Segmentation typing tools and conjoint simulators: who owns them?

Who owns a segmentation typing tool?#

A segmentation typing tool is usually owned in layers: the client owns its segment solution and the classification rule fitted to its data, and the agency owns the methods and software used to produce them. The typing tool itself, a short question set with a scoring rule that assigns new respondents to segments, sits right on the line between the two.

Disputes start when a client switches agencies and asks for everything needed to keep typing its customers, or when an agency wants to reuse a segmentation approach for another client in the same category. Both questions are answered by the contract, not by research industry custom.

Conjoint work raises the same issue in a different form. The utilities belong to the study, but the simulator that turns them into share predictions is often a piece of agency software the client has never seen inside.

Client-specific model versus firm algorithms and software#

The cleanest way to see the split is component by component. Most agency agreements land close to the pattern below, but a broad assignment clause can move items from the right column to the left.

How work-made-for-hire and assignment clauses decide it#

Work-made-for-hire and assignment clauses decide ownership in different ways. Under US copyright law, a commissioned work counts as made for hire only if it fits a statutory category, such as a compilation, a contribution to a collective work or a test, and both sides sign a written agreement saying so. When a work qualifies, the commissioning party is treated as the author and owns the copyright unless a signed writing says otherwise.

Simulator code and estimation software often do not fit those categories, so client contracts that rely only on work-made-for-hire language may not transfer them. That is why many research agreements add an assignment of all work product, which can sweep in agency tools unless a background IP clause carves them out.

The statistical idea behind a typing rule is not protected by copyright. The coefficients, code and documentation are where ownership questions actually bite.

Delivery format shapes what the client can do#

Delivery format decides what a client can do with a tool even after ownership is settled. A typing tool delivered as a spreadsheet with visible coefficients lets the client score new respondents anywhere, while a simulator hosted by the agency gives the client use, not possession, for as long as the hosting agreement runs.

Locked workbooks deserve a specific clause. Client analysts routinely try to unlock formulas to add a scenario or a competitor, and a simulator delivered without terms on that point invites the client to treat the embedded logic as its own.

Clauses to settle before fieldwork#

The clauses to settle before fieldwork are the ones both sides skip when a project is urgent. Adding them to the proposal or statement of work costs little and avoids renegotiating when the client changes agencies.

Trackers need extra care because they run for years. A typing tool fitted in the first wave and refit in later waves can absorb agency improvements that neither side documented, so record each refit, what changed and who paid for it.

• 1Background IP: name the agency's pre-existing methods, code, design tools and templates.
• 2Deliverables: list what the client receives and in which format.
• 3Assignment scope: limit any assignment to study-specific outputs.
• 4Embedded tools: license the client to use any agency component built into a deliverable.
• 5Respondent data: state who controls it, how long it is kept and under which panel terms.
• 6Normative use: whether the agency may add de-identified results to cross-client norms.
• 7Transition: what the agency hands over if a tracker or segmentation moves to another supplier.

Illustrative: a quant agency hands over a conjoint study#

Illustrative: a fictional quantitative research agency built its own conjoint simulator platform and has delivered many studies through it. A long-standing consumer durables client moves its research to a larger agency and asks for the simulator source code along with its data.

The agency's master agreement assigns study deliverables to the client but lists the simulator platform as background IP. The agency delivers the utilities file, the experimental design, the respondent-level data the client already owns, and a standalone workbook simulator under a perpetual internal-use license. It keeps the platform code.

The client's new agency rebuilds a simulator from the utilities file. Because the deliverables list was specific, the handover takes a short call instead of a legal dispute.

Respondent data and how SourceX approaches research records#

Respondent data is the most restricted part of any quant project. Panel provider terms, respondent notices and client agreements usually limit it to the study, and open-ended answers can identify people, so agencies should assume respondent-level files are excluded from reuse unless every layer allows it.

SourceX approaches research records through the SourceX five-step transaction: Supply, Rights, Preparation, Approval and Delivery. Agency-owned material, such as method documentation, design decisions and internal project reviews, is the usual starting point. Under the SourceX Enterprise Data Value Framework, domain expertise and uniqueness raise value while privacy burden and preparation cost reduce net value, which is why de-identified methods records often outrank raw survey files.