Logistics Technology IP Checklist: Shipment Data Rights, Platform and API Terms, Carrier and Shipper Agreements, Benchmarking Outputs, and Customs Interfaces

By Casey Scott McKay · 2025-11-03

This checklist audits the intellectual property position of a freight, logistics, or supply chain technology business in the order the questions arise. It starts with the data clause, because almost nothing in this sector is defensible by patent and the property position is created by agreement rather than found in doctrine. It then works through the four platform terms, integration ownership and API dependency, benchmarking products and their competition law exposure, rate files and departures, counterfeit interdiction, contract logistics exit, telematics and condition records, customs interfaces, and the small technical filing programme. Gate items mark where work should stop.

IP and Technology > Information Technology | Checklist | Published 3 November 2025 - Updated 17 February 2026 | Casey Scott McKay - marksy.us

Summary. This checklist audits the IP position of a freight, logistics, or supply chain technology business in the order the questions arise. It starts with the data clause, because almost nothing in this sector is defensible by patent and the property position is created by agreement rather than found in doctrine. It then works through the four platform terms, integration ownership and API dependency, benchmarking and its competition law exposure, rate files and departures, counterfeit interdiction, contract logistics exit, telematics records, customs interfaces, and the small technical filing programme. Gate items mark where work should stop.

Keywords: logistics checklist · three category data clause · platform permitted use · aggregation threshold · competitive restriction · exit and deletion · integration ownership audit · API dependency · benchmarking source audit · rate file controls · departure process · counterfeit intake · contract logistics exit · telematics evidence · customs retention

How to use this checklist

| Phase | What it produces | Who runs it | Gate | |---|---|---|---| | 1. Data clause | A three-category allocation deployed everywhere | Counsel | Clause settled before the next enterprise deal | | 2. Platform terms | Four terms resolved per relationship | Commercial and counsel | Exit term present | | 3. Integrations | An ownership audit and a new template | Counsel and engineering | Statement of work template changed | | 4. Benchmarking | Source audit, thresholds, competition review | Counsel and product | Nothing launches without review | | 5. Rate files | Marking, access control, and departure process | HR and counsel | Exit process run every time | | 6. Counterfeits | Intake, record, and escalation process | Operations and counsel | No inspection duty assumed | | 7. Contract logistics | Exit and transition terms | Commercial and counsel | Transition assistance defined | | 8. Telematics | Access and preservation terms | Counsel | Evidence preservation present | | 9. Customs | Retention and confidentiality positions | Compliance and counsel | Statutory retention addressed | | 10. Filing | A small, technical programme | Counsel | Eligibility screened honestly |

The matter. A freight technology company operates a visibility platform with eighty carrier integrations and two hundred shipper customers, has just launched a lane rate benchmark built from platform data, runs a small brokerage on the side, and lost two commercial managers to a competitor last quarter. The integrations were built by three outside firms under statements of work. The data clause was adapted from a software agreement. Nobody has read the source agreements underlying the benchmark.

Phase 1. Rewrite the data clause into three categories

[ ] Allocate customer transactional data to the customer: bookings, rates, origins and destinations, commodities, volumes, party identities, and terms of sale. It describes the customer's business, and fighting for it spends credibility needed elsewhere.
[ ] Take broad rights to operational data — status events, exceptions, handling records, system telemetry, performance measurements — as an irrevocable, perpetual, worldwide licence for any purpose relating to the provider's products and services.
[ ] Own derived analytics outright: transit time distributions, carrier scores, dwell analytics, and predictive models, with the customer's benefit delivered through service improvement.
[ ] State survival expressly for data already collected, since a provider required to purge its history when a customer leaves has no history.
[ ] Add an evidence preservation obligation triggered by notification of a claim, with a named recipient inside the business.
- Why. Condition and status records decide cargo claims, they have the shortest default retention, and an obligation with no trigger and no named owner is an obligation nobody performs.
[ ] Address aggregate use in the same clause, with a stated threshold rather than the words "aggregated and anonymised."
[ ] Answer the security and access questions specifically: where the data goes, who can see it, and for how long.
[ ] Understand why contract does the work. A record that a container left port on the fourteenth is a fact, and Feist Publications, Inc. v. Rural Telephone Service Co. settles that facts are not owned. Database copyright is thin. The property position here is created by agreement.
[ ] Confirm the business actually receives what the clause grants. Providers routinely negotiate broad rights and never build the pipeline, leaving a clause with no asset behind it.
[ ] [Gate] The three-category clause is settled before the next enterprise negotiation, not drafted inside it.

What the records actually are

Before allocating anything, be precise about the categories, because a clause written against a vague notion of "shipment data" will handle several of them badly.

Transactional records. Booking, rate, origin and destination, commodity, weight, dimensions, party identities, terms of sale, and payment. Commercially sensitive to the shipper because they reveal sourcing, volume, and pricing — and visible to the forwarder, the carrier, the broker, the customs agent, and the bank simultaneously.

Status and location records. Where the unit is, when it moved, whether it is late. Generated by carriers, terminals, telematics devices, and container-mounted sensors that may belong to a fifth party.

Condition records. Temperature, humidity, shock, tilt, and door events. Decisive for pharmaceutical and food shipments and generated by devices with contested ownership.

Documentary records. Bills of lading, packing lists, certificates of origin, phytosanitary certificates, and customs declarations. These have legal effect independent of their information content — a bill of lading is a document of title — and every party keeps a copy.

Derived records. Transit time distributions, carrier performance scores, lane rate indices, dwell analytics, and predicted arrival estimates. Built from the above, and where the commercial value concentrates.

[ ] Confirm the data clause distinguishes all five, since documentary records carry retention duties and legal effect that a general data provision handles poorly, and derived records need an ownership answer the other four do not.
[ ] Identify which records the business generates versus receives, because rights to received data depend on the supplying party's grant and rights to generated data do not.
[ ] Map who else holds each category, since the co-generated and jointly visible character of freight data is what makes this sector's ownership questions unlike others.

Phase 2. Fix the four platform terms

[ ] Define permitted use of ingested data. May the platform use a participant's data beyond providing the service to that participant?
- Trap. Silence means yes in practice, because the platform will and the participant will not know.
[ ] Set the aggregation threshold by market structure. State the minimum number of contributing parties and transactions per output cell, and suppress cells below it. On a lane with three carriers, an aggregate rate discloses each carrier's price to the others.
[ ] Define the output. Whether an aggregate may appear in a free market report, a paid product, an internal pricing model, or a training set. These are four concessions and customers assume they are granting the first.
[ ] Impose a competitive restriction. May the platform enter the participant's business using knowledge derived from serving it? This is the term nobody asks for and everybody needs, and it matters when a visibility provider becomes a broker.
[ ] Settle exit: deletion, return, or retention on termination.
- Why. If a participant can negotiate only one of the four, this is the one, because it is the only one whose absence cannot be remedied afterwards.
[ ] Add an identifiability notice mechanism — how a customer raises a concern that an output identifies it, an obligation to suppress, and a timeframe — since otherwise the customer's only remedy is termination.
[ ] Address model training expressly, because silence is read as permission by anyone sophisticated.
[ ] State whether aggregate outputs already produced survive termination. They almost always must, and saying so avoids a dispute at exit.
[ ] Confirm the architecture supports segregation where restrictions are promised, since a contractual promise the systems cannot keep is a breach waiting for an audit.
[ ] [Gate] No platform agreement is signed without the exit term resolved.

Carrier and shipper agreements, specifically

The two counterparty types need different treatment, and a single template applied to both produces a clause that protects against neither.

Carrier agreements. The carrier generates status, exception, and performance data about its own service, and it has a legitimate interest in not having that data used against it or shared with its competitors.

[ ] Define what the platform may do with carrier performance data, distinguishing use in service delivery to the booking shipper from use in scorecards visible to other shippers from use in aggregate market products.
[ ] Address whether the platform may broker freight in competition with the carrier's direct customers, which is the competitive restriction question from the carrier's side.
[ ] Settle rate confidentiality, since carrier rates disclosed to a platform that also brokers are rates disclosed to a competitor.
[ ] Address telematics, where the carrier owns the device and the shipper wants the data.

Shipper agreements. The shipper's transactional data reveals its sourcing, volumes, and pricing, which are among its most sensitive commercial facts.

[ ] Confirm the shipper's data is not visible to its competitors through any aggregate product, at any threshold.
[ ] Address whether the provider may approach the shipper's suppliers or customers using knowledge gained from handling its freight.
[ ] Settle what happens to the shipper's history on termination, since a five-year record of lane volumes and rates is the shipper's negotiating position with every carrier.
[ ] Address the shipper's own audit rights with a defined trigger and scope.
[ ] Do not use one template for both, because the interests are genuinely opposed and a clause drafted to balance them protects the party who drafted it.

Phase 3. Audit integration ownership and price the API dependency

[ ] List every integration — carrier, terminal, customs system, customer enterprise system — and identify who built it and under what agreement.
[ ] Read each statement of work for ownership. Where an integrator built it and the statement of work is silent, the integrator keeps it.
- Why. Integrations are the network and the network is the asset. This is the highest-value single change available in a logistics technology business, and it applies prospectively from the day the template changes.
[ ] Change the statement of work template this quarter: assignment where achievable, and at minimum a perpetual, sublicensable licence covering reuse with other customers.
[ ] Secure reuse rights explicitly, since an integration owned but not reusable does not compound.
[ ] Read the API terms the business depends on, recording rate limits, permitted uses, redistribution rights, derived product restrictions, and termination rights.
- Trap. Termination of API access has ended companies. A provider reserving the right to terminate at will has retained an option to eliminate a competitor, and the model should be priced accordingly.
[ ] Do not rely on copyright in interface specifications. Google LLC v. Oracle America, Inc. resolved the leading dispute on fair use grounds while declining to decide copyrightability, so a position depending on interface control should rest on contract and technical measures.
[ ] Understand the limits of access statutes where data is obtained without an agreement, since Van Buren v. United States narrowed the reach of the exceeding-authorised-access theory and pushes the dispute back to contract.
[ ] Check standards participation in freight message standards bodies for disclosure and licensing obligations.
[ ] Confirm configuration and workflow export rights in every bought system, in a documented format, with a test extract performed during the term.
[ ] [Gate] No new integration is commissioned under the old statement of work template.

Software the business buys

[ ] Confirm configuration and workflow ownership and export for transport management, warehouse management, and customs filing systems, in a documented format with a test extract during the term.
- Why. A deployment configured over three years is an asset, and a licence silent on export has converted the customer's configuration work into the vendor's asset — discovered during a competitive retender.
[ ] Read the vendor's rights to operational data from the deployment, since vendors increasingly build analytics products from customer telemetry and the concession is larger where the vendor also serves competitors.
[ ] Settle data return on termination in a usable format, because a five-year operational history retained by the vendor is a head start withheld from the next provider.
[ ] Obtain escrow with tested build verification where a small vendor's system carries an operation, since an untested deposit is a comfort letter rather than a continuity plan.
[ ] Require a bill of materials per release from vendors and from the internal build alike, with licence identification.
[ ] Check whether the vendor claims ownership of connectors built to its platform, which is the integration ownership problem arriving through a different document.
[ ] Confirm the right to permit third-party support after a defined period, and the access to tooling needed to exercise it, since 17 U.S.C. § 1201 governs circumvention of access controls and its triennial exemption process makes any technical gating an unstable entitlement rather than a settled one.

Phase 4. Clear the benchmarking product

[ ] Audit the source agreements for secondary use permission. Most benchmarking products are built on data collected for another purpose, and the usual answer is that nobody looked.
[ ] Test whether the aggregation genuinely anonymises given the market structure, lane by lane rather than by a general rule.
[ ] Bring competition counsel in before launch.
- Trap. Information exchange between competitors about prices and capacity attracts sustained regulatory attention, and a product letting competitors observe each other's pricing in near real time is a problem regardless of how the contracts read.
[ ] Design the mitigations in: aggregation, delay, historical rather than forward-looking data, independent administration, and no participant-identifiable output.
[ ] Own the right things. The methodology, weighting, and selection are authorship and trade secret material; the underlying observations are facts. Claiming the whole product overstates, and treating it as unownable underestimates.
[ ] Document the methodology in a form that supports both the trade secret claim and any regulatory explanation.
[ ] Record the contributor benefit position — product access, fee reduction, or nothing — since nothing is legitimate if stated and a grievance if discovered.
[ ] [Gate] Nothing publishes without the source audit, the threshold design, and the competition review.

[ ] Confirm the aggregation threshold is enforced by the system rather than by a policy, since a suppression rule that depends on an analyst remembering it will fail on the day the output is most sensitive.
[ ] Test the benchmark against a known participant's data before publication, checking whether that participant could recognise itself in the output. If it can, the threshold is wrong regardless of what the clause says.

Phase 5. Protect the rate file and manage departures

[ ] Name what is protectable: rate cards, carrier scorecards, customer volume and margin profiles, and account-specific terms. These derive value from not being generally known and cost real money to build, and 18 U.S.C. § 1839 protects them where reasonable measures were taken.
[ ] Be honest about what is not. The identity of companies importing from Asia is not a secret, nor is the name of a company's logistics manager, nor are the personal relationships an employee built.
- Trap. Courts are unsympathetic to employers who conflate the categories, and an over-claimed case is a lost case.
[ ] Mark the rate file and control access by role rather than by default.
[ ] Maintain an access log, since a bulk download in a final week is the evidence that decides cases.
[ ] Name the categories in confidentiality agreements rather than reciting generalities.
[ ] Run a structured exit every time: access revoked on the day, devices returned and imaged, an interview walking the categories, and a signed acknowledgement.
[ ] Know the covenant position per jurisdiction, since restrictive covenant law varies enormously and is moving toward greater restriction in several states.
[ ] Obtain an onboarding acknowledgement from every incoming hire, which is the first document produced when the claim arrives from the other direction.
[ ] Confirm the federal claim is available, noting that 18 U.S.C. § 1836 supplies it and contains an express limit on enjoining employment based merely on what a person knows.
[ ] [Gate] No commercial departure proceeds without the exit process.

Phase 6. Build the counterfeit interdiction programme

[ ] Understand the exposure. Contributory liability attaches to a party who continues supplying services to one it knows or has reason to know is infringing, following Inwood Laboratories, Inc. v. Ives Laboratories, Inc. as applied to service providers.
[ ] Recognise the warehouse operator's sharper position, since fulfilment operators hold the goods, are identifiable, and are within the jurisdiction when the seller is not — and whether storage and shipment is itself infringement under 15 U.S.C. § 1114 turns on control over the transaction.
[ ] Include terms permitting refusal of shipments and disclosure of shipper details on lawful request.
[ ] Establish a notice intake process with a named owner, and record what was reported and what was done.
[ ] Set an escalation threshold for repeat shippers and apply it.
[ ] Agree cooperation protocols with brand owners that do not commit the provider to policing beyond its capability.
[ ] Do not agree to inspect for infringement.
- Trap. A provider that assumes an inspection duty has taken on something it cannot discharge and has created the knowledge that founds liability.
[ ] [Gate] No cooperation agreement commits the business to an inspection obligation.

Marketplace and disintermediation terms

Where the business operates a marketplace, its model rests on connecting parties who could find each other unaided. Five items follow.

[ ] Draft the circumvention clause with restraint. A clause purporting to prevent a shipper from ever dealing with a carrier it already knew will not be enforced; one confined to parties actually introduced through the platform, for a defined period, has a chance.
[ ] Build depth rather than relying on the clause. A marketplace whose value is introduction is fragile; one whose value is performance history, settlement, insurance, and dispute resolution has embedded itself in the transaction. Counsel supports that by ensuring the platform owns the records it creates.
[ ] Protect the verification records — carrier authority, insurance status, safety history, past performance — assembled at real cost. Terms restricting export of verification data are more defensible than terms restricting contact.
[ ] Settle rating ownership and republication, and provide a dispute and response process where ratings carry narrative rather than numbers, since defamation exposure is real.
[ ] Authorise the policing use expressly, because detecting fee circumvention relies on data the platform holds about both sides and that use should not be left to inference.

Brand and impersonation

[ ] Register in the service classes and jurisdictions where the network operates, not only where it is headquartered.
[ ] Monitor for impersonation rather than competitive use, since freight fraud works by fictitious pickups, near-identical invoicing entities, and cloned quotation sites.
- Trap. Speed matters more than remedy here. A registrar complaint that removes a cloned site in a day is worth more than a judgment in a year, and the claims under 15 U.S.C. § 1114 and 15 U.S.C. § 1125 are the backstop rather than the plan.
[ ] Put real quality control into agent and partner agreements, since a network whose partners use the brand without control is exposed to a naked licensing argument under 15 U.S.C. § 1064.
[ ] Check that named service products are delivered as promised by every partner using the mark, because a divergence is a brand problem and a liability problem at once.

Phase 7. Fix contract logistics exit terms

[ ] Read the exit provisions before the service levels. The service levels govern the relationship; the exit provisions govern whether it can end.
[ ] Define transition assistance: a period, rates, and deliverables. Without those three, the exit clause does not permit exit.
[ ] Specify the data export format and schema documentation, and perform a test extract during the term so the format is known to work before it is needed.
- Trap. "Reasonable assistance" produces a data dump nobody can load.
[ ] Allocate the system configuration, which encodes the operating design in executable form and determines whether the customer can move without rebuilding.
[ ] Require standard operating procedures as actually performed, not as originally written, since the two diverge within months of go-live.
[ ] Allocate operating design, performance data, and continuous improvement work expressly, including whether the provider may deploy an improvement developed at the customer's site elsewhere.
[ ] Allocate automation separately across equipment, control configuration, and integration, because the right answer differs across the three and nobody will agree it under exit pressure.
[ ] Address personnel transfer arrangements at signing, since employment law varies by jurisdiction and is unpleasant to discover at exit.
[ ] [Gate] No contract logistics agreement is signed without defined transition assistance.

Phase 8. Settle telematics and condition records

[ ] Identify the claimants for each device: the device owner, the asset owner, the cargo owner, and the monitoring service. Nobody resolves this in advance and it becomes acute during a claim.
[ ] Give the cargo owner full access to data about its own shipments, which is where the commercial interest is strongest and the contractual position weakest.
[ ] Let the device operator retain rights to operate and improve the service.
[ ] Address aggregate use expressly with a stated threshold, since monitoring providers accumulate lane-level intelligence about which routes damage cargo from customers' shipments.
[ ] Impose evidence preservation on whoever holds the record when a claim is notified.
- Why. Condition data determines whether pharmaceutical or food cargo is saleable and who pays. A party that cannot obtain the record cannot prove its claim, and access rights settled after the loss are settled too late.
[ ] Address performance use — whether a shipper may use carrier-generated telematics in a rate negotiation — which carriers will resist and which is worth having.
[ ] [Gate] No monitoring arrangement operates without an evidence preservation term.

Phase 9. Handle customs, instruments, and compulsory retention

[ ] Address statutory retention expressly in customer agreements, stating that customs recordkeeping obligations prevail over a customer deletion instruction and defining what the filer may and may not do with the archive.
- Trap. A forwarder holds years of customer commercial information because the law requires it, and a data clause silent on compulsory holding creates a conflict nobody can resolve.
[ ] Advise import customers on confidential treatment of manifest data, which is available in some jurisdictions and which most importers do not know they can request.
[ ] Check screening tool terms for permitted use, retention of screening results, and derived data, since a counterparty database built from screening outputs may rest on terms that do not permit it.
[ ] Fix broker data terms at appointment, including whether the broker may use aggregate filing data commercially, since the data accumulates from day one.
[ ] Treat electronic transferable record systems as registries, whose legal effect depends on the model law's adoption in the relevant jurisdictions and whose operator holds real commercial power.
[ ] Respect third-party document terms for certificates of origin, phytosanitary certificates, and inspection reports issued by chambers, agencies, and inspection companies.
[ ] Note where documentary credits are involved, since a data error in a shipping document becomes a financial dispute rather than an operational one.
[ ] Record trade programme obligations that require sharing documented practices about the participant and its partners.
[ ] [Gate] The retention conflict is resolved in the template before the next customer agreement.

Phase 10. Make the filing decision, and keep it small

[ ] Screen honestly. A claim to matching loads to trucks, optimising a route, or notifying a customer of an exception is a business method in technical vocabulary, and 35 U.S.C. § 101 with Alice Corp. v. CLS Bank International is where it fails.
[ ] File where a real technical improvement exists: a specific data structure, a named failure-mode handling, a real-time constraint approach, or hardware-involved innovation in scanning, sortation, or sensing.
[ ] Confine claims to a single actor, since logistics methods span shippers, carriers, platforms, and terminals and a claim no one entity performs runs into Limelight Networks, Inc. v. Akamai Technologies, Inc..
[ ] Test breadth against 35 U.S.C. § 112 where claims recite optimisation across conditions the specification does not describe.
[ ] Keep technical documentation of accused systems current, because an early eligibility motion is the most cost-effective response to the non-practising entity assertions this sector attracts against tracking, routing, and notification functions.
[ ] Protect the rest by secrecy and speed: pricing models, matching heuristics, and network intelligence are unobservable from outside and their value lies in accumulated data.
[ ] Confirm contributor assignment for internally written software, including contractors.
[ ] Inventory open source, since providers delivering software as an appliance or on-premises component may be distributors with obligations they have not considered.
[ ] [Gate] No filing programme is approved before the integration audit is complete.

A note on order

The phases are ordered by how much of the business each change reaches and how quickly.

The integration audit and template change come first because they apply prospectively from the day they are made. Every week of delay adds integrations owned by an outside firm, and unlike most items on this list the cost of delay is linear and permanent — an integration built under the old template is not retroactively fixed by a new one.

The data clause is second for the same reason inverted: it cannot be deployed to existing customers except at renewal, which means the rollout runs over a full renewal cycle. Starting a month late costs a year of coverage, and starting after a large customer's renewal costs that customer's term.

Benchmarking is placed third but becomes first if the product has already launched, because an unreviewed product in market is an accruing exposure rather than a pending decision.

Rate file controls and the departure process are template and process work that applies immediately, which is why they can sit in the middle without cost. They are also the phase most often skipped, because nothing forces the timing until someone resigns.

Counterfeit intake, contract logistics exit, telematics, and customs are each triggered by external events — a brand owner's notice, a retender, a cargo claim, an audit — and each is far cheaper to have in place beforehand than to construct under the event.

The filing programme is deliberately last. A business that files patents while its integrations belong to somebody else has spent its budget on the least likely risk, and the gate on Phase 10 exists to prevent exactly that sequencing.

[ ] Check insurance cover for intellectual property claims, since general liability policies routinely exclude them and specialist cover is rarely bought in this sector despite the assertion volume against tracking and notification functions.
[ ] Confirm any lender's security description reaches the data assets, because the integration library and the operational history are frequently the most valuable items in the package and the least clearly described in it.
[ ] Record which contracts survive a change of control and which grant termination rights on one, since a platform business acquired by a carrier will find that several shipper agreements contemplated exactly that.

Outcome. A business that has run this checklist can say who owns the eighty integrations that constitute its network, whether its benchmark rests on data it was permitted to use, what happens to five years of shipper history if a customer leaves, and whether two departed commercial managers took the rate file or only their contacts. Those four answers determine what the business is worth and what it can safely do next, and none of them is a patent question.

Key Authorities at a Glance

| Authority | What it settles | Phase | |---|---|---| | Feist Publications, Inc. v. Rural Telephone Service Co. | Facts unprotectable; thin compilation copyright | 1 | | 18 U.S.C. § 1839 | Reasonable measures element of trade secret status | 5 | | 18 U.S.C. § 1836 | Federal civil misappropriation claim and its employment limit | 5 | | Google LLC v. Oracle America, Inc. | Reimplementation of a functional interface was fair use | 3 | | Van Buren v. United States | Narrow reading of exceeding authorized access | 3 | | Alice Corp. v. CLS Bank International | Two-step eligibility framework | 10 | | 35 U.S.C. § 101 | Patentable subject matter | 10 | | Limelight Networks, Inc. v. Akamai Technologies, Inc. | Single-actor requirement for direct infringement | 10 | | 35 U.S.C. § 271 | Acts of infringement | 10 | | 35 U.S.C. § 112 | Written description and enablement | 10 | | Inwood Laboratories, Inc. v. Ives Laboratories, Inc. | Contributory liability for continuing to supply a known infringer | 6 | | 15 U.S.C. § 1114 | Infringement of a registered mark | 6 | | 15 U.S.C. § 1125 | False designation of origin and false advertising | 6 | | 15 U.S.C. § 1064 | Cancellation, including for abandonment | Agent network licensing | | 17 U.S.C. § 102 | Copyright subject matter; ideas and methods excluded | 4 | | 17 U.S.C. § 103 | Compilations and derivative works | 4 | | 17 U.S.C. § 106 | Exclusive rights including distribution | 10 |

The five things people get wrong

One: letting the integrator own the integrations. Eighty carrier integrations are the network, the network is the asset, and each one was built by somebody under a statement of work that said nothing about ownership. Three years later the integrator holds a library of connectors it can offer to a competitor, and the company that funded them has a licence at best. Changing the template takes an afternoon and applies to everything built afterwards, which is why the change should happen before anything else on this list.

Two: writing "aggregated and anonymised" and calling the aggregation term done. Adjectives are not enforceable. On a lane served by three carriers an aggregate rate discloses each carrier's price to the others, and a customer who discovers this has a grievance the contract does not answer. Put numbers in the clause — minimum contributors, minimum transactions, suppression below threshold — and put a notice mechanism next to them so a customer who believes an output identifies it has a remedy short of termination.

Three: launching a benchmark without reading the source agreements. Benchmarking products are built from data collected to operate a platform or clear a shipment, and whether those collection agreements permit the secondary use is a question nobody asks until a customer's counsel does. The competition law question is worse, because it is not answered by drafting at all: a product that lets competitors observe each other's pricing in near real time is a regulatory problem however carefully the contracts read.

Four: conflating the rate file with the relationships. When a commercial manager leaves, the employer wants to stop everything and usually claims everything, which loses the case that could have been won. The rate cards, scorecards, and margin profiles are protectable and the relationships are not. An employer that marked the file, controlled access, logged it, named the categories, and imaged the laptop has a real claim about a real secret. One that claims its customer list is confidential because the customers are its customers has an argument a court will dislike.

Five: spending the budget on patents. Routing and matching claims face the eligibility screen and mostly lose, and the ones that survive are narrow enough that a competitor can work around them. Meanwhile the integrations are owned by outsiders, the rate file is on a shared drive, and the exit clause in the largest contract logistics agreement has no transition assistance. The budget belongs in contracts, employment documentation, and takedown infrastructure, and a general counsel who reverses that has protected the business against its least likely risk.