Sources reviewed November 15, 2026.
An overseas buyer enters `1,234` in a quantity field. A salesperson sees one thousand two hundred and thirty-four. A buyer using a decimal comma may have intended one point two three four.
That difference can reach a quotation, a cut length or an order minimum. The safest place to resolve it is the inquiry record, while the buyer can still confirm the intended quantity and unit.
An English-language website can receive visitors who use several numeric conventions. Page language, shipping country and numeric locale are useful context. Each field still needs a clear interpretation and a reviewable value before the business acts on it.
Formatting and parsing are separate operations
Formatting starts with a known value and produces display text for a chosen locale. Parsing starts with text and attempts to recover a value under a declared set of rules.
For the fictional value 1234.5, formatted to two fractional digits, common locale outputs include the following.
| Locale | Display text | Grouping and decimal convention in this example |
|---|---|---|
| en-US | 1,234.50 | Comma grouping, decimal point |
| de-DE | 1.234,50 | Point grouping, decimal comma |
| fr-FR | 1 234,50 | Narrow no-break space grouping, decimal comma |
The French grouping character in this example is U+202F. A visual inspection alone can miss the difference between it and an ordinary space.
Unicode CLDR describes locale-specific number symbols and patterns. Its parsing guidance discusses ambiguous inputs and recommends echoing the interpreted number back in the user's locale for confirmation. Lenient parsing and confirmation are separate design considerations. Any application policy that rejects ambiguous input should be specified and tested as that application's policy. Unicode CLDR number specification, version 79
Reproduce the display before designing the input
This small JavaScript example formats a known number. Run it in a modern JavaScript environment with the relevant locale data. ECMA-402 specifies the locale and digit options used by this API. ECMA-402 NumberFormat specification
```js
const value = 1234.5;
const locales = ["en-US", "de-DE", "fr-FR"];
for (const locale of locales) {
const formatter = new Intl.NumberFormat(locale, {
minimumFractionDigits: 2,
maximumFractionDigits: 2
});
console.log(JSON.stringify({
locale,
display: formatter.format(value)
}));
}
```
The input value is already known here. Building the reverse operation requires an input contract, locale handling and ambiguity tests. Deleting every comma or changing every comma into a point encodes a policy that can silently change a buyer's meaning.
Use representative test cases before choosing that policy. Under an explicitly declared en-US interpretation, `1,234` can represent 1234. Under an explicitly declared de-DE interpretation, it can represent 1.234. A country dropdown alone may leave ambiguity because the buyer may work with specifications written for another market.
Give each numeric field a business meaning
A product inquiry often contains several kinds of number-like text.
| Field | Meaning to preserve | Example review question |
|---|---|---|
| Order quantity | Count and purchasing unit | Does 120 mean pieces, packs or metres? |
| Required length | Measurement and unit | Is 1.25 expressed in metres or millimetres? |
| Part identifier | Exact identifier text | Are the leading zeros part of the code? |
| Telephone number | Contact identifier | Does the international prefix remain intact? |
| Target price | Currency and commercial basis | Is the amount per piece, per pack or per shipment? |
| Material fraction | A technical expression | Does 1/2 denote a nominal size or a calculated quantity? |
A field's declared type governs its handling. A part number such as `001234` needs its leading zeros. A telephone number needs its contact syntax. A measurement needs a unit and often an allowed range. Record those rules in the form specification and in the handoff instructions used by sales.
Keep unit conversion as a separate, visible step. Once the buyer has confirmed a length of 1.25 metres, conversion to 1250 millimetres is straightforward. The original entry, confirmed interpretation and converted value should remain traceable. An earlier guide explains how to preserve product facts across market-specific content. Xindar guide to localization and product facts
Preserve both the entry and its interpretation
A useful inquiry record stores the raw text alongside the proposed numeric meaning. The following fictional JSON object illustrates one possible application contract.
```json
{
"raw_text": "1.234,50",
"declared_locale": "de-DE",
"quantity": {
"value": "1234.50",
"unit": "mm"
},
"parse_state": "awaiting_buyer_confirmation"
}
```
The decimal string is a deliberate choice for this example contract. Its scale, permitted precision, validation and downstream conversion need specification. Other systems may use a suitable decimal type with an equally explicit contract.
The record supports a readable confirmation such as “You entered 1.234,50 mm. We interpreted this as 1234.50 millimetres.” Give the buyer an edit action and retain the confirmation state. Sales can then distinguish a proposed interpretation from a quantity the buyer has approved.
When an input is ambiguous, preserve it for review. A workflow might ask the buyer to select a numeric convention, enter the value again with an example, or confirm an expanded reading. Choose the policy according to the field's commercial consequence and test it with intended users.
JSON syntax does not carry a numeric locale
JSON's number grammar uses a full stop for a fractional part. Grouping punctuation belongs in a string representation or another explicitly defined field. RFC 8259 provides the grammar that a receiving system can validate. RFC 8259, section 6
These examples illustrate distinct data structures.
| JSON text | Interpretation |
|---|---|
| `1234.5` | A JSON number |
| `"1,234"` | A string whose numeric interpretation needs context |
| `[1,234]` | An array containing two numbers |
| `{"length":1,234}` | Invalid object syntax |
A file can pass syntax validation while still containing an ambiguous numeric string. Add semantic tests for the application contract, including its locale, unit and confirmation rules.
At an integration boundary, ask the receiving developer to show exactly which fields contain display text and which contain machine values. Retain an example request and its interpreted record in the acceptance evidence. A successful HTTP response alone gives little detail about how the value was understood.
CSV quoting solves the field boundary
CSV often carries inquiry data into a spreadsheet or a sales workflow. A comma inside a field can be protected with double quotes under the common format documented by RFC 4180, an Informational RFC. Embedded double quotes have their own escaping rule. RFC 4180, section 2
Consider this fictional record.
```csv
raw_text,declared_locale,unit
"1,234",de-DE,mm
```
A conforming reader can preserve `1,234` as one field. Deciding whether that string means a decimal or a grouped integer still requires the declared interpretation. Spreadsheet import settings can also affect how a preserved string becomes a displayed cell value.
Test the complete path. Export the record, open it through the actual receiving workflow, and compare the original text, intended value, unit and identifier characters. Keep the raw export so the comparison can be repeated after software or regional settings change.
Build a small acceptance set around costly mistakes
Use fictional values during design. Select real customer examples only through the appropriate permission and data-handling process.
| Test case | Expected evidence |
|---|---|
| Same `1,234` entry under two declared locales | Two explicit interpretations, each available for review |
| French narrow no-break space | Exact text survives capture and export |
| Identifier `001234` | Leading zeros remain intact |
| Quantity with an unselected unit | Record remains in the defined review state |
| A valid JSON numeric string | Syntax success and separate semantic validation are visible |
| Quoted CSV numeric text | One field survives export and actual import |
| Buyer edits a proposed quantity | Latest approved value and prior raw entry can be traced |
Assign an owner to the interpretation policy, a technical owner to parsing and storage, and a sales owner to the confirmation step. The acceptance record should show the actual form, saved record, exported file and handoff received by the person preparing the quotation.
Xindar's current export-marketing scope includes website inquiry paths and human handoff. A project brief can specify these input and confirmation requirements, then require a demonstrated receiving record before acceptance. The form's implemented behaviour remains a project-specific deliverable. Xindar export-marketing services
Questions from overseas sales teams
Should an English website force en-US number input?
A site can choose a documented convention and make it visible through labels, examples and confirmation. Evaluate the choice with the buyers who will use the form. Preserve their original entry and provide a correction path.
Can we infer locale from the buyer's country?
Country is useful context. Buyers can use different language and numeric settings, especially when copying an engineering specification. Explicit field rules and a readable confirmation reduce reliance on that inference.
What should we test after a spreadsheet import?
Check the field boundary, interpreted value, unit, decimal precision and identifier characters. Compare the imported record with the saved inquiry and the original export bytes.
Where does this affect GEO work?
AI visibility can bring a buyer to an English technical page. The inquiry that follows needs to preserve what the buyer intends to purchase. Track visibility, visits, submitted inquiries and commercially confirmed requirements as separate stages of that path.
