Get surveys

GET{{API_BASE_URL}}/ticketinghub/odata/v1/Surveys?$select=SubmissionInternalId,AnswerInternalId,SelectionOrdinal,QuestionText,AnswerDisplayValue,TicketInternalId

What visitors told you, joined back to who they are. One row per selected option — read the grain note before you count anything.

Because survey answers join to the ticket, any answer can be segmented by anything the ticket knows: article, event, statistics group, country, or how the ticket was issued.

Grain and key


Grain	one row per selected option (selection grain)
Natural key	`MandatorInternalId`, `AnswerInternalId`, `SelectionOrdinal`
Page size	500
Columns	32
Watermark	`LastRefreshedAt` — incremental loading possible

The columns most extracts start with

Column	Type	Notes
`SubmissionInternalId`	int	One completed survey. Count distinct for responses.
`AnswerInternalId`	int	One answer to one question. Count distinct for answers.
`SelectionOrdinal`	short	Which selected option this row represents.
`QuestionText`	string?	The question as the visitor saw it.
`QuestionType`	string?	The question's form type. Pass-through — no fixed list.
`IsMultiSelect`	bool	Whether several options were possible.
`IsRequired`	bool	Whether the question was mandatory.
`Answer`	string?	The raw lookup key.
`AnswerDisplayValue`	string?	The label the visitor saw. Report on this one.
`IsCustomAnswer`	bool	A free-text answer rather than a chosen option.
`AnswerSource`	string?	Where the answer came from. Pass-through.
`SubmittedAt`	datetimeoffset	When it was answered.
`Language`	string?	Which language they answered in.
`TicketInternalId`	int	Join key back to `Tickets`.
`LastRefreshedAt`	datetime	Watermark. No offset — do not convert.

Traps specific to this dataset

Selection grain is the trap. A multi-select answer is not one row with a comma-separated list; it is one row per selected option. Someone ticking four boxes produces four rows. Counting rows overcounts, sometimes by a lot.

You want Count
Responses (completed surveys) distinct SubmissionInternalId
Answers (question-level) distinct AnswerInternalId
Selections (option-level) rows
Group on AnswerDisplayValue, not Answer. Answer holds the raw lookup key, which is stable but meaningless on a chart. AnswerDisplayValue is the label the visitor actually read.
Free-text answers need separate handling. IsCustomAnswer rows contain whatever the visitor typed — do not group on them as if they were a controlled vocabulary.
No visitor context here. Name, contact data and article are deliberately not in this dataset — join Tickets.

Joining

Join to Tickets on MandatorInternalId, TicketInternalId.

Request

No request body.

$select(string): the columns to load.
$filter(string): an OData predicate. This dataset carries LastRefreshedAt, so it can be loaded incrementally.
$orderby(string): sort order.
$top(integer): maximum rows, capped at 1000.
$count(boolean): include the total count.
$skiptoken(string): server-generated paging token.

Response

@odata.context (string): the metadata URL describing this payload.
value (array): the page of selection rows.
@odata.nextLink (string, optional): the next page; absent on the last one.
@odata.count (integer, optional): present only when $count=true.

The grain will catch you out

One row is one selected option, not one answer. Somebody ticking four boxes on a multi-select question produces four rows.

You want	Count
Responses (completed surveys)	distinct `SubmissionInternalId`
Answers (question level)	distinct `AnswerInternalId`
Selections (option level)	rows

Group and report on AnswerDisplayValue, the label the visitor actually read — Answer holds the raw lookup key, which is stable but meaningless on a chart.

Notes

Page size is 500. @odata.count counts selections, so it is not your response count.

QuestionType, AnswerSource and Characteristic are pass-through columns — their value sets come from your own survey configuration rather than from a fixed list. See Datasets → Column values.

Authentication#auth

This endpoint does not require authentication.

Headers#headers

AuthorizationBearer {{ACCESS_TOKEN}}

Required. A client-credentials token carrying the scope ticketinghub-api and no subject. See Authentication.

Acceptapplication/json

Optional. JSON is returned by default.

Query parameters#query

$selectSubmissionInternalId,AnswerInternalId,SelectionOrdinal,QuestionText,AnswerDisplayValue,TicketInternalIdoptional

Comma-separated list of columns to load. Carry TicketInternalId so the answers can be joined back and segmented; report on AnswerDisplayValue rather than the raw Answer key.

$filterLastRefreshedAt gt 2026-01-01T00:00:00Zoptional

OData predicate. This dataset exposes LastRefreshedAt, so a watermark filter gives you an incremental load.

$orderbySubmittedAtoptional

Sort order. Note that it also changes the order the keyset walk runs in.

$top1000optional

Maximum number of rows. Capped at 1000; a higher value is rejected with 400. A smaller value lowers the page size but cannot raise it above 500.

$counttrueoptional

Include the total row count in @odata.count. Because of the selection grain this counts ticked options, not responses — count distinct SubmissionInternalId for that.

$skiptokenoptional

Server-generated keyset paging token. Note that this dataset has a three-part key, so the walk runs in alphabetical key order.

$skipoptional

Offset paging. Supported but discouraged — use @odata.nextLink instead.

Responses#responses

200OKGet surveys

You want	Count
Responses (completed surveys)	distinct `SubmissionInternalId`
Answers (question-level)	distinct `AnswerInternalId`
Selections (option-level)	rows