Spend analysis data from procurement consulting: what can be reused

Who owns a spend cube built from client data?#

A spend cube built from client data usually belongs to the client as a deliverable, while the tools the firm used to build it may stay with the firm. The raw inputs, such as accounts payable extracts, purchase orders, invoice lines, purchasing card transactions and the supplier master, are client confidential information under almost every engagement agreement.

The ownership line runs through the contract's IP clause. Firms that defined their taxonomy, classification rules, enrichment logic and supplier mapping tables as background IP can usually keep improving them across engagements. Firms working on client paper that assigns all work product to the client may find that even their improved mappings belong to that client.

A useful test is to ask what the firm brought to the first day of the engagement. The taxonomy, the rules engine and the analyst playbook existed before the client's data arrived; the cleansed vendor list and the finished cube did not. Writing that inventory into the statement of work avoids the argument later.

What in a spend engagement can usually be reused?#

The reusable parts of a spend engagement are mostly methods, not data. The table shows typical treatment; your agreements decide the actual answer for each client.

Three tests before any cross-client reuse#

Cross-client reuse of spend data has to pass three tests: aggregation, antitrust review and contract consent. A benchmark that fails any one of them should stay inside the engagement that produced it.

Run the tests in that order. Aggregation design is cheap to change early, antitrust review needs a concrete design to assess, and contract consent is easiest to confirm once you know exactly which fields and clients a benchmark will draw on.

• 1Aggregation: results combine enough clients that no single client's spend or prices can be inferred.
• 2Antitrust caution: price information is old enough, broad enough and anonymous enough that it cannot help competitors coordinate.
• 3Contract consent: each contributing client's agreement allows aggregated or de-identified use, or the client has consented in writing.

Aggregation thresholds and suppression#

Aggregation thresholds set the minimum number of contributing clients per benchmark cell and limit how much any one client can dominate it. The exact thresholds are a policy choice made with counsel, but they should be written down and applied the same way every time.

Disclosure risk applies to companies as well as people. NIST SP 800-188, written for government statistical releases, aims to limit disclosure risks to individuals and establishments while still allowing meaningful statistical analysis, and its approach can inform commercial benchmark design.

• 1Set a minimum contributor count per cell and suppress cells below it.
• 2Apply a dominance rule so no single client accounts for most of a cell.
• 3Coarsen periods, regions and categories until cells clear both rules.
• 4Remove supplier names from price benchmarks unless the supplier has agreed otherwise.
• 5Log each release: sources, rules applied and who approved it.

Why supplier price data needs antitrust caution#

Supplier price data needs antitrust caution because sharing current, specific prices among competitors may help them coordinate, and antitrust laws may apply even when a consultant is the channel. Procurement benchmarks often go to clients that compete for the same inputs, and the suppliers named in them also compete with one another.

Lower-risk benchmarks use historical rather than current data, report ranges or indexes rather than specific prices, cover enough contributors that no one's terms can be reverse engineered, and are run by staff outside the clients' commercial teams. Have antitrust counsel review the design before the first release, not after a client asks for more detail.

Supplier-side confidentiality adds a second layer. Many supply agreements bar the buyer from disclosing pricing, so a client that shared its contracts with the consultant could itself be exposed if those prices surface in another client's benchmark.

Which contract clauses decide reuse?#

Reuse is decided by a handful of clauses in the master services agreement and statements of work. Read them for each client before any record enters a shared tool, benchmark or knowledge base.

Illustrative: an indirect spend consultancy draws the line#

Illustrative: a fictional procurement consulting firm has classified indirect spend for mid-market manufacturers and distributors for many years. Its classification library maps invoice descriptions to a category taxonomy, and its analysts built supplier name normalization tables along the way.

The partners want to reuse that library in a new analytics product and ask what else could be licensed. Contract review shows most engagements reserved the taxonomy and rules as firm IP, while client extracts had to be deleted at close. Several old client files were still on an analyst's laptop.

The firm deletes the retained extracts and documents the deletion. It keeps only the taxonomy, the rules and public supplier names, drops price fields from every reusable table, and adds an aggregated data clause with written thresholds to its standard agreement for new clients.

How SourceX approaches procurement consulting records#

SourceX approaches procurement consulting records by separating firm-authored methods from client data before anything else. In the Rights step of the SourceX five-step transaction, client extracts, prices and bids are excluded by default, and only material the firm owns or has written permission to use moves on to Preparation.

Under the SourceX Enterprise Data Value Framework, a well-maintained taxonomy with documented classification decisions draws on domain expertise, data cleanliness and clear rights. Supplier price data usually fails on rights, because client and supplier confidentiality terms attach to it, and the aggregation needed to make it safe adds preparation cost that often outweighs its value. Each approved package carries a SourceX Evidence Packet recording provenance, licensing rights, permitted use, the privacy record and release authorization.