In my work with retail leadership teams, I keep hearing versions of the same frustration: We can see the disruption coming. We just cannot move fast enough to respond. That is not primarily a forecasting problem. It is an optionality problem. We have published a report on optionality and invite you to download it for free here. We also offer you a retail optionality stress test, Are You Efficient—or Trapped? After you have read the report, select one major business commitment and take the test. The assessment will show whether your organization can recognize when conditions have changed and respond before the damage reaches the P&L, or whether the business has optimized away the very flexibility it now needs. The point is not a perfect score. It is a better executive conversation about assumptions, alternatives, authority, funding, and the real cost of being unable to change course. Take the stress test.
What’s a proven competitive edge in a disruptive marketplace? And the answer is: Optionality as a strategic advantage.
The Speed of Change
Retail has spent decades removing slack from the operating model. Fewer suppliers. Bigger buys. Longer leases. Leaner inventory. Standardized store prototypes. Centralized technology platforms. Every line item inspected, consolidated and negotiated until the company looked beautifully efficient on a spreadsheet—and occasionally less so in real life. Then, the world declined to cooperate.
Consumer preferences now move faster than capital plans. Trade policy can change product economics before inventory clears the port. A store may need to shift from selling space to fulfillment hub to service center before the lease reaches its first renewal. AI can detect changes, generate scenarios and recommend action faster than many organizations can approve a meeting to discuss them. The annual plan may still run twelve months. The customer apparently missed that meeting.
Salesforce reports that agentic search at the beginning of shopping journeys increased 200 percent year over year. Yet only 28 percent of commerce organizations currently use agentic AI, while another 44 percent expect to adopt it within six months. That mismatch matters to boards and C-suites because optionality often requires decisions that only they can authorize. It may mean paying a modest premium for an alternate supplier, reserving capacity that is not fully utilized, running two operations in parallel, staging capital or accepting a slightly higher unit cost in exchange for dramatically lower trapped cost.
The board’s job is not to choose the backup supplier. It is to decide whether the company is willing to pay to have one before the primary supplier fails.
Efficiency Works Until It Becomes Fragility
The pursuit of efficiency is not irrational. Large buys reduce unit cost. Supplier concentration improves purchasing leverage. Standard prototypes accelerate construction. Long leases can secure favorable economics. Lean inventory frees working capital. But every efficiency decision contains an assumption: Demand will materialize. The supplier will remain reliable. The channel will keep performing. The store format will stay relevant. The technology will continue meeting the need. Reversing the decision will not be necessary.
The trouble begins when the assumption expires before the commitment does. Gartner found that 72 percent of supply-chain leaders revisited final approvals for major network decisions at least once, and more than half revisited them three or more times. Gartner recommends that business cases account for the cost of recurring turbulence, network diversification and moving capital later. It also recommends defining how long the assumptions behind a strategy are expected to remain valid—and establishing what would trigger management to stop, reverse or repurpose the investment.
That is an unusually honest prescription for modern planning. The goal is not to pretend the first decision will remain perfect. It is to make the cost of being wrong visible before everyone becomes professionally, financially and emotionally attached to the answer. Optionality creates complexity. Multiple suppliers require coordination. Smaller initial orders can raise unit cost. Flexible leases may cost more. Modular technology requires governance.
Optionality is not free. Neither is being trapped. The question is whether the premium required to preserve flexibility is smaller than the cost of unwinding a decision after its assumptions change.
Optionality Is Not Indecision
Corporate strategy has long recognized the idea of a real option: Management retains the right, but not the obligation, to defer, expand, contract, switch or abandon an investment when new information becomes available. Retailers create option value when they can reduce an inventory commitment, move product across channels, add or exit a supplier, resize a store, exercise a lease right, replace one technology component without rebuilding the stack or stage capital rather than committing it all before the evidence arrives.
This does not mean keeping every choice open forever. Keeping every door open is how you heat the whole neighborhood.
Smart optionality means making a clear primary decision, identifying the assumptions behind it and preserving a practical alternative when the downside of being wrong is significant. Management must define the trigger, owner, authority and funding required to act. Without a trigger, an option is not a strategy. It is a souvenir from the strategy meeting.
AI Can See the Turn
“The turn” is the point at which evidence shows that the assumptions behind a major commitment have materially changed. Demand is moving. A supplier is becoming risky. Inventory is accumulating in the wrong channel. A store needs to perform a different job. A platform is no longer delivering the expected value. AI can help the business see the turn sooner. It can model scenarios, identify supply risk, improve inventory visibility, flag underperforming promotions and monitor the assumptions behind major investments.
Optionality determines whether the company can make the turn. A retailer can identify a demand shift perfectly and still be unable to respond because the merchandise commitment cannot change, the supplier agreement provides no alternate path, inventory cannot move across channels, the lease does not permit another use, or the organization requires six approvals and a ceremonial sacrifice before acting.
Deloitte found that 30 percent of surveyed retailers already use AI for supply-chain visibility, with that figure expected to rise to 41 percent. Fifty-nine percent anticipate a positive return from AI-driven supply-chain initiatives within twelve months. Yet Gartner reported that 67 percent of supply-chain digital investment was being allocated to AI while 55 percent of chief supply-chain officers remained unclear about the return.
That is the spend-first, define-value-later school of innovation. AI can tell the retailer that the forecast is wrong. Optionality determines whether it can do anything about it. Knowing you are trapped sooner is useful. It is not the same thing as having an exit.
The Five Optionality Operating Forms
Optionality does not always mean expensive redundancy or a warehouse full of “just in case” inventory. In many cases, it is built into the product, contract, network, sequence or supplier relationship.
There are five practical forms:
- Wait: Delay an irreversible commitment until better information arrives.
- Switch: Move to a qualified supplier, component, geography, route, channel, technology model or partner.
- Scale: Expand or contract capacity, labor, inventory, fulfillment, space or capital without rebuilding the operating model.
- Reroute: Redirect product, orders, supply, fulfillment or customer access across channels, locations or markets.
- Repurpose: Convert an asset, facility, supplier, store, distribution center, workflow or technology capability to another use.
An option is not real because it appears in a contingency plan. It must be visible, qualified, available, executable and fast enough to matter. Otherwise, it is just a comforting noun.
Where Optionality Shows Up
· Merchandise and Supply: Pay for Response Time
Retailers have historically rewarded merchants for buying more at a lower unit cost. That math works until demand changes, the trend passes, or the product lands in the wrong market. Nobody receives a medal for achieving the lowest unit cost on inventory the customer does not want. In my experience, they earn an exit.
AI can improve demand sensing and inventory positioning. It cannot renegotiate a rigid purchase commitment after the merchandise is already on the water. The analytical capability and the commercial option have to be designed together.
The cheapest product can become the most expensive once markdowns, storage, transfers, cancellations, air freight, tariffs and missed selling time are included. The supplier savings look less thrilling after the emergency logistics bill and the apology tour.
Inditex offers a current example of a retailer that treats responsiveness as part of its integrated operating model. In fiscal 2025, it reported a 58.3 percent gross margin and 6 percent net income growth while continuing to open, enlarge, refurbish and absorb stores. Those results do not prove flexibility alone caused the performance. They do show that scale and adaptability are not opposites.
· Channel Strategy: Distribution Is Not a Religion
Retailers have a history of falling in love with the channel strategy of the moment. First stores were everything. Then ecommerce was the future. Then direct-to-consumer became the preferred path to higher margins and customer data. Wholesale was treated as the less glamorous relative near the back of the family photo. The customer was not consulted about these organizational preferences.
AI has now arrived with another route to market. Retailers must weigh third-party traffic and personalization against commoditization, disintermediation and loss of customer data. Where should checkout occur? Does loyalty follow the customer into an AI-mediated transaction? What happens when one intermediary becomes too powerful?
Nike’s fiscal 2026 results illustrate why distribution should not become an ideology. Full-year wholesale revenue increased 6 percent while Nike Direct revenue declined 6 percent. That does not prove one channel decision caused the company’s wider challenges. It does show that channel mix must remain responsive.
Direct-to-consumer is not sainthood. Wholesale is not failure. AI-mediated commerce is not automatically progress. The goal is profitable customer access without casually giving away the customer.
· Real Estate: A Smarter Forecast Does Not Make a Rigid Lease Flexible
CBRE reported that U.S. retail availability remained at 4.9 percent in the second quarter of 2026, while asking rents increased 2.4 percent amid historically low construction. Scarcity can encourage long commitments. Sometimes that is right. But a low rent does not compensate for the wrong market, an inflexible box or a lease that assumes the store will perform one function for ten years.
AI can improve site analysis and capital modeling. It cannot retroactively negotiate a termination right, add an expansion option after the adjoining space is leased or make an inflexible prototype support a different customer journey. Renewal and termination rights, expansion and contraction options, sublease provisions, alternate-use flexibility and phased capital are strategic and financial assets—not legal details to inspect after the decision.
A store prototype should be a platform, not a tattoo.
· Technology: Avoid Tomorrow’s AI Hostage Situation
Technology is where AI creates some of retail’s greatest optionality—and greatest risk. Salesforce found that only 27 percent of commerce organizations had fully unified customer data. Among multichannel organizations, just 2 percent reported no significant failure points. One vendor. One implementation. One version of the truth. One impressive contract. One throat to choke. And eventually, one uncomfortable renewal conversation.
Technology optionality means retaining and moving data, connecting new capabilities through usable interfaces, changing models or providers where practical, auditing AI actions, maintaining human override and rollback and exiting a vendor without losing the intelligence accumulated inside it. An agent automating a rigid workflow may make it faster without making the business more adaptable. Congratulations, you have automated the trap.
The CFO Needs Four Numbers, Not One
Most capital-allocation processes reward the apparently certain answer. A conventional business case compares upfront cost, expected revenue, expense, payback and return. Those measures matter, but they can understate the value of changing course.
Every major commitment should separate four economics:
- Option premium: What does it cost to preserve the alternative?
- Exercise cost: What will it cost to activate it?
- Trapped cost: What happens if conditions change and no viable alternative exists?
- Time to exercise: Can the option be activated before the damage occurs?
Suppose a single-source component costs $10 while maintaining a qualified secondary source raises the effective cost to $10.20. The additional 20 cents should not automatically be labeled inefficiency. It is the option premium. The question is whether that premium is justified by the probability and consequence of needing the alternative.
This does not require fake precision. It requires making the trade-off visible. Saving 20 cents while creating ten dollars of trapped cost is not procurement excellence. It is a very tidy way to lose money.
The Board Mandate
Boards should ask not only which forecast management believes, but which alternatives management has preserved if that forecast is wrong.
- Are we funding better intelligence while eliminating the ability to respond?
- Which commitments carry the greatest trapped cost?
- Where have we mislabeled an option premium as inefficiency?
- Who has authority and funding to act before the damage reaches the P&L?
The board does not need to design the supply network, choose the AI model, or negotiate the lease. It does need to authorize the trade-offs that preserve maneuverability and hold management accountable for knowing when to use them. The annual planning cycle is a management process. It is not a law of nature. Retailers do not need a plan for every future. They need a business that can survive being wrong about one.
AI will not eliminate uncertainty. It will make changing conditions more visible, decisions faster and actions increasingly automated. That creates enormous value—but only for retailers that have preserved something worth acting with flexible commitments, usable data, adaptable formats, multiple routes to market and clear decision rights.
AI can detect the change. It can model the alternatives. It may even help execute the response. But it cannot conjure a backup supplier, a flexible lease or a technology exit path after the business spent years optimizing those choices away.
Better intelligence is useful. It is not a substitute for maneuverability. The winning retailer will not necessarily have the most accurate forecast. It will have the shortest, most affordable distance between intelligence and action. That is why optionality is not the enemy of efficiency. Optionality is the competitive edge.
Put the Premise to the Test
Download our full report here. Then choose one major commitment—an inventory buy, supplier relationship, lease, technology platform, channel strategy or capital investment—and take the Retail Optionality Stress Test. The goal is not to earn a perfect score. It is to determine whether your organization can recognize when the assumptions behind that decision have changed and respond before value is destroyed—or whether years of “efficiency” have quietly boxed you in.
Knowing you are trapped earlier is useful. Having a credible way out is considerably better.

