So You Think You Want to Reorg
Recently, someone asked us a simple question with no simple answers: "How should you restructure a global organization?"
Here was our response (in four parts):
First, we don’t assume that structure is actually a useful lever for organizational effectiveness or responsiveness. If it were, re-orgs would be worth the effort. Yet when McKinsey studied 25 corporate re-orgs, they found that 80% failed to deliver the intended value and 60% noticeably reduced productivity. Every year, there is some "miracle cure" org chart being bandied about and every year companies destroy their productivity chasing what turns out to be nothing but a rehashing of the classic matrix model (we're looking at you, Spotify Squads).
When the world went digital, companies responded by drawing "IT" as its own organizational vertical and hiring Chief Digital Officers to shoulder the responsibility for change. Instead of accelerating change, this silo'ing internally of a pervasive external force only slowed or blocked organizational change. But drawing a new box, inventing a new title, and adding another dotted line can feel like a fast, simple fix. And even though we have mountains of data to discourage it, the world keeps training managers to problem solve with an org chart. We don't outright ignore structure (it's crucial to make sure your structures are aligned with your strategies), but we don't lead with it in our practice.
Talent (acquiring, developing, exiting) and methodology (how work is initiated, distributed, and completed) are far more significant levers for organizational effectiveness and responsiveness. If you see your company slowing or falling behind, these are the domains to obsess over. Find the teams already succeeding here in your organization and amplify their practices. Trial leading-edge ways of working from other firms. Measure, experiment, and repeat. When we embed with clients, this is the bulk of our work. Because companies usually haven't focused here, there are often immediate and profound wins to be had.
Second, we have found that the most effective way to knit a scaling organization together (and keep it together amidst change) is intense customer-centricity. We believe that you shouldn’t have a team, department, or business unit without a stated customer or customers. Most often, those customers will be internal audiences (e.g. HR serves team leads as a customer group). This requires a mindset shift, but it pays enormous dividends. We train teams to investigate what their customers need using user-centered research models (design thinking, jobs to be done, etc.) and we even measure teams to an internal net promoter score (e.g. HR being rated by team leads in how well their serve their needs).
Third, and here’s where it gets nerdy, the question of where work lives (by department, region, etc.) in an organization should be driven by three things: a) the quality of talent you have and b) the level of complexity of the task, and c) the amount of complexity created by geographic differences. Quick definitions: Simple work is defined as repetitive/routine tasks (e.g. customer support). Complicated work is less routine, but with enough time and expertise a limited set of solutions are revealed (e.g. ux design). Complex work is defined by novelty, uncertainty, and ultimately a solution set far too numerous to possibly vet one by one (e.g. market strategy). Simple work can be easily dispersed (geographically and horizontally) if you have information systems to distribute and refine best practices. Complicated work is often organized into structures like centers of excellence, but it’s critical to ensure these structures are easy to interface with across the organization (mostly they aren’t). People fulfilling complicated work must also have access to continuous learning and development. Complex work should be localized (e.g. developing a market strategy for China should be developed by teams in China) and completed by centralized teams (teams with remote workers often struggle with complex work). Complex workers need systems that provide them with heaps of data and insights they can use in pattern recognition. All of this is academic noise, however, if your talent isn't capable of the work.
Most large companies are rushing to automate simple and complicated work. Financially, this makes sense but it might not be strategically sound. For example, years ago Zappos proved that human customer service could be a strategic advantage in a crowding market.
Fourth, while the first three points encompass the core elements of global organizational design you can still get it wrong if you ignore these basics:
- Work hard to create a unified organizational culture with strongly shared values and a shared approach to work across all regions (GE does many things in many places, but there's always a core GE-ness to it)
- Invest in face-to-face meetings among leaders in different regions (humans still bond faster and develop deeper levels of trust in person)
- Keep teams small and lean (no more than 13 people on any given team)
- Centralize information and obsess over how teams interface (human-to-human interfacing is just as critical in a modern organization as APIs)
Whew. Long answer, huh? In fact, it could be much, much longer given the unique challenges within any single organization. This answer isn't meant to be a one-sized-fits-all solution, but it is meant to provoke further questions while exposing our own biases (every partner has biases). When we do engage with a client, we use a change model that we've developed that helps us refine and adapt our approach. If you want to keep going down this rabbit hole, our change model is actually public.