The Diversification Decision: When Entrepreneurs Should Enter a New Industry — and When They Should Not

Diversification is the most romanticized strategic move in business and the most reliably destructive when done for the wrong reasons. The conference-stage version of the story sounds inevitable: a great operator sees an adjacent opportunity, applies their playbook, builds a second engine of growth. The version I have actually lived through is much messier — a long argument with myself, my team, and the balance sheet about whether a new industry is genuinely a chance to compound capital or a flattering excuse to dilute attention.

After thirty years of building across wine, real estate, hospitality, water, electricity, mobility, marine, and US distribution, the most useful frame I have found for the diversification decision is not "what industry do we want to enter" but "do we have a right to win in this industry, and at what cost to the businesses we already own?" The two questions are inseparable. Most failures of diversification are not failures of the new business; they are failures of the original one, which quietly starved while the founder chased the new shiny thing.

## Why Diversification Tempts Almost Every Successful Entrepreneur

The instinct to diversify is, in my experience, a function of three pressures that build up at the same career stage.

The first is **boredom with the original business**. By the time a company has been profitable for five or ten years, the operating challenges have become familiar. The founder has solved them once and now spends most of their time supervising people who are solving variations of the same problems. A new industry promises new puzzles. That promise is real. It is also, by itself, a terrible reason to deploy capital.

The second is **opportunity-set inflation**. Successful operators see deal flow that less successful operators do not see. Brokers, advisors, and former colleagues bring opportunities to the door. Each looks plausible on its own. Aggregated, they create the illusion that "we are seeing more opportunities" is the same as "we should act on more opportunities." It is not.

The third is **identity drift**. The founder begins to think of themselves not as "the wine person" or "the real estate person" but as "an entrepreneur" — a more flattering and more dangerous self-image. Identity as a generalist makes it easy to talk oneself into industries one does not understand.

Diversification done well is a deliberate, almost reluctant decision. Diversification done badly is the natural product of a successful operator following their own restlessness.

## The Three Questions That Decide the Outcome

Before our group enters any new industry, I now force three questions to be answered in writing, with real evidence, before any commitment is made. The discipline is uncomfortable. It is also the difference between the diversification moves I am proud of and the ones I would undo.

### 1. The Right-to-Win Question

What, specifically, do we bring to this industry that the existing players do not? The honest answers are narrower than founders want to admit. Capital alone is almost never enough — every industry has well-funded incumbents. Generic operating talent is also not a moat — most industries already have professional management. The right-to-win answer has to be specific: a distribution relationship, a brand the new market values, an operating discipline transferable from the original business, a structural cost advantage, or an unusual willingness to operate at lower returns for a longer time horizon than public-market competitors.

When we entered hospitality with Palacio de Manzanos in Haro, the right-to-win answer was not "we are good operators." It was "we already own the wine, the heritage, the regional relationships, and a property that cannot be replicated, so the marginal cost of producing a luxury hospitality experience around it is structurally lower than for any new entrant." That was a real answer. If the answer had been "the hospitality industry has good margins," it should have stopped the project.

### 2. The Adjacency Question

How close is the new industry, operationally, to what we already do? Not in terms of customer-facing similarity — those overlaps are usually superficial — but in terms of the management muscles required.

Real estate, hospitality, and wine, in our case, share a deeper operating logic than they look like from the outside: long capital cycles, brand-driven pricing power, regional concentration of expertise, and an unusual tolerance for slow returns. That overlap is what made the diversification credible. Mineral water and electricity were further from the core; the case for those rested on different reasoning, which I will come back to.

The mistake I have watched many entrepreneurs make is overestimating adjacency. A wine producer entering a beverage that is "also bottled" is not in an adjacent business. A retail brand entering "another retail category" is often in a different business entirely. The customer-facing surface fools the founder. The operating reality does not.

### 3. The Capacity Question

Do we have the management bandwidth to actually run a new business without degrading the existing ones? The dirty secret of diversification is that the founder's time is the binding constraint, and there is never as much of it as the deal memo assumes.

A new business demands disproportionate founder attention in the first three to five years — not the management time the model assumes. If the existing businesses are still in a phase where they require active leadership, the diversification will either be done badly or will pull the founder away from the original at exactly the moment they should be deepening it. I have made this mistake and watched friends make it. The capacity question is the one most often answered with optimism instead of evidence.

## When the Answer Is Yes

There is a narrow set of conditions under which I now think diversification is genuinely value-creating for an entrepreneurial group.

- **A defensible right to win.** Specific, articulable, and supported by an asset or capability the founder already controls.

- **Genuine operating adjacency.** The management muscles required overlap meaningfully with those already trained in the existing businesses.

- **Excess capital and excess management capacity.** The original businesses generate cash they do not need, and the leadership team has bench strength beyond what the original requires.

- **A long time horizon.** The new industry can be entered at a pace that matches the founder's actual life and energy, not a pace dictated by deal flow.

When those four conditions are present, diversification compounds the group. When even one is absent, it usually subtracts.

## When the Answer Is No

The cases where the answer is no are more common than the cases where the answer is yes, and saying no is the harder discipline.

I now treat the following as red flags that pull me out of the deal regardless of how attractive the underlying opportunity looks:

- **The thesis depends on the founder personally running the new business.** If we cannot identify a credible operator before signing, we are not buying a business; we are buying ourselves another job.

- **The original business is not yet "boring."** If the existing platform still requires active leadership to grow, diversifying is borrowing attention from the place it is most productive.

- **The "right to win" reduces to "we have capital."** Capital is the easiest moat to outbid. If it is the only one, the returns will be median at best.

- **The deal has urgency.** Genuinely attractive industries do not require sprint-speed decisions. Urgency is almost always a tactic, and the side that owns the calendar wins.

The single best filter I have found is the question: "Would I still want to do this if it took five years longer than I think?" If the answer is no, the project is being driven by impatience, not by strategy.

## What the Public Cases Teach

Three well-known examples are useful, less because they prove a rule than because they illustrate where diversification works and where it fails.

When LVMH built its portfolio under Bernard Arnault — wine and spirits, fashion and leather, watches and jewelry, perfumes and cosmetics, selective retail — the connecting tissue was not "luxury industry" in the abstract. It was a specific operating model: heritage brands acquired at moments of weakness, capitalized for very long horizons, run by category specialists, and sold through controlled distribution channels the parent company increasingly owned. That set of management muscles is what made diversification compound. LVMH would be a much weaker group if it had simply bought "good brands" without that model.

When General Electric, under Jack Welch and then Jeff Immelt, expanded into financial services through GE Capital, the strategic logic looked sound on a slide and fell apart in a real downturn. The reason was a misread of the adjacency question. Industrial cash flows and financial-services balance sheets behaved nothing alike under stress. The diversification did not extend the moat; it imported a wholly different risk profile that the parent company was structurally unprepared to manage. The eventual divestiture cost shareholders more than a decade of compounding.

When Berkshire Hathaway, under Warren Buffett and Charlie Munger, diversified from textiles into insurance, then into a wide range of consumer and industrial businesses, the logic was different from both of the above. The shared model was not industry adjacency at all — it was capital allocation discipline, decentralized operating autonomy, and a willingness to hold for decades. The diversification worked because the binding constraint was capital, not founder attention, and because the operating model was deliberately designed to scale across unrelated businesses without pulling the principals into any of them.

The pattern across the three cases is the same. Diversification works when the connecting tissue is real and operationally specific. It fails when the connecting tissue is abstract — "we are good operators," "this is a growing industry," "these markets are converging."

## What I Have Learned Across Eight Verticals

In our own group, the diversification moves I am most confident about are the ones where the right-to-win answer was specific and where the original business was healthy enough to run without me before the new business absorbed real attention. The moves I would re-examine today are the ones where I underestimated how much management capacity the new business would require, or where I told myself a story about adjacency that did not survive contact with operating reality.

Three things I would tell a younger version of myself standing in front of a new industry:

- The new business will need at least double the founder time you currently estimate. Plan for that, not for the model.

- The original business does not stand still while you are away. If it has not yet been built to operate without you, build that first. Then diversify.

- Boredom with the existing business is information, but it is not a strategy. The cure is usually inside the same industry — a new geography, a new product, a new layer of the value chain — not a different one.

## Key Takeaways

- Diversification is romanticized in business literature; in practice it succeeds only when the connecting tissue between businesses is specific, operational, and defensible

- The three questions that decide the outcome are: do we have a right to win, is the new industry genuinely adjacent operationally, and do we have the management capacity to run it without degrading what we already own

- The most common failure mode is not the new business failing — it is the original business stalling while the founder's attention is diverted, which is invisible until it is severe

- A real right-to-win answer is narrow and specific (a distribution relationship, an irreplaceable asset, a transferable operating model) — capital and generic operating talent are not moats

- Boredom with the existing business, opportunity-set inflation, and identity drift toward "generalist entrepreneur" are the three internal pressures that push founders into bad diversification decisions

- The strongest filter is the five-year delay test — if the project would not still be worth doing if it took five years longer, it is being driven by impatience rather than strategy

- The pattern across LVMH, GE Capital, and Berkshire Hathaway is consistent: diversification compounds when the operating model is real and specific, and dilutes when it rests on abstract industry-level reasoning

The most important thing I now believe about diversification is that the question is rarely about the new industry at all. It is about whether the original business is strong enough, run well enough, and free enough of the founder to allow a second business to be built without taking the first one down. When that condition is met, a group can compound across decades. When it is not, no industry, no matter how attractive on paper, will save the entrepreneur from the math.