Should your global firm centralize digital operations?

Recently, a heated argument divided the leadership at a large sporting goods retailer. The retailer had launched a digital transformation initiative, but no one could decide whether to centralize or localize their digital talent and activities. Some argued that digital activities should be put in just one central hub, pointing to examples like Booking.com, the $15B global leader in travel accommodations, which collocates over 1,700 developers in one location to optimize the tens of thousands of A/B tests that keep them in the lead. But others argued digital needs to be local, given the nature of retail, arguing that each country needed its own digital team to adapt to the local needs. They justified their arguments with examples from Amazon, which has prospered in the U.S., but struggled in countries, like the Netherlands, where they don’t have warehouses. Yet others suggested a middle ground of organizing digital operations around regions.

After a long debate, proponents of the local strategy prevailed. But they soon realized that many of the differences between countries were more imagined than real — and the fragmentation of their digital team made it hard to build the digital expertise needed to really compete. In the end they scrapped the local plan and centralized almost everything.

The dilemma at the sports retailer is a microcosm of the debate about digital transformation, all around the world, across different industries: Should leaders centralize their digital activities or leave them local? Each company has their own set of caveats: their ownership structure, their customer base, the demands of local regulation all shape how they make this decision. But how do you cut through this noise to the factors that matter most? In a cross-case comparison of 50 companies undergoing digital transformation, we identified five factors to help companies weigh the pros and cons. Understanding them can help your organization choose the right balance to get the most out of your digital transformation.

The major driver to centralize: economies of skill

As a general rule, our evidence suggests that, when possible, you should try and centralize everything digital. Why? Digital business requires specialized technical skills, as well as short development cycles empowered by high interaction. All of that is made harder by distance. Centralizing your digital activities enables critical specialization, faster learning, scalable IT solutions and overcomes digital talent shortage.

Economies of skill explains why digital first players like Uber or Booking.com are willing to move offices a dozen times if needed to allow everyone to be as close together as possible, learning from each other. Think about it like this: If you have one person in charge of online marketing in each country, they will likely also be in charge of other things as well, like PR or regular marketing. But if you bundle those 40 people together, suddenly they can start to specialize into critical roles such as UX conversion specialist, traffic specialist, CRM campaign lead, social marketer, SEO optimizers, and other roles you didn’t even realize were important. Just as importantly, when they are close together, they can learn from each other, and coordinate more easily in morning standup meetings, speeding up cycles of experimentation that power growth.

Finally, talent is hard to attract and retain. A lone marketer will likely feel lonely and disconnected, but when they are working with other likeminded people, they are more likely to stay engaged. In the fashion industry, conglomerates like LVMH thrive on building internal communities of practice that contain likeminded experts across a range of domains. When these experts have a sense of a community, they are more likely to stay and not move to competition.

The three drivers to localize

Even though it makes sense to centralize in general, there are three forces that you should consider that may temper our advice: customer journeys, business model needs and physical infrastructure.

Localize around customer journey archetypes

If your customer journeys differ significantly between operational areas, you may need to localize to some degree. But be sure these differences are real, rather than perceived. For example, the retailer in the introduction initially argued that every country has a different customer journey. But in reality, these differences were only skin deep: Customers buy running shoes in the U.S. in a very similar way to Germany. This is why they initially decentralized, only then to bring all the activities centrally.

Across the companies we studied, in about 80% of cases the underlying customer journeys were the same, 15% of customer journeys were similar (meaning they can be solved through a configurable central solution), and less than 5% proved country-specific (customer driven, but also regulation-driven, such as Portugal having 5 VAT zones). Notably, B2B tends to have more “real” differences than B2C.

How do you deal with differences between customer journeys? Can you group your customer journeys around archetypes? For example, at first glance, a global beverage manufacturer selling primarily B2B faced an immense heterogeneity in customer journeys across their many markets. In the U.S. customers might order using an online platform whereas in Mexico they might demand a face-to-face representative. The difference in customer journeys required localization, but they found these many different journeys aggregated around three archetypal journeys — face-to-face, wholesale, and hybrid selling — around which they could capture some of the benefits of centralization and some of localization.

Localize for high-frequency business models

Different business models have different clock speeds, different revenue models, and different demands to win customer business, which can create pressures to localize. Specifically, some business models have long cycles of customer interaction where you work to win new customers on a monthly or annual basis, like platform-based and subscription-based business models, whereas others are much more fast-paced, requiring effort to win customer business on a weekly business, such as food retailing.

Fast-paced business models often require more localization to allow rapid response, local customization, and even deviation from the rules. Likewise, B2B relationships such as selling pet food to vets, container ships to shipping companies, or professional services, often requires some local presence, even if just a sales force. Lastly, customer demands may drive localization, such as the need for a physical presence to generate trust. Surprisingly, sometimes less physical presence is needed than expected. A study by a global consumer bank showed that in retail banking, very few branches were needed in the market to generate the needed trust for customer to do business with them. As for the sports retailer, although they initially categorized themselves a high-frequency retailer requiring a physical presence, in fact customers only purchased a few times a year and like the banks, a physical presence proved less critical with time.

Localize for high dependency on physical assets

The tie to physical assets, such as supply chain or route-to-market, can necessitate some level of local customization. An extreme example is producing beverages. Producers of these products tend to have a decentralized organizational structure, as the product needs to be manufactured locally. The ratio between price point and weight of product is too low to allow for shipping across countries. If your digital business is dependent on these physical infrastructures, be it product, logistics, route-to-market, decentralization of some parts of your organization might be required. But many things not dependent on physical assets can be digitized by region or even centrally.

The red herring driver: language and regulation

Often local language, or local regulations, get cited as grounds for localizing digital activities. But in the cases we studied, more often than not, these were distractions from the real drivers. Often such differences — in terms of digital — can be managed from a centralized location or sourced with a partner.

Again, to be clear, we are speaking in terms of digital, not product development or regulatory and compliance. But when it comes to digital, often more can be done centrally than might be realized. For example, the global bank ING has one global overview of legal requirements and one IT backlog. This has facilitated quicker delivery and more integrated products across their geographies. Likewise, Booking.com operates across language and regulatory boundaries with fully centralized customer service and marketing teams. Sometimes, local regulations and language require a small local team, but be sure the benefits of localization outweigh the loss of economies of skill!

How to apply the framework

To apply the framework, it can help to do two things. First, think about these forces in terms of having a high, medium, or low impact. Second, think about choices not as central vs local, but as the “degree” of centralization, or the “types” of centralization (i.e., which activities are amenable to centralization? Are there regional archetypes if not country-level archetypes?).

As an example, before using this framework, the global beverage company mentioned earlier operated at a local level with each country having a nascent digital team. But using the framework they realized the potential benefits of economies of skill, both for attracting digital talent to an older incumbent and developing their capabilities. Next, they analyzed their customer journeys, which differed on each country and realized that they bore critical similarities. In some countries, sales reps sold directly to retail outlets and for these countries they developed smartphone apps for sales reps and point of sale tools for retailers that allowed them to gather and share data about customer trends. Instead of selling purely on relationships, sales reps could now go in and give retailers data-based advice about ways the retailer could increase their sales. In other countries, they sold directly to wholesalers and for these they build digital transfer tools to better manage the relationship. The remaining countries proved a hybrid of these tool kits.

Although the company localized the tools for these customer journeys archetypes, they developed them with the central digital team, co-creating them and testing them in partnership with local units. Likewise, wherever possible they centralized activities like data warehousing, but they did empower local teams to do their own data analysis, using best-in-class, off-the-shelf tools as well as provide them reports and insights centrally.

In conclusion, ever since companies started operating outside their domestic markets they have had to deal with a question of centralizing or localizing their business activities. In the past such choices were made based on physical infrastructure and local culture. But as digital lowers barriers to create, connect, and transact, how to make these choices is shifting. Economies of skill put a premium on centralizing digital activities, yet considerable differences in customer journeys, high frequency business models and dependence on physical infrastructure will pull the company towards local solutions. When you understand these forces and evaluate the pros and cons of centralization, you can make a more informed decision for where to locate your precious digital talent.

This article is written by Nathan Furr, Andrew Shipilov, and Jur Gaarlandt. It has first been published in Harvard Business Review.