What is it that explains the sudden turn in fortune of these Chinese automakers in such a short period, in an industry that is more than a century old?

It has happened because, as we have argued in previous Fortune articles, the rules of competition have changed as the world fragments from three disruptive forces: the fracturing of the geopolitical consensus; the onward march of digitization,;and the rise of “deep tech” innovation. And the consequences for traditional global companies are profound: The power is shifting to local consumers; the global playground is becoming a multitude of fast-evolving local “mini-playgrounds”; and the winners are those local companies, even newcomers, who meet the differentiated needs of each mini-playground with speed, responsiveness, and innovation. 

We call this new basis of competitiveness “fractal advantage.” Global scale and pedigree are no longer guarantees of success in this new world. 

It was not that global players were too late with their EV models in the fast-growing Chinese market. But the Chinese EV companies were first to the market with a disruptive, superior solution to the needs of the digitally sophisticated Chinese customers. This calls for a rethink of the traditional innovation proces. And without the traditional mind-set and engineering legacy of their global rivals to hold them back, the Chinese players pivoted quickly to this new approach we call fractal innovation.

For global incumbents, the innovation process that produced winners in the last century could well become the biggest roadblock to success in the present one—unless they master fractal innovation, and change not just how they innovate, but also what they create.

Creating fractal customer solutions, not global products

Today, global companies create global products for an idealized global consumer. They do this by adopting what we call a “product-out” approach: A centralized R&D team converts customer need analysis into new products that are designed to satisfy the averaged needs of the global customer without breaching certain performance, quality, and importantly scale-driven cost and profit-margin thresholds.

By contrast, fractal companies create customer solutions that are local, customized, even personalized, and focused not only on meeting the current needs but also on anticipating the future needs of customers. They do this by adopting what we call a “customer-in” approach: They consider the entire customer journey (which starts long before and ends long after the purchase of the core product) and they draw on the local customer’s unique and real-time data.

It sounds simple. But it’s not: It leads to a complex set of design choices. To explain what we mean, let’s continue with the example of the electric vehicle (although there are many products that show a similar pattern of development). 

Originally, cars were simply physical products designed to move from Point A to Point B. Then, when they were built with software to improve functionality and performance, they were transformed into “computers on wheels.” Now the most modern cars—and especially electric vehicles—are becoming “smartphones on wheels”: complete “solutions” offering customers an ever-growing range of digital micro-services for personalizing their driving experience. But it is only those solutions that are designed for cost-efficient localization to markets, personalization to customers, and adaptation to future needs that become the winners. The Chinese EV players are winning because they have designed a better solution for Chinese customers. 

Four ways to build a fractal innovation process

At first glance, it would seem to be an impossible task to combine cost-efficient global design with costly local customization. But the fractal innovation process allows companies to square the circle. To follow their example, traditional global companies will need to make four critical alterations to how they innovate.

1. From the ‘fixed’ product to the ‘versatile’ solution

Cars used to be designed as integrated “fixed” product structures. Now, EVs are designed as “versatile” three-tiered structures—with a hardware layer, an embedded software layer (sometimes called “intelligent hardware”), and a micro-services layer. Each layer is designed in such a way as to allow for quick, cost-efficient and independent (that is, without having to change the design of the other layers) “manipulation” using three design levers—modular (for the hardware), cloud (for the software), and API-based design (for the micro-services).

With such a flexible structure, designers can make decisions about which parts should be designed for global scale and which for speedy localization, customization and adaptation. Also, by integrating cloud technology with API-based design, they can collaborate with external and local market-based designers who are close to customers and ensure that the functionality and applications can be updated frequently, instantaneously, and cost-efficiently. 

2. From product designers to solution architects

Given the complexity, making the right design choice is not easy and calls for a new class of designers called “solution architects.” These are experts who have design thinking and diplomatic skills to mediate between the often conflicting needs of technology and quality-driven R&D, the product management team that wants to meet every need of every local customer, and the finance team that wants to minimize costs. 

These experts must decide on things like: Where should the processing intelligence that drives the functionalities and, ultimately, the revenues be put? Should it be in the hardware, or in the embedded software, or in the cloud? How much should the company design for current versus future needs? For example, should microchip capacity be sufficient only to meet current processing needs (and thereby make the product cost efficient) or should it include excess capacity to meet future, sometimes unknown, processing needs (which would decrease the risk of early obsolescence but increase the upfront costs)? The extent to which companies answer these questions better than their competitors determines the winning potential of their customer offer.

3. From closed design to open innovation

To deliver these kinds of versatile solutions, global companies must cultivate a new corporate mindset. Many traditional companies have a “closed” approach to innovation: They decide what their customers need and try to design or control everything. By contrast, fractal companies are “open” innovators: they work closely with customers to identify in real-time any “friction” points along the customer journey and any emerging customer needs. Then, they run “live” design experiments by drawing on an army of external innovators to design new services, functionalities, and applications. 

Open innovation has been a key to the success of the Chinese EV automakers. For example, SAIC Motor, the largest Chinese carmaker, partnered with Alibaba, the Chinese digital leader, to develop Banma, an open infotainment platform. Today, this is the default internet solution for Chinese EVs and, as SAIC says, shifts the car’s role “from simple travel tool to intelligent mobility terminal.” To replicate this kind of success, traditional global companies will need to overcome their fear of losing control (particularly in relation to intellectual property and quality issues), partner with other companies and customers, and decentralize their design teams by sending key personnel to key markets.

4. From global gross margins to return on capital employed

Finally, global companies need to come to terms with the new economics of fractal innovation. Generally, they have a simple metric for any new product: Does it deliver the gross margin target expected by the financial market? Again, let’s look at the car. To hit the target, the central R&D team tries to maximize economies of scale by designing a globalcar with around 90% of the fixed cost spread over the projected number of vehicles sold around the world. The problem is that this approach makes it difficult, if not impossible, to design fractal solutions that square the circle by being localized for different markets, flexible and innovative enough to meet difficult-to-anticipate future needs, and marketable at lower price points. 

Fractal companies solve the problem by taking account not only of gross margins but also of another target: the return on capital employed (RoCE). Utilizing the layered, modular structure, they design not for the global average but for a de-averaged set of eight to 10 “local-design” markets. While this does limit the fixed cost that can be globally scaled to about 70% or 80%, it also allows the companies to capitalize on lower localized development costs and thereby increase the RoCE, introduce more competitive pricing, and boost the chances of higher sales volumes. 

It’s time to change

The global competitive landscape has changed. Decisively. To win, global companies can no longer afford to offer an average global product using the traditional innovation process, however well-designed. Instead, they should design for fractal solutions to meet fast changing local needs. Doing this cost-efficiently isn’t easy—but it will be the difference between winning and losing in today’s fast changing, fragmented world. 

Arindam Bhattacharya is a senior advisor at Boston Consulting Group. He is a former managing director and senior partner at BCG and a cofounder of the BCG Henderson Institute. He is also the former head of BCG India.

Hans-Paul Buerkner is a managing director and global chair emeritus at BCG, and the former CEO and chairman of BCG.