide among four priorities. The form must accept a deliberate zero, distinguish it from a blank answer, and reject a total of 99 or 101. It must also save the final category values, not just a total displayed on screen.
Use ChatGPT or Claude connected to Walla to request the interaction as a Custom Field. The guide below focuses on question wording, exact validation, saved-answer checks, and a small reporting example. All categories, allocations, and records are illustrative; this question records preferences rather than authorizing expenditure.
1. Write a question with a clear allocation target
Question and instructions you can adapt
How should we allocate next quarter’s discretionary budget?
Distribute exactly 100 budget points across the four areas below. Use whole numbers from 0 to 100. Enter a value for every area; enter 0 if you would allocate no points to it. Your total must equal 100 before you can submit.
Category | Meaning in this example |
|---|---|
Customer Research | Learning about customer needs and testing assumptions. |
Product Development | Building and improving the product. |
Marketing | Reaching prospective customers and communicating the offer. |
Training | Developing the team’s skills. |
These points express a preferred allocation of the hypothetical budget. They are not currency amounts, actual spending records, or an approved budget. Replace the categories and definitions with ones your respondents can distinguish.
Keep the same category wording for every respondent in a reporting period. If the question concerns current spending instead of a preferred future allocation, change the wording explicitly.
2. Define what counts as a valid answer
All four categories must be answered. A blank is incomplete; a typed 0 is a valid category value.
Each value must be a whole number between 0 and 100, inclusive.
The sum of the four entered values must equal exactly 100.
A negative or fractional value is invalid even when the displayed sum happens to be 100.
Show points allocated and points remaining, or the amount over 100. Do not silently rebalance the last category or scale the answers to force a total of 100.
If a valid answer becomes invalid after an edit or Reset, it must not submit the earlier valid values.
Start the inputs blank rather than prefilled with zeros. The display may count unfilled categories as contributing nothing to its running total, but the validation must still treat those categories as unanswered.
3. Copy this build request into ChatGPT or Claude
Connect Walla first and replace [workspace name or ID]. The English help center explains the connection and the Custom Field review process. Walla connection guide / Custom Fields guide
Use Walla to build a reusable Custom Field named “Budget Allocation 100” in [workspace name or ID]. Check for a suitable existing field first. If reusing one would not preserve the intended four-category behavior and saved-value structure, explain the difference. Do not change existing production form placements.
Ask: “How should we allocate next quarter’s discretionary budget?” Show these four labeled numeric inputs: Customer Research, Product Development, Marketing, and Training. Explain that respondents must distribute exactly 100 budget points, use whole numbers from 0 to 100, and enter 0 explicitly when allocating nothing to a category.
Start every input blank. Treat blank and zero differently. Require all four values, require each to be an integer from 0 to 100 inclusive, and require their sum to equal exactly 100. Do not coerce a blank into a completed zero answer, round a fractional value into validity, automatically rebalance another category, or normalize the values to 100.
Show a running allocated total and either points remaining or points over the limit. If an input is missing, say that every category needs an answer even if the other entries already total 100. Give a clear inline error for invalid category values and totals.
Use Walla’s supported Custom Field value and validation contract so the full form cannot accept an invalid allocation as a valid required answer. When a previously valid allocation is changed to an invalid one, clear its stale submitted value. Reset must clear all four inputs and return the answer to incomplete.
Request this saved answer structure: questionVersion equal to budget-v1, researchPoints, productPoints, marketingPoints, trainingPoints, and totalPoints. All point properties must be numeric integers, including legitimate zeros. Save the final category values and their computed total, not only a completion flag or display label. Show the actual output schema you implement and explain any difference.
Keep category labels and errors readable on mobile. Make the number inputs and Reset action usable with a keyboard, with visible focus. Do not collect personal information in this field or implement a spending approval workflow.
Build the review preview. Summarize the validation and saved-value behavior, identify unmet requirements, and wait for my approval before registering the build. Do not publish a form.
These are specifications to verify against the generated field. A plausible preview does not establish that the full form blocks invalid answers or that the category values are stored correctly.
4. Test the exact-total rule and the category rules together
Use the category order Research, Product, Marketing, Training when reading the examples below.
Entered values | Calculated total | Expected result |
|---|---|---|
40, 30, 20, 10 | 100 | Valid: all four values are entered and the total is 100. |
100, 0, 0, 0 | 100 | Valid: preserve all three explicit zeros. |
0, 0, 0, 100 | 100 | Valid: the fourth category may receive all points. |
40, 30, 20, 9 | 99 | Invalid: 1 point remains. |
40, 30, 20, 11 | 101 | Invalid: 1 point over the limit. |
−1, 51, 25, 25 | 100 | Invalid: negative value, despite the total of 100. |
25.5, 24.5, 25, 25 | 100 | Invalid: fractional values, despite the total of 100. |
Additional case | Expected result |
|---|---|
100, blank, blank, blank | Incomplete: the three blanks are not explicit zeros. |
101, 0, 0, 0 | Invalid: category above 100 and overall total above 100. |
0, 0, 0, 0 | Invalid total of 0, although every individual value is in range. |
Reset after a valid answer | Incomplete: no earlier allocation can pass through. |
Change 40, 30, 20, 10 to 40, 30, 20, 9 | Invalid immediately; do not submit the older total-100 answer. |
Copy this validation-review request
Review Budget Allocation 100 using the full validation rules, not only the sum. Valid examples are 40/30/20/10 and 100/0/0/0. Reject 40/30/20/9, 40/30/20/11, negative values, and fractions. Also reject 100 with three blanks, even though the running numeric total is 100.
Test changing a valid answer to an invalid one and pressing Reset. Confirm that the previous valid payload cannot be submitted. A deliberate numeric zero must remain valid and must not be treated as missing.
Fix any mismatch without silently redistributing or rounding the respondent’s answers. Preserve the agreed property names and types unless a necessary change is explained. Provide a revised preview, report only tests actually performed, and wait for approval before registration or publication.
5. Approve the field and collect fictional test responses
Approve the reviewed build, then add its registered version to a separate draft form. Field registration and publishing the form are distinct steps.
I approve the reviewed Budget Allocation 100 build. Register that version and create a new draft form named “Budget Allocation Validation Test” in [workspace name or ID].
Add a required short-text question called “Test case ID”, followed by a required question using the four-category budget allocation field we approved. Reuse that registered version rather than creating another copy.
Read back the form ID, allocation question ID, registered version, required settings, and actual saved-value schema. Keep the form unpublished. Explain any placement step that cannot be completed with the available tools.
If you finish placement in the editor, use the Custom Fields tab while adding the question. Check both the required setting and the allocation field’s own validation.
Publish only “Budget Allocation Validation Test” with form ID [test form ID] and return its respondent link for fictional submissions. Do not publish or modify another form.
BA-001: enter 40, 30, 20, 10 and submit the whole form.
BA-002: enter 100, 0, 0, 0 and submit. This checks that explicit zeros survive storage.
BA-003: first enter a valid allocation, then change it to 20, 30, 25, 25 before submitting. This checks that the final edited distribution is saved.
BA-004: enter a valid answer, lower its total to 99, and attempt to submit. It must not be accepted as a valid allocation.
BA-005: enter 100 and leave three categories blank. Attempt to submit; the required incomplete answer must not pass.
Repeat the important cases on a phone, navigate away from the question and back, and verify that the current values and validation messages stay consistent. These cases are instructions for testing your build, not a report of an already executed test.
6. Inspect the values stored in Walla
Find the valid test records in the form’s Response Sheet and inspect each response’s allocation answer. Search, response details, and CSV or XLSX download can help you check the collected data. Read the English Response Sheet guide
The expected table below uses the answer design requested here. Inspect the actual registered output schema before interpreting the stored value.
Saved property | BA-001 | BA-002 | BA-003 |
|---|---|---|---|
questionVersion | budget-v1 | budget-v1 | budget-v1 |
researchPoints | 40 | 100 | 20 |
productPoints | 30 | 0 | 30 |
marketingPoints | 20 | 0 | 25 |
trainingPoints | 10 | 0 | 25 |
totalPoints | 100 | 100 | 100 |
Copy this saved-response verification request
In form [test form ID], inspect the stored Budget Allocation 100 answers for BA-001, BA-002, and BA-003 using the actual output schema. Do not use preview state or default properties.
Expect BA-001 to contain researchPoints 40, productPoints 30, marketingPoints 20, trainingPoints 10, totalPoints 100. BA-002 should contain 100, 0, 0, 0 and totalPoints 100, with numeric zeros preserved. BA-003 should contain the final edited values 20, 30, 25, 25 and totalPoints 100. All three use questionVersion budget-v1.
Where returned types are visible, check that the point values are numeric integers. Recompute the sum from the stored category values rather than trusting totalPoints alone. Check whether a completed response was accepted for invalid BA-004 or incomplete BA-005.
Report response identifiers and any missing, stale, coerced, or mismatched value. Do not edit or delete responses. If a value cannot be inspected with the available tools, leave it unverified and explain what I should check in the response details or export.
A total of 100 does not prove the distribution is correct. Compare every category with the final input, especially BA-003 after editing. Also check that blank values were not replaced with zeros.
Inspect the actual downloaded representation before building spreadsheet formulas. A structured answer may be serialized in a cell rather than automatically expanded into category columns. If you transform the data, keep its response ID and question version.
7. Summarize only valid, comparable allocations
For a report, define a valid response as four present numeric integers in range whose sum is 100, using the same question version and category meanings. Separate test records, incomplete responses, and invalid data before calculating category averages. Do not repair an invalid total by silently scaling its values.
Category | Average points in three fictional valid responses |
|---|---|
Customer Research | 53.33 |
Product Development | 20.00 |
Marketing | 15.00 |
Training | 11.67 |
This illustrative summary uses BA-001, BA-002, and BA-003 only. Its denominator is three valid responses. Zero is included as a real allocation, not removed as a missing value. The averages describe this small fictional set and are not evidence about your organization’s priorities.
State the question wording, category definitions, version, valid response count, and exclusion rules alongside a real report. Keep preferred allocations distinct from actual spending or approved budgets.
8. Keep the question and its validation aligned
When you change category meanings, the target total, or the allowed precision, revise the instructions and tests together. Record a new question version when the change affects how answers should be interpreted, and check which registered field version is placed in each form.
Before collecting real responses, confirm that invalid totals are blocked in the full form, deliberate zeros are saved, and final edits replace earlier values. Then repeat the storage check after a revision that affects the answer structure or validation.
Use the English Walla connection guide and Custom Fields guide to build and refine the question through conversation.
Open Walla to build your budget allocation question








