Think orthogonally

AI makes strategic ideas cheap. Campaigns, business models, partnerships, and product variants can be generated within minutes. Yet this speed creates a new problem: many answers sound different while resting on the same familiar patterns.

Subscription. Community. Platform. Personalization. Freemium. Influencers. Gamification. Any of these options may be useful. But when every team works with similar models, sources, and prompts, apparent creativity quickly converges on a polished middle.

The professional counter-question is therefore not: How do we generate even more ideas?

It is: How do we reach a perspective outside our industry routine without losing relevance, safety, and evidentiary discipline?

Orthogonal thinking does not copy someone else’s action. It extracts the action’s operating mechanism and rebuilds it for the problem at hand.

Today’s average strategy is exceptionally well written

Generative AI lowers the cost of articulation. That is progress, but it does not guarantee strategic originality.

Models often produce the options most accessible in their training data, the current conversation, and common business frameworks. The output can feel clear, complete, and professional while creating very little strategic distance.

The risk is not merely a bad idea. It is plausible convergence: many teams receive variations of the same answer and mistake linguistic variety for structural variety.

Orthogonal does not mean random

Orthogonal thinking is not unconstrained brainstorming. It changes the direction of observation while keeping the strategic question stable.

Suppose a project asks how a difficult-to-explain AI service can be sold credibly. The problem remains. What changes is the source of the solution logic: instead of examining only software companies, the team studies museums, festivals, public-health campaigns, games, logistics, or live demonstrations.

Distance should make new mechanisms available. It must not sever the connection to the real customer problem.

Why guerrilla marketing is a useful source domain

Guerrilla marketing often operates under three conditions familiar to early AI projects: limited budget, low awareness, and a crowded attention market.

Its strongest examples do not compensate for missing reach by becoming merely louder. They change the context in which people experience a message. A place becomes a medium. An action becomes admission. A competitor becomes a trigger. A product capability becomes a public demonstration.

That is why the field is strategically interesting. It provides not only campaign ideas but patterns for resource use, value exchange, distribution, participation, and proof.

The visible action is not the strategic core

What remains in memory from famous activations is often the spectacle: a vending machine, an unusual app rule, a provocative condition of entry. Copying that surface usually produces a weaker repetition.

The strategic core sits deeper: Which existing infrastructure was reframed? Which abstract statement became physically immediate? Which action connected attention with a relevant value? Which friction made participation more interesting than passive viewing? Which proof emerged during use?

Only these neutrally described mechanisms are transferable.

Mechanism before surface

A surface is what people see. A mechanism is what changes behavior or perception.

“An activation near a competitor” is a surface. “Use the competitor’s reach as a trigger for one’s own entry point” is a mechanism.

“A shocking vending machine” is a surface. “Translate an invisible problem into an immediate decision” is a mechanism.

“An unusual ticket” is a surface. “Turn a transaction into value-aligned participation” is a mechanism.

This distinction prevents creativity from becoming costume.

Analogical transfer depends on structural similarity

Research on analogical problem solving has long shown that people do not automatically transfer solutions from a source problem to a target problem. Recognizing the shared structure is especially useful; surface resemblance can be a distraction.^[1]

For strategy work, a festival and a B2B software product do not need to look alike. What matters is whether they perform a comparable task — for example, turning participation into identification or making an abstract benefit provable at the moment of decision.

Orthogonal research is therefore not an excursion into arbitrariness. It is a search for distant cases with a compatible causal structure.

The AI paradox: more creativity, less collective diversity

A study by Anil R. Doshi and Oliver P. Hauser published in Science Advances found in a writing task that generative AI could improve individually rated creativity and quality, particularly for less creative participants. At the same time, AI-assisted outputs became more similar to one another.^[2]

This tension matters for strategy. AI can help a team articulate more useful options, yet it can reduce collective variety if everyone relies on the same obvious cues.

The answer is not to remove AI from creative work. It is to design distance, source selection, and mechanism extraction explicitly.

Case 1: Whopper Detour — someone else’s footprint becomes your trigger

In “Whopper Detour,” users could unlock a one-cent Whopper through the Burger King app while near a McDonald’s location. The One Club award archive describes more than 14,000 included locations and reports strong app downloads and use.^[3]

The interesting lesson is not “offer a discount at a competitor.” The mechanism is:

Use the physical or digital presence of a dominant actor as the trigger for your own access point.

For an AI project, this could mean turning the moment of an established workflow, an export file, an industry event, or a widespread tool into the entry point for the new service — legally, transparently, and without manipulating someone else’s systems.

Case 2: UNICEF Tap Project — an abstract problem becomes immediate

UNICEF USA describes a “Dirty Water Vending Machine” activation associated with the Tap Project. A machine in New York offered bottles visibly labelled with diseases instead of beverage varieties. No one was meant to drink contaminated water; the installation made the distance between everyday water consumption and the lack of safe water physically immediate and close to a donation decision.^[4]

The mechanism is:

Translate an abstract problem into a concrete decision situation that is understood before it has to be explained.

This is particularly powerful for AI products. Instead of merely claiming error reduction, time savings, or data risk, a project can show a real before-and-after process, let people experience a decision path, or make uncertainty visible inside the workflow.

Case 3: Wacken — the transaction becomes value-aligned participation

Wacken Open Air linked blood donation with the opportunity to receive festival tickets in its official “Pay with your blood” campaign. The initiative started in 2024 and expanded to several German cities in 2025. The organizer reports that the available tickets were allocated within two weeks.^[5]

The transferable mechanism is not the dramatic wording. It is:

Turn the value exchange into an action that fits the community’s identity and serves a real social purpose.

For other projects, that action might be qualified data sharing, a community contribution, a documented learning step, a repair service, or a joint experiment. The action must create genuine value rather than merely functioning as an artificial hurdle.

Three cases, not three templates

None of these cases should be reproduced. Their effects depended on brand, timing, place, audience, partnerships, and execution.

What can be adopted is the reasoning operation: investigate a real action, separate visible form from operating mechanism, state the necessary conditions, translate the mechanism into several project areas, test the smallest safe version.

Quality does not come from resemblance to the source. It comes from the fit between mechanism and one’s own problem.

A small library of transferable mechanisms

The table is not a recipe book. Every mechanism requires fresh translations and its own boundaries.

Orthogonality can change every strategy surface

Treating guerrilla marketing only as advertising inspiration leaves most of its value unused.

Orthogonality is not a department. It is a second angle on strategic decisions.

Product design: the feature becomes the proof

Many AI products explain their value in slides before anyone experiences it. An orthogonal approach asks whether the feature itself can become the demonstration.

An analytics product might turn a small real data sample into a verifiable decision brief within minutes rather than merely showing a dashboard. A creative system could place original material, transformation, and quality review side by side. An agent could expose uncertainty as explicit approval points instead of hiding it.

The product does not become louder. It becomes more capable of proof.

Price and value exchange: money is not the only relevant action

A price remains a monetary amount. Yet the route to access can include other value-creating actions.

A pilot might receive a reduced price when the customer contributes structured process data, committed review time, or permission for a credible case study. A community edition might be tied to documented contributions or joint testing. A research offer could grant access in exchange for qualified feedback.

These models must not conceal labor or data extraction. Contribution, consideration, rights, and exit must be clear. Otherwise creative exchange becomes unfair extraction.

Onboarding and distribution: context becomes the channel

The most effective entry point often sits not on a new landing page but in a moment that already exists.

A project might begin where users already export a file, perform a quality check, write a brief, prepare a tender, or attend an event. The existing context supplies relevance; the new offer removes a visible friction exactly there.

This is different from unauthorized intrusion into external systems. The mechanism uses a legitimate handoff moment, not infrastructure without consent.

Partnerships: change roles instead of merely adding reach

Conventional partnerships combine two audiences or sales channels. Orthogonal partnerships also ask whether one partner’s resource can take on a new function in the shared model.

An association may provide not only reach but trusted curation of test cases. A university may contribute not only research but independent quality review. An event may serve not only as a stage but as a controlled live environment. A customer may become not only a buyer but a co-designer of a reference standard.

The new role must create transparent value for both sides.

Funding and stakeholder communication: proof instead of assertion

Conversations with investors, grant bodies, or internal decision makers can also be designed orthogonally.

Instead of delivering a long future narrative, a team can make one central assumption testable live: real raw data in, visible transformation, defined review, measured time, documented errors. Or it can present three deliberately different product decisions and let stakeholders experience the trade-off.

The unusual moment is not there for show. It compresses evidence and understanding.

Experience before advertising message

Strong unconventional strategies give people something to do, discover, compare, or decide. The message emerges from the experience.

This matters for AI projects because their promises are often abstract. “More efficient,” “smarter,” or “personalized” remains weak until a customer experiences a difference.

A good orthogonal option therefore answers: Which action makes the strategic value tangible without requiring ten slides of explanation first?

The transfer method starts with a stable question

Before searching for distant examples, the target problem must be fixed.

A useful framing includes the concrete decision, audience and use context, the central friction, the desired behavior or outcome, non-negotiable conditions, and the current default approach.

Example: “How can we prove the value of an AI review assistant to compliance leads during the first session, without uploading sensitive data or completing a full integration?”

This precision keeps later creativity attached to the problem.

Then search deliberately in distant domains

Research should run on at least two tracks.

Track A: near reality

Which solutions, prices, distribution patterns, and forms of proof are actually common in the project’s industry? Real examples and reliable sources define the baseline.

Track B: distant mechanisms

Which other fields solve a structurally similar task under different conditions? Festivals, games, museums, public campaigns, sports, logistics, hospitality, disaster response, and science communication may all be useful.

The tension between proximity and distance makes the transfer productive.

Formulate the mechanism as a neutral verb

A good mechanism statement avoids brand names and props. It describes a causal operation.

Weak: “We do something like Whopper Detour.”

Stronger: “We turn the moment when the audience uses an established alternative into a voluntary, legitimate entry point to our immediately comparable value.”

Weak: “We build a viral vending machine.”

Stronger: “We make a previously abstract consequence visible in a concrete decision situation and offer a directly relevant action.”

Verbs such as redirect, invert, embody, concentrate, reveal, exchange, couple, demonstrate help strip away the surface.

One mechanism needs several translations

The first idea is often a disguised copy. A team should therefore produce at least five translations per mechanism across different strategy surfaces.

“Immediate proof” might become a product feature that exposes its quality review live, pilot pricing tied to a measurable result, onboarding that begins with a real problem rather than a product tour, a partnership with an independent review role, or an investor conversation built around a reproducible mini-test.