Outsourcing is approved as a transfer of cost and executed as a transfer of capability; only the second one fails to reverse.
A board approves a contract manufacturing arrangement. The paper compares unit cost, capital avoided and time to market. It is a competent paper, and it describes a transaction.
What follows over three years is not a transaction. It is a migration — of specifications, then tooling, then process settings, then the working knowledge of why the settings are what they are, then the second-tier relationships held only by whoever placed the orders. Each step is small, commercially sensible, and approved by nobody in particular.
At the three-year review the unit economics are as forecast, and the question that matters was never on the paper: if this relationship ended badly, what could we resume, where, and how long would it take? In many organisations resumption is not available at any price the business would pay.
Whether to outsource at all is a separate decision, taken on its own terms in [Related article: Is Your Make-or-Buy a Cost Decision or a Capability Decision?]. This is about what crosses the boundary once it is made, and which protections are narrower than those relying on them believe.
The Strategic Context
Manufacturing outsourcing is a knowledge transfer wearing the clothes of a procurement transaction. Commercial terms are negotiated by people thinking about price, volume and lead time; the protective instruments are drafted afterwards, treated as a closing formality and assessed for completeness rather than fit — yet their value depends on three things settled early and commercially: sequence, jurisdiction and scope.
Underneath sits an asymmetry no drafting removes. Your counterparty acquires, cumulatively and irreversibly, the ability to make your product. You acquire, contingently, a right of action if they misuse it. One is a capability, held permanently and exercised at will; the other a claim, exercisable in a forum, at a cost, over a period, against a defendant with assets in a jurisdiction you did not select.
Four Confident Beliefs That Do Not Hold
The first is that the confidentiality agreement is the protection. It is an obligation, not a barrier: the remedy is damages for a loss hard to quantify and harder to detect, and it works on the counterparty's incentives, not their capability.
The second is that registration produces worldwide protection. It does not, and that gap is the most consequential one here.
The third is that the brand sits on the balance sheet as a valuable intangible and is governed like other assets of comparable worth. For a brand you built yourself the premise is false, and a governance failure follows.
The fourth does the most damage: that sequence is a technicality and signing can catch up with commercial reality. Once technical detail has been disclosed it cannot be undisclosed, and an instrument signed afterwards addresses a state of affairs that no longer exists.
Reframing the Issue
Replace "is our intellectual property protected?" with "what have we kept recoverable, and within what period?"
Protection is a legal state. Recoverability is an operational one, and it determines what happens if the relationship fails. The reframe puts everything transferred on one axis: could we resume this ourselves, or with an alternative supplier, and how long would it take? Some items cannot be recovered at any speed, and the response is not to protect them harder but to price the dependency honestly.
What Actually Crosses the Boundary
Six distinct things move, and the conventional checklist reaches only the first two.
Documented specification is recoverable, because you hold a copy. Tooling is conditionally recoverable, depending on title, where the equipment sits, and how straightforward removal would be from a foreign site against an unwilling holder.
The other four differ in kind. Process knowledge — settings, sequence, the compensations applied when material varies, the operator's sense of what running well looks like — is the largest transfer and the one no document records, because it does not exist until it is created inside the counterparty's plant, on their equipment. Yield learning accrues wherever production happens. Upstream relationships migrate quietly, because the party placing orders is the party the sub-tier knows. And the capability to serve a competitor, or become one, is permanently created and breaches nothing unless you contracted against it.
Confidentiality obligations protect what is provable; registrations what is registrable; supply agreements what is documented. Tacit capability is none of these, which is why the checklist can be completed in full while the substantive exposure is untouched.
Two protections need no court: refreshing the design on a deliberate cadence, and advancing the technology faster than a copyist can follow. Both impose a rhythm on the roadmap that must be funded; treating either as free is how the control quietly stops happening. They make the copy obsolete rather than preventing it — staying ahead rather than excluding.
Protections That Stop at a Border
Registered rights are national grants. A trade mark registered in Australia gives rights in Australia. There is no global trade mark registration, and no instrument creates one. [FACT CHECK REQUIRED]
The Madrid Protocol is routinely described as a world registration. It provides a single international application through which an applicant may designate member countries; each designated office examines that designation under its own law and may accept or refuse it. [FACT CHECK REQUIRED] The output is a bundle of national rights obtained through one channel — a procedural convenience, not an expansion of territory.
Coverage is therefore a map, and the jurisdictions left off it are usually omitted on cost grounds — which correlate closely with where low-cost manufacturing sits, and so with where a copy would be made and first sold. An organisation registered in its home and major sales markets but not where its goods are made has protected where it would sue and left the place of origin unprotected. [FACT CHECK REQUIRED]
Two Australian specifics are commonly misunderstood. Falsely representing that a trade mark is registered is an offence — section 151 of the Trade Marks Act 1995 (Cth). [FACT CHECK REQUIRED] Treating the ® symbol as a general confidence signal, applied everywhere because the mark is registered somewhere, creates exposure in every market where the mark is not registered. Separately, a registered design in Australia must be certified following examination before it can be enforced. [FACT CHECK REQUIRED] Registration alone produces a right that cannot be asserted, at exactly the moment a copy reaches market.
All of this requires professional legal verification for your own circumstances and jurisdictions. ERANORTH is neither a law firm nor a financial adviser, and nothing here is legal advice.
The pattern is not confined to intellectual property: executives routinely rely on protective structures whose perimeter is narrower than the confidence invested in them — [Related article: What Limited Liability Actually Excludes] examines the same misplaced confidence in the corporate form.
The Asset That Does Not Appear in Your Accounts
Under AASB 138, and IAS 38 internationally, internally generated brands, mastheads, publishing titles, customer lists and similar items may not be recognised as intangible assets. [FACT CHECK REQUIRED] The reasoning is that the expenditure creating them cannot be separated from the cost of developing the business as a whole. Brand value reaches a balance sheet when it is acquired, not when it is built. That acquirers pay far above net tangible assets shows markets price brands, not that your accounts do.
The governance consequence is the point. The asset most exposed by outsourcing carries a value of nil: no impairment test, no capital allocation and, in most organisations, no named owner at executive level. Assets with a book value attract governance automatically — a register, a custodian, an insurance line, a board question when the number moves. Assets without one attract governance only when somebody decides to create it, a decision rarely made whose absence is invisible because nothing in the reporting appears to be missing.
The remedy is to build what the accounts will not: a named executive owner, a register of rights by jurisdiction with renewal dates, an annual review of coverage against where goods are made and sold, and a defence budget that exists before it is needed.
What the Instruments Can and Cannot Do
Legal instruments protect prospectively and almost never retrospectively. A confidentiality agreement signed after a disclosure does not restore confidentiality: the information's status was fixed when it left the building. An application filed after a mark or design has been published may be compromised by that publication or by a prior applicant. [FACT CHECK REQUIRED] In each case the instrument could protect, and was deployed where it could not.
Ordinary commercial process defeats it. Selection requires technical evaluation, evaluation requires disclosure, and disclosure happens across meetings with several candidates, most of whom will not be appointed. By the time counsel papers the arrangement, the most consequential disclosure has been made.
Three rules follow, none difficult once somebody owns them. Confidentiality obligations bind before technical evaluation begins, with every candidate, not only the winner. Ownership of improvements is settled before development starts, because a competent manufacturer will improve your process, and silence there becomes contested exactly when you want to leave. Registrations are filed before the product is shown, in jurisdictions of manufacture as well as sale.
Even done perfectly, all of it resolves to one mechanism: detect a breach, establish it, obtain a remedy in a forum with jurisdiction over a defendant holding assets. Each step carries a cost and a probability below one, and those probabilities multiply. The alternative is to arrange the counterparty's economics so the behaviour you want is what they would choose anyway — [Related article: Restructure the Payoff and Let Self-Interest Do the Enforcing] develops that mechanism, which belongs in this conversation and almost never is.
Decision Framework
Test the arrangement on recoverability rather than protection.
| What is transferred | Recoverable | What determines recovery | Where to act |
|---|---|---|---|
| Documented specification | Yes | Whether we hold the controlled copy | Version control held by us |
| Tooling and fixtures | Conditionally | Title, location, jurisdiction | Right of removal; drawings in escrow |
| Process settings and know-how | Rarely | Whether our people are present | Resident engineering presence |
| Yield and learning curve | No | Where production occurs | Dual-source, or accept it knowingly |
| Upstream supplier relationships | Rarely | Who holds second-tier contracts | Contract directly with critical sub-tiers |
| Capability to compete with us | Never | Nothing | Price it in at approval |
Two thresholds are worth adopting. If the honest answer to "how long to resume elsewhere?" exceeds what your customers would wait, you hold not a supply arrangement but a single point of failure with a contract attached. And if the only evidence your rights are enforceable is a certificate, verify the enforcement precondition in each jurisdiction before relying on it. [FACT CHECK REQUIRED]
From Strategy to Execution
The immediate work is an inventory: what has been transferred, to whom, under which instrument, and on what date relative to the disclosure it was meant to cover. The dates are the finding. Alongside it, map registered rights by jurisdiction against the manufacturing and sales footprint, gaps named rather than implied.
The medium-term work is capability. A resident engineering presence where production happens is the most effective mechanism for retaining process knowledge, and invariably the first cost cut once the arrangement is judged to be working. A modest internal capability — a pilot line, a prototyping capacity — buys the ability to resume, to specify credibly, and to tell whether a supplier's quoted constraint is genuine.
The long-term positioning is a standing decision about which process steps stay inside the enterprise, because that is where advantage is created and learning compounds. Those steps are non-outsourceable regardless of unit economics — a strategy decision, not a sourcing one, recorded as such, because sourcing decisions get made repeatedly by people who were not in this conversation.
Signals to Monitor
Watch the direction of technical initiative. When process improvements originate with the manufacturer and arrive as proposals rather than instructions, capability has already moved. That is not misconduct but the predictable result — and the earliest reliable indicator available.
Watch for your product's characteristic features appearing in adjacent categories or markets you do not serve. Treat a lapsed registration as an incident, not an oversight. Watch what happens when the manufacturer's supplier changes a material: if you learn afterwards, you no longer control the specification you believe you own. And watch for proposed reductions in your on-site presence — decisions about recoverability made without anyone recognising a decision.
Questions for the Leadership Team
- If our principal manufacturing relationship ended without cooperation in ninety days, what could we produce, where, and how long would restoration take — and has anyone verified that rather than estimated it?
- In which jurisdictions are our marks and designs registered, and does that map include everywhere our goods are made? Where it does not, was that a decision or an omission?
- For each significant supplier disclosure, was the confidentiality obligation signed before or after? What did we disclose during selection to candidates we did not appoint?
- Who at executive level owns the brand and registered rights as assets, given that our accounts assign them no value and no custodian?
- Which protections have we assumed rather than verified — and if verification came back negative, what would we do differently tomorrow?
Closing Perspective
The instinct is to strengthen the instruments: a firmer agreement, wider registration, a budget for enforcement. Those are worth having, and they address the smaller half of the problem. The larger half is that capability accumulates wherever the work is done, and no document reverses that.
The choice is not how much protection to buy but how much irreversibility to accept knowingly, in exchange for the cost and capital outsourcing releases — and then to spend enough on the recoverable parts to keep the irreversible ones bounded.
An organisation that has made that trade deliberately can live inside it for decades. One that believes its agreements made the trade unnecessary discovers its real position only when the relationship is already ending — the one moment at which nothing can still be changed.
About EraNorth Insights
EraNorth Insights publishes practical analysis on strategy, projects, operations, transformation and decision intelligence for professional and organisational use. About EraNorth.