Transformation looks clean in retrospect.
There is the old company. Leadership sees the market changing. A new strategy is created. Capabilities are built. The business moves from one model to another.
That is almost never how it feels while you are inside it.
For a long period, both versions of the company exist at once.
The legacy model still has customers, revenue, employees, processes, budgets and expectations. The new model represents where leadership believes the company has to go—but it is smaller, less proven and usually more expensive than it looks on a strategy slide.
That is the dangerous middle.
The legacy business is not the enemy.
One of the easiest mistakes in transformation is to talk about the old model as though it is simply obsolete.
Usually it is not.
It may still produce most of the cash. It may hold the customer relationships. It may contain the people who understand the industry best. It may finance the very transformation intended to change it.
The problem is not that the legacy business has no value.
The problem is that its current value can give it disproportionate power over the future.
Research by Charles O’Reilly and Michael Tushman has long described this tension as the need to exploit the existing business while simultaneously exploring new ones. Their work on ambidextrous organizations argues that the two often require different structures, processes and cultures—but must still be integrated at the senior-leadership level.
I think the executive difficulty is even more practical than the theory sometimes makes it sound.
You are deciding which business gets the next dollar.
Which team gets the strongest people.
Which customer promise gets protected.
Which process gets rebuilt.
Which legacy revenue you are willing to disrupt before the replacement is fully mature.
Established revenue, mature processes, known economics and strong claims on near-term resources.
Early economics, new capabilities, incomplete proof and a need for protected investment.
The problem is not choosing “old” or “new.” It is deciding when the evidence is strong enough to move resources while both models still matter.
The new model is structurally disadvantaged.
The existing business knows how to defend itself.
It has historical revenue. Established forecasts. Managers with operating authority. Customers asking for continuity. Metrics built around its economics.
The new model has none of those advantages at first.
It can look inefficient precisely because it is being compared with a system that has had years to optimize itself.
MIT Sloan has described the strategic challenge for incumbents as doing the old and new business simultaneously, especially when the emerging technology can threaten the economics of the established model. HBR research on ambidextrous leadership makes the same tension explicit: established operations naturally make stronger claims on near-term resources because their returns are visible now.
This creates a predictable failure mode.
Leadership announces transformation but allocates resources using the economics of the legacy business.
The new model is then asked to prove itself while being starved of the capabilities required to become competitive.
Many transformations do not fail because leadership chose the wrong destination. They fail because the future business never receives enough protected resources to become a credible alternative to the present one.
Transformation becomes a capital-allocation problem.
This is where I think a lot of transformation language gets too soft.
Culture matters. Communication matters. Vision matters.
But at some point, transformation becomes resource allocation.
Headcount moves. Budget moves. Technology spend moves. Management attention moves. Incentives change. Some work stops.
If none of those things move, the transformation is probably still a presentation.
I lived a version of this moving a traditional media organization toward a digital-first model.
The work was not “launch digital.” It was deciding which capabilities the company needed to become good at: digital acquisition, CRM, analytics, technology, product, data and new revenue models. Those investments had to happen while the legacy business was still operating and still mattered.
Over time, digital moved from roughly 20% of the revenue mix toward 70%. Restructuring and operating-model redesign also produced approximately $2.8 million in annualized operating-cost reductions.
The result mattered. But the part I remember most clearly is the middle—the period when the old model still had stronger evidence and the new model needed disproportionate attention anyway.
Cost reduction and transformation are not opposites.
Another mistake is assuming transformation always requires layering new cost on top of the old organization.
Sometimes it does temporarily.
But if the operating model is genuinely changing, leadership should also be willing to ask what no longer needs to exist.
Which systems are duplicative? Which workflows belong to the old model? Which roles exist primarily to coordinate work that technology now handles differently? Which products consume complexity without creating enough value?
Transformation becomes economically credible when the new capabilities are paired with choices about what the business stops carrying.
The hardest decision is often what to protect.
New initiatives need protection from legacy economics, but not immunity from accountability.
That line matters.
A future-state initiative should not survive forever because it is strategically fashionable. Leadership still has to define what evidence would justify more investment, what evidence would change the thesis and what conditions would cause the company to stop.
Protection should create room to learn—not permission to avoid economics.
I would manage the middle with two scorecards.
The existing business and the emerging business should not always be evaluated the same way.
The legacy scorecard should emphasize current economics: revenue, margin, retention, productivity, cash generation and customer commitments.
The new-model scorecard should include different questions: adoption, learning velocity, unit-economics trajectory, capability maturity, customer pull, repeatability and how quickly uncertainty is being retired.
Eventually the two scorecards converge.
Early on, forcing them to be identical can kill the future before it has had a chance to become efficient.
The senior team has to host the contradiction.
This is where transformation stops being a program-management problem and becomes an executive-leadership problem.
The old and new models will make legitimate, conflicting claims on resources.
There is no spreadsheet that removes the need for judgment.
Senior leadership has to decide which contradictions are temporary, which legacy economics must be protected, which new capabilities deserve insulation and when the balance needs to shift.
Research on organizational ambidexterity repeatedly puts this responsibility at the senior-team level. The separate operating models can coexist only when leadership actively orchestrates the tradeoffs rather than expecting the organization to resolve them on its own.
What I may be wrong about.
Not every business needs a dramatic transformation.
The word itself is overused. Sometimes disciplined continuous improvement is the better strategy. Sometimes the legacy model has more life than leadership assumes. Sometimes the “future” model is simply fashionable technology with weak economics.
The danger cuts both ways.
A company can protect the old model too long.
It can also abandon a durable business too early because transformation sounds more visionary than optimization.
The job is not to prefer new over old.
It is to allocate against the evidence while understanding that the evidence arrives at different speeds.
Questions I would ask in the middle.
Evidence & sources
The transformation framework and operating conclusions are mine. These sources provide useful external grounding for the tension between exploiting the present and building the future.