Manual or contracted survey routes for marketplace research, compared on evidence fields

Marketplace seller software survey tools should be chosen only after the population, questions and data route are fixed. This desk comparison names no vendor because no exact current edition has been tested against the required privacy, accessibility, security and export evidence. A manual route remains a valid baseline.

The common study asks verified staff at an England merchant about one recent marketplace listing, stock or order exception. It collects no live customer order, payment or product-safety record.

Compare routes, not marketing labels

Research date: 6 September 2026. Method: Four delivery routes were compared on identical fields. Scope: One bounded staff survey for a named seller account. Exclusions: Prices, vendor rankings, assumed features, customer panels and general ecommerce surveys. Hands-on status: None. Conflicts: No provider funded the work. Ranking: Non-ranked, with no winner.

Route	Instrument and contact	Data and access	Accessibility evidence	Export and deletion
Manual, buyer-run	Versioned form or interview sheet sent through an approved channel	Buyer controls the file and permissions	Buyer tests each offered format	Buyer proves deletion and retains an approved analysis file
Buyer-built form	Merchant hosts the questionnaire and response store	Internal owners document hosting, logs and account access	Complete form and assistive-technology test required	Structured export, purge and restore tested by buyer
Contracted survey service	Exact service and edition must be named	Controller, processor, subprocessors, transfers and roles require evidence	Obtain a current conformance statement, test scope and unresolved barriers	Demonstrate complete export, account closure and deletion evidence
Research agency route	Agency administers a defined instrument	Contract must separate recruitment, fieldwork, transcription and analysis access	Recruitment and participation adjustments need an owner	Return or deletion of contact, response and recording data checked separately

No route receives credit for a field marked unknown. Costs and service thresholds stay blank until current contractual evidence exists.

Hold the instrument constant

The Government Analysis Function questionnaire design guidance says questions should be clear, avoid assumptions and be designed around how responses will be used. Keep the wording, order, response options and eligibility logic identical when comparing routes. Otherwise a mode change is mixed with a tool change.

Test the screener and questionnaire with synthetic records. Include refused or skipped items, an incomplete response, duplicate submission, withdrawal and correction. Do not ask for a customer's name, address, payment data or full order history merely because a form can accept it.

Inspect the data relationship

The ICO's current controller-processor contract guidance describes required processing details, documented instructions, confidentiality, security, subprocessors, rights assistance, return or deletion, and audits. The page is under review after DUAA changes. A privacy specialist must verify the live position and the actual supplier chain.

NCSC supplier-assurance questions cover network, cloud, access and data controls. Treat answers as supplier statements until evidence is inspected for the contracted service. A security badge does not prove safe configuration.

GOV.UK guidance on managing research data recommends collecting the minimum contact information, limiting access and deleting details when no longer needed. Its setting is government user research, but those practical controls suit this comparison.

Choose a route only after it passes the buyer's synthetic test and every independent gate. If privacy roles, accessible participation, security evidence or a usable export remains unknown, the decision is hold, even when the interface appears convenient.