# Continuous Market Testing: Why Business Evolution Must Never Stop

> Continuous market testing helps businesses detect changes in demand, technology, channels, and economics before a successful strategy becomes obsolete. The history of McDonald's salads shows why failed tests are conditional evidence, not permanent verdicts.

- Canonical page: https://mtfinstitute.com/insights/continuous-market-testing-business-survival/
- Content type: Article
- Editorial category: Articles & Analysis
- Publisher: MTF Institute of Management, Technology and Finance
- Author: MTF Institute Editorial Team- Published: 2026-07-26
- Updated: 2026-07-26
- Language: English
- Topics: Entrepreneurship, Strategy, Innovation, Change Management, Market Research, Business Transformation

## Continuous Market Testing: Why Business Evolution Must Never Stop

A successful business cannot treat strategy as a permanent answer. Customer behavior, technology, distribution, regulation, costs, and competitors keep changing. Continuous market testing gives leaders a disciplined way to detect those changes early, run affordable experiments, and evolve before a once-successful business model becomes a liability.

The goal is not to chase every trend. It is to make market sensing and business experimentation part of normal management.

## What is continuous market testing?

Continuous market testing is a recurring management process for:

1. monitoring changes in customers, channels, technology, competitors, and economics;
2. converting signals into explicit business hypotheses;
3. testing those hypotheses with limited cost and clear success criteria;
4. scaling, adapting, pausing, or ending initiatives based on evidence; and
5. retesting important assumptions when market conditions change.

This definition matters because a failed experiment is not an eternal verdict. It answers a narrower question: did this offer, for this audience, through this channel, at this price and time, work well enough to continue?

The history of McDonald's salads is a useful illustration.

## The McDonald's salad story: what actually happened?

The half-remembered version of the story is simple: McDonald's tested salads in the 1980s, customers rejected them, strategists criticized the company for moving beyond burgers, and salads later became a major success.

The documented history is more interesting.

| Period | What happened | What the evidence means |
|---|---|---|
| 1984-1986 | McDonald's tested packaged salads in selected U.S. markets. By April 1986, the test covered about 1,000 restaurants in six markets. | The company was exploring demand beyond its traditional burger occasion. |
| 1987 onward | Salads entered the broader U.S. menu, but performance varied across formats and periods. | An initial rollout did not settle the question permanently. |
| 1999-2000 | McSalad Shakers were tested and launched in a portable cup format. | McDonald's changed the delivery format, not just the ingredients. |
| 2003 | Premium Salads launched as meal-size products. McDonald's reported selling approximately 150 million during the year. | A category that had struggled could work when timing, product design, positioning, and operations changed. |
| 2013 | Then-CEO Don Thompson said salads accounted for only 2-3% of sales. | A successful launch did not guarantee durable strategic importance. |
| 2015-2016 | McDonald's changed its lettuce blend after customer feedback and continued local product testing. | The company kept refining the offer rather than treating it as fixed. |
| 2020 | Salads left the national U.S. menu during operational simplification. | Complexity, demand, speed, waste, and supply-chain economics matter alongside consumer interest. |
| 2024-2026 | McDonald's USA leadership said customers were not looking to the chain for salads, and the July 2026 U.S. menu still has no salad category. | The current U.S. decision is to focus resources elsewhere. |
| 2026 in Portugal | McDonald's Portugal lists mixed, tuna-and-pasta, feta-and-pasta, and selected island salad options. | The correct decision can differ by geography and customer context. |

Sources include [contemporary reporting on the 1980s tests](https://www.washingtonpost.com/archive/business/1986/04/10/mcdonalds-tosses-up-a-salad/4a37a920-d7c2-43b0-a2e5-dd6c1042db84/), McDonald's reporting on [the 2003 Premium Salad launch](https://www.aeca.es/old/comisiones/rsc/biblioteca_memorias_rsc/memorias_empresas_extranjeras/mcdonalds.pdf), its account of [later customer-led product changes](https://corporate.mcdonalds.com/corpmcd/our-stories/article/new_salad_blend.html), the current [U.S. menu](https://www.mcdonalds.com/us/en-us/full-menu.html), and the current [Portuguese salad range](https://www.mcdonalds.pt/produtos/saladas).

The conclusion is not that McDonald's should always sell salads. It is that management should keep asking whether salads—or any other offer—fit the current market, operating model, brand, and economics.

## A failed business test is conditional evidence

Executives often interpret test results too broadly.

- A weak test becomes "customers will never want this."
- A strong test becomes "customers will always want this."
- A failed channel becomes "the product has no market."
- A successful launch becomes "we have found a permanent growth engine."

Each conclusion removes context from the evidence.

A business experiment should be recorded as a set of conditions:

> For customer segment A, in market B, at price C, through channel D, with product version E and operating model F, the result was G during period H.

When one of those variables changes materially, the old conclusion may no longer apply.

McDonald's salads worked differently across the United States and Portugal. Premium Salads in 2003 did not produce the same result as salads a decade later. The 2020 decision also included an operational context: restaurants were simplifying menus during a pandemic.

The lesson is strategic humility. Data is essential, but data always comes from a particular environment.

## Strategy requires both consistency and experimentation

The apparent conflict between "stay focused" and "keep experimenting" is false.

A company needs a stable core: the customer promise, capabilities, operating standards, and economic logic that make it distinctive. It also needs an exploratory edge where assumptions can be tested without destabilizing the core.

James G. March described this as the tension between [exploitation and exploration](https://pubsonline.informs.org/doi/abs/10.1287/orsc.2.1.71). Exploitation improves what the organization already knows. Exploration searches for new possibilities. Too much exploration creates noise and waste; too much exploitation can make a company highly efficient at serving a market that is disappearing.

McDonald's current strategy illustrates the balance. Its [2023 annual report](https://corporate.mcdonalds.com/content/dam/sites/corp/nfl/pdf/2023%20Annual%20Report_vf.pdf) emphasizes commitment to familiar core products while also describing the ability to test and scale innovations as customer preferences evolve. Its [2025 annual report](https://corporate.mcdonalds.com/content/dam/sites/corp/nfl/pdf/MCD%202025%20Annual%20Report.pdf) describes customer-insight-led innovation and a U.S. beverage pilot across more than 500 restaurants.

Focus defines where the company plays. Experiments reveal how the game is changing.

## Why famous disruption stories are often oversimplified

Polaroid, Nokia, BlackBerry, and Kodak are frequently presented as companies that simply ignored change. Their histories are more complicated, and that complexity produces better management lessons.

### Polaroid: demand can return in a different cultural context

Polaroid says the rise of digital technology led it to cease instant-film production in 2008. Yet enthusiasts saved the last factory through the Impossible Project, enabling the later rebirth of the Polaroid brand.

The original mass-market logic had weakened, but a new combination of nostalgia, physical experience, creative identity, and community support made instant photography meaningful again. A category can decline as a mainstream utility and return as a differentiated experience.

### Nokia: losing a category is not the same as disappearing

Nokia did not vanish. In April 2014 it [sold substantially all of its Devices & Services business to Microsoft](https://www.nokia.com/newsroom/nokia-completes-sale-of-substantially-all-of-its-devices--services-business-to-microsoft/) and continued in network infrastructure, technology development, and licensing.

The smartphone disruption was devastating to Nokia's handset leadership, but the company itself reconfigured around other capabilities. The lesson is not merely "notice the trend." Leaders must decide which capabilities can migrate, which assets should be sold, and where the organization can still earn an advantage.

### BlackBerry: experimentation must reach the business model

BlackBerry states that its transition from smartphone manufacturer to software company was complete in 2016. It ended internal hardware development, moved device production to partners, and focused on software and secure communications. Legacy BlackBerry operating-system services ended in 2022.

BlackBerry did adapt, but adaptation did not preserve its former smartphone position. Detecting change is only the first step. Timing, platform ecosystems, product execution, developer support, switching costs, and unit economics determine whether a response can win.

### Kodak: revival is real, but the mechanism matters

Kodak is neither the vanished company of popular mythology nor simply a restored consumer-camera giant. The company said its film business was growing in 2024 and that it was investing in capacity. It reported $1.069 billion in revenue for 2025 across its current businesses.

Kodak-branded instant-print cameras also exist, but [Kodak's own product page](https://www.kodak.com/en/consumer/product/cameras/instant-print/printomatic-instant-print-camera/) states that C+A Global manufactures and sells the Printomatic under license. This is adaptation through brand licensing and partnerships, not a simple return to the old vertically integrated model.

The strategic point is important: evolution may mean a new product, a new audience, a licensing model, a partnership, a divestiture, or a new definition of the core business.

## How to build a continuous market-testing system

Occasional brainstorming workshops are not enough. Continuous experimentation needs an operating system.

### 1. Create a market-signal radar

Track signals in five groups:

- **Customer behavior:** purchases, retention, complaints, search intent, usage patterns, and unmet needs.
- **Distribution:** acquisition cost, channel reach, algorithm changes, partner performance, and sales-cycle length.
- **Technology:** new capabilities, falling production costs, substitutes, standards, and infrastructure.
- **Competition:** new entrants, pricing, positioning, business models, and adjacent-category moves.
- **Economics and environment:** margin, input costs, regulation, demographics, labor, and macroeconomic pressure.

Assign owners and a review cadence. A dashboard without a decision meeting is only decoration.

### 2. Maintain a hypothesis register

For every material assumption, record:

- what the organization believes;
- the evidence supporting it;
- what change would make it doubtful;
- the next test;
- the decision threshold;
- the owner and review date.

This prevents forgotten assumptions from becoming invisible doctrine.

### 3. Protect an exploration budget

Core operations will usually win every resource debate because their returns are immediate and measurable. Exploration benefits arrive later and are uncertain. Leaders therefore need to reserve money, people, data access, and management attention for tests.

There is no universal percentage. The appropriate allocation depends on industry volatility, cash position, regulation, and the cost of experiments. What matters is that exploration is intentional rather than funded only when spare capacity appears.

### 4. Test the smallest decision, not the biggest dream

An experiment should reduce uncertainty before the company makes a larger commitment.

Examples include:

- a landing page before a full product build;
- a limited geographic rollout before a national launch;
- a manual service before workflow automation;
- a partner channel before an internal sales team;
- a time-limited menu item before permanent operational changes;
- a pricing test before redesigning the entire offer.

[Discovery-driven planning](https://hbr.org/1995/07/discovery-driven-planning) applies the same logic: convert assumptions into milestones and release resources as evidence improves.

### 5. Measure customer value and business economics together

A popular experiment can still be a bad business. Track:

- adoption and conversion;
- repeat use and retention;
- willingness to pay;
- gross and contribution margin;
- acquisition and service cost;
- operational complexity;
- cannibalization of existing revenue;
- brand and customer-experience effects.

McDonald's salad story cannot be understood through unit sales alone. Fresh ingredients, preparation, waste, kitchen speed, supply-chain requirements, and the ability to attract a group that might otherwise veto the restaurant all affect the decision.

### 6. Define scale, adapt, stop, and retest rules before launch

Every test needs four possible outcomes:

- **Scale:** evidence meets the threshold and operations can support growth.
- **Adapt:** demand exists, but the product, price, channel, or process needs revision.
- **Stop:** the hypothesis is not supported under current conditions.
- **Retest later:** the result depends on a variable likely to change.

Predefined rules reduce the risk that senior enthusiasm keeps weak projects alive or that one disappointing result kills a valuable option forever.

### 7. Separate global learning from local decisions

Multimarket companies should share evidence globally but decide locally where customer behavior and economics differ.

The U.S. absence and Portuguese presence of McDonald's salads are not contradictory. They are evidence that standardization has limits. The reusable capability is not one universal menu; it is the ability to test, learn, localize, and operate each choice economically.

### 8. Review the portfolio, not only individual experiments

Leaders should ask:

- Are we only improving current products?
- Are we testing new channels and business models?
- Which old failures deserve retesting because conditions changed?
- Which successful products now show weakening signals?
- Are experiments concentrated in one comfortable area?
- Can the organization scale a successful test without damaging the core?

This portfolio view connects experimentation to strategy rather than treating it as a collection of isolated ideas.

## How often should a business test the market?

There is no universal calendar, but the cadence should match the speed of change and the reversibility of decisions.

- Digital acquisition, pricing, and messaging can often be reviewed weekly or monthly.
- Product and customer-segment assumptions may be reviewed quarterly.
- Major capability, geography, and business-model choices may need formal semiannual or annual reviews.
- A major external change—new regulation, technology, competitor, channel policy, or cost shock—should trigger an immediate review.

The most important rule is continuity. Trend monitoring should happen before a crisis makes change unavoidable.

## The real lesson: build the capability to change your mind

The best decision is not permanent consistency and not permanent novelty.

It is the ability to:

1. defend a profitable core;
2. notice when its assumptions are weakening;
3. test alternatives without betting the company;
4. learn from both positive and negative results;
5. revisit old conclusions when conditions change; and
6. scale only when demand, operations, distribution, and economics align.

Dynamic-capabilities research frames advantage in changing environments as the ability to identify opportunities and reconfigure organizational capabilities. That is what continuous market testing makes operational.

McDonald's did not prove that salads are good or bad. It demonstrated that the answer can change by year, format, country, customer occasion, and operating system.

A resilient company is not one that predicts every trend. It is one that keeps producing evidence before the market produces a verdict.

## Build the executive capabilities behind continuous evolution

Continuous market testing is cross-functional. Leaders must connect strategy with finance, marketing, product decisions, operations, technology, AI, people, and change management. A promising experiment can fail when any one of those disciplines is missing.

The [Advanced Executive Program in Management & Business Administration](https://mtfinstitute.com/programs/advanced-executive-management-business-administration/) from MTF Institute is designed for managers, founders, and specialists who need that integrated view.

The 100% online, self-paced professional program includes approximately 480 hours across eight pillars: leadership and strategy; financial acumen; marketing and customer centricity; operations and supply chains; digital transformation and AI; human capital; an integrated strategy capstone; and executive career development.

Its capstone requires learners to connect market analysis and business validation with financial modelling, go-to-market planning, operations, and a board- or investor-style presentation—the same integration required to turn experiments into strategic decisions.

[Explore the curriculum, current tuition, and enrollment terms](https://mtfinstitute.com/programs/advanced-executive-management-business-administration/).

## Frequently asked questions

### Why is continuous market testing important?

Continuous market testing helps a company detect changes in demand, technology, channels, competition, and economics before those changes undermine the existing business model. It replaces permanent assumptions with regularly updated evidence.

### Does a failed market test mean the idea is bad?

No. It means the tested version did not meet its threshold under specific conditions. The customer segment, timing, price, channel, geography, product design, or operating model may change the result.

### Should a company focus on its core or experiment with new products?

It should do both. The core funds the business and protects its distinctive promise. A controlled exploration portfolio tests how the market is changing and creates options for future growth.

### What should a business experiment measure?

Measure customer response and business viability together: conversion, retention, willingness to pay, margin, acquisition cost, service cost, operational complexity, cannibalization, and brand effects.

### When should an old failed idea be tested again?

Retest when a material assumption changes—for example customer behavior, technology cost, distribution access, regulation, product format, brand position, or geographic market.

### Are McDonald's salads available in 2026?

They are not part of the current national U.S. menu. McDonald's Portugal currently lists a salad category and several salad products, demonstrating how menu decisions can differ across markets.


## Citation

When citing or summarizing this material, link to the canonical HTML page: https://mtfinstitute.com/insights/continuous-market-testing-business-survival/