Capital is expensive again. With the era of near-zero rates over, every dollar now carries a real opportunity cost. At the same time, the demands on capital are climbing. To keep up in the AI race, firms need to make investments into AI tools, tech stacks, and upskilling; to manage geopolitical volatility, they need to retool their supply chains. Yet for many firms, effective capital allocation remains elusive.
So what separates the best capital allocators? We analyzed nearly 600 publicly listed multi-segment companies with at least $5 billion in market value over 2015 to 2024 to find out. (See the sidebar, “Identifying the Best Capital Allocators.”) When we compared each firm’s value with a synthetic sum-of-the-parts benchmark built from the pure-play peers in its segments, just 30% end up being worth more than the sum of their parts. (See Exhibit 1.) These top performers stand out not just on valuation but also achieve 1.6 percentage points higher returns on their investments and grow their assets faster than the rest.
One reason effective capital allocation is so rare is that it punishes the very instincts that serve leaders well almost everywhere else. The instinct to be even-handed leads them to spread capital equitably across their businesses, rather than backing the most promising ones. The instinct to double down on success makes their allocation backward-looking, feeding proven businesses while starving tomorrow’s winners. And the instinct to be decisive leads them to overreact to temporary setbacks.
The outperformers act the opposite: first, they are selective about where they invest; second, they reallocate capital to the businesses with the highest potential for future value creation; third, they stay committed through adversity.
Be Selective, But Not Mechanical
CEOs often try to be even-handed: capital is spread in proportion to the size of each business, or to last year’s budget. This approach looks reasonable on the surface: it is fair, administratively simple, and keeps political maneuvering to a minimum. But it is a trap. Spread evenly, capital may lead to underinvestment in some businesses and overinvestment in others.
The best allocators do the opposite: they concentrate, backing the segments where a dollar does the most and starving the ones where it does not. In our data, this shows up as an 18% wider spread in investment intensity across outperformers’ segments than across the bottom third group’s segments in our sample. (See Exhibit 2.)
Being selective is not the same as being rigid. Indeed, the opposite of pro-rata funding would be to install a pecking order, which mechanically dictates allocation. Top performers avoid this: we found no evidence in our data that their investments into the same segments are significantly correlated from year to year. Rather, top performers regularly review priorities in capital allocation—concentrating investments in any given year, and re-concentrating in the next.
How can this strategy be put into practice? At the portfolio level, the most useful tool is to give each business an explicit role, which captures what you expect of it—and therefore how to fund it: for example, a growth business is backed to expand, even while current returns are thin; by contrast, a business set to defend gets only what it needs to hold its position.
This, of course, is the logic behind the classic growth-share matrix, turned into a funding rule. The catch is that roles must be revisited, or they harden into the very pecking order we warned against. Shell demonstrates a long-running version of this discipline: for decades, it has used scenarios to steer where it commits capital across its businesses, revisiting that portfolio direction as the environment changes.
Finally, for portfolio roles to be effective, incentives must match them: a leader asked to build a growth business cannot be measured only on this year’s P&L, or they will quietly manage it like a cash engine.
Back Potential, Not Past Winners
The second instinct to override is to double down on success. When a business is thriving—earning high returns, growing fast, or generating cash—it feels obvious to give it more. This looks evidence-based and prudent. Yet, it means that capital is allocated based on gravity, rather than strategy: cash is pooled where it is generated, or where it has been deployed successfully in the past—rather than where it would work best.
Top performers break the link: within their portfolios, the correlation between how much cash a business generates and how much investment it receives is 27% lower than among the bottom third of performers. Moreover, we found no tendency among outperformers to pour capital into their highest-return or fastest-growing segment―the backward-looking move that instinct pushes toward. In other words, top performers manage to not confuse where value was created with where it will be created. (See Exhibit 3.)
Three habits keep capital pointed at future potential.
Anchor investment decisions in the strategic plan rather than last year’s budget. Weigh each investment on its strategic value versus the cost of not acting. Conversely, judging investments purely based on internal rates of return should be avoided, as it mechanically favors the proven businesses that will continue to deliver in the short term.
This reasoning is apparent today in how leading tech companies think about their AI spending. Alphabet’s Sundar Pichai has argued that “the risk of underinvesting is dramatically greater than the risk of overinvesting”; Meta’s Mark Zuckerberg has made much the same case. Their logic is strategic: weighing the cost of being left behind rather than a near-term return. Of course, the caveat is that “strategic” thinking needs to be followed by results.
Weigh investment requests against one another, rather than approving each on its own merits. Otherwise, the risk is that any reasonable case can clear the hurdle. Comparing business cases across segments turns a string of go/no-go calls into real selection.
Amazon runs a version of this in its annual operating plan process: each business submits its plan into a single review, where leadership decides which operating resources to fund and which to cut, rather than approving each on its own merits.
Tie funding to milestone-gated tranches—with clear evidence and kill criteria. Add to what proves out and stop what does not. Beyond limiting risk, this is what makes reallocation routine, since exit becomes the automatic consequence of a missed milestone rather than a political decision. Many successful pharma companies have established a version of this practice in their R&D pipelines. For example, Pfizer continuously prunes its R&D programs, including late-stage ones acquired at high cost, and reinvests the capital into its most promising bets. Rigorous post-mortems on major capital projects should then be conducted in order to create a feedback loop, making the next allocation decision better informed than the last.
Hold Your Nerve, Even When Times Are Tough
The third instinct to override is decisiveness. When a business you back stumbles, the natural move is to act: trim its funding or move to exit. It feels responsive to stay close to performance and refuse to throw good money after bad. But reacting to every swing means stop-start investing, and its hidden costs—capability decay, lost momentum, the expense of restarting later—often outweigh what the reaction saves.
The best allocators hold through the dip: their investment in a given business is about 30% less volatile from year to year than the bottom third’s, a sign that they respond to signals deliberately rather than reflexively. Top performers recognize that short-term performance is a poor guide to the future value of a business. (See Exhibit 4.)
But their discipline goes further: even when a segment faces a sustained, multi-year decline, top performers are less likely to pull the plug—cutting its investment in only about half of cases, against roughly seven in ten among the bottom third of capital allocators.
That restraint rests on a distinction most companies blur: treating a business that is in decline as if it were a burden and had no way to contribute to the portfolio. A declining business can still generate cash, optionality, talent, and customer relationships if it is given a deliberate role and managed for those outcomes rather than reflexively starved. For example, if an entire industry faces challenges and competitors are cutting back investments, holding capital levels in an exposed segment steady may enable winning market share in the next upturn.
In sum, the key is to not confuse a signal with a verdict, or reaction with discipline. This can be implemented in two steps:
First, decide in advance what would mark a genuine structural break rather than a temporary dip. The test is whether the economics of a business have shifted for good (for example, demand migrating to a substitute) or whether this is a cyclical trough or a one-off stumble that will pass. Setting those markers before a bad quarter arrives is what separates “this business is having a bad year” from “the world has changed under this business,” and keeps the call from being made in the heat of the moment.
A marker is a specific, observable threshold, not a feeling—and the most telling ones reach beyond financial results into the operational drivers of competitiveness: a substitute technology taking share for several years running rather than one soft quarter; a cost gap to the best competitor that stays open after the obvious fixes; a core product that has fallen two generations behind. Lagging financials confirm a structural break only after it has happened; operational signals like these flag it while there is still time to act.
Second, for a business in genuine structural decline, assign it a clear mandate, rather than letting it bleed out by default. Each mandate should carry its own investment expectation: a segment to be harvested is kept running efficiently for the cash it still produces, with new investment held to what sustains that output; a “turnaround” case is a time-boxed bet, with real capital and hard milestones, to repair the economics before a set deadline; an “exit” means selling the business down while it still has value to someone, or winding it down. The throughline is that decline becomes a deliberate choice with capital to match, not a slow starvation.
Resisting Your Instincts
The instincts to be even-handed, back what works, and act decisively serve leaders well almost everywhere else, but they tend to mislead in capital allocation, making investments mechanical, backward-looking, and reactive. Today, with capital expensive again and the demands on it climbing, leaders need to pivot to be selective, back potential, and hold their nerve in times of trouble to make the most of capital allocation.
Identifying the Best Capital Allocators
For each company, we constructed a synthetic “sum-of-the-parts” valuation benchmark. In every business segment, we took the median enterprise-value-to-sales multiple of pure-play peers in that segment′s industry (at the three-digit SIC level), then sales-weighted those segment multiples into a synthetic enterprise value for the whole firm. We compared actual value to this synthetic benchmark as a log ratio, averaged it across each company’s qualifying years, and converted it to a percentage. The companies that are worth more than the sum of their parts (roughly 30% of our sample) are the outperformers; the bottom third on this measure are the comparison group.
Practice 1. Investment intensity is a segment’s investment relative to its own asset base. We measured how unevenly a company spreads it with the median absolute deviation of intensity across segments: how far each segment sits from the company’s own median, in percentage points. We also checked whether a company’s ranking of fastest-growing segments persists year to year by assessing rank correlation of per-segment asset growth.
Practice 2. To assess the link between cash generation and capital allocation, within each company-year with at least three segments, we correlated each segment’s share of EBITDA with its share of investment, then took the median across years. To test the relationship between past success and current allocation, we measured whether companies steer extra growth into their own highest-return or fastest-growing segments (compared to the company’s other segments). Because these correlations are measured across each firm’s own segments, and because the bottom and top thirds are benchmarked against pure-play peers in each segment, the finding reflects allocation behavior rather than any industry composition effects.
Practice 3. Steadiness is measured as the coefficient of variation of each segment’s year-to-year asset growth, taken as the median across a company’s segments. To assess reactions to structural decline, we defined a “decline event” as a segment with three consecutive years of falling ROA, then measured how often investment intensity is lower three years in than at the start.
