Diversification Is Not a Strategy: 5 Tests Before You Add Another Business to the Group

In April 2024, General Electric finished cutting itself into three pieces. The company that spent most of the twentieth century arguing that a diversified group could be run better than the sum of its businesses concluded that its aviation, healthcare and energy divisions were each worth more standing alone. GE Aerospace, GE HealthCare and GE Vernova now trade separately. Siemens had already spun out Healthineers and Siemens Energy. Tyco, ITT, DowDuPont, all of them dismantled by the same logic.

Meanwhile Berkshire Hathaway owns insurance, a railroad, utilities, furniture stores and a candy company, and the market has never asked it to break up.

The difference between those two outcomes is not how many businesses you own. It is whether you can prove you are the best owner of each one.

Our group was founded in Azagra in 1890 and today runs eight divisions in more than 75 countries, from wine and real estate to hospitality, mineral water, electrical installations and luxury mobility. People assume a portfolio like that came from a diversification strategy. It did not. It came from five generations of individual decisions, and the only reason it holds together is that each one had to survive the same set of questions. Here are the five I use before adding anything else.

First, understand what the market is telling you

A conglomerate discount, as Investopedia defines it, is what happens when investors value a diversified group below the combined worth of its individual businesses. Academic work on financial conglomerates published in ScienceDirect found that the diversification itself causes the discount, rather than troubled firms simply diversifying into more promising areas. The default assumption in capital markets is that adding an unrelated business subtracts value.

McKinsey frames the same evidence more usefully. A corporation with multiple businesses should not be worth less than the sum of its parts, and when it is, the cause is usually a performance problem or a communication problem rather than the structure itself. That is the honest version. Diversification does not destroy value by arithmetic. It destroys value when the owner cannot explain what the second business gets from being owned by the first.

If you cannot answer that question in one sentence, you do not have a portfolio. You have a collection.

Test 1: The better-owner test

The only defensible reason to buy a business in another industry is that you can make it worth more than any other realistic owner can. Not that it is a good business. Not that it is cheap. That you specifically add something.

Ask what you bring that a competing buyer does not: a distribution channel the target cannot access alone, purchasing scale, a brand it can borrow, land, a licence, a relationship with a regulator, patient capital that lets it invest through a downturn. Write the answer down before you look at the price.

If your honest answer is "we have cash and they need cash," you are not a better owner. You are a bank charging equity prices for a debt product.

Test 2: The shared-capability test

Synergy is the most abused word in corporate finance because it is almost never specified. Force it to be concrete. A real link between two businesses is one you could describe to a warehouse manager: the same customers, the same sales force, the same logistics network, the same supplier, the same brand, the same technical skill.

Our wine business in Spain and our distribution company in the United States share a customer list, a shipping lane and a brand. That is a real link. Our hospitality business in Haro and our wine business share a place, a story and the same visitor. Also real. When we looked at businesses that shared nothing but a spreadsheet, we passed, and the two or three we did not pass on early enough taught us the expensive version of this lesson.

A synergy you cannot describe as a specific, repeatable transaction between two divisions is not a synergy. It is a hope with a number attached.

Test 3: The attention test

Capital is not the scarce resource in a diversified group. Senior attention is.

Every new business consumes a fixed amount of the leadership team's calendar regardless of its size, and it consumes disproportionately more when it is unfamiliar. A division that is 5% of revenue can easily take 30% of the executive committee's agenda in its first two years, and every hour it takes comes out of the businesses that already pay for everything.

Before adding a vertical, name the person who will run it without you. If that name is your own, the honest answer is that you are not buying a business, you are buying yourself a second job, and the core will pay for it. This is the failure mode nobody models, because it never appears in the deal file.

Test 4: The capital test

Every euro has a next-best use. The relevant comparison for a new business is never "will this return more than our cost of capital." It is "will this return more than putting the same money into the businesses we already understand."

That comparison is brutal, and it should be. Expanding a proven line usually carries lower execution risk than entering an unknown industry, so a new vertical needs to clear the internal alternative by a real margin, not match it. We treat that gap as a hurdle, not a rounding error, and it kills most ideas at the first meeting, which is exactly what a good test is supposed to do.

Test 5: The cycle test

The classic argument for diversification is that it smooths earnings. Research on German diversified firms published by Springer notes a genuine version of this effect: diversification can lower a group's default probability and raise the market value of its debt, an insurance effect across divisions.

The catch is that this only works if the businesses actually fail at different times. Wine and hospitality both suffer when travel and discretionary spending collapse. Real estate and construction turn together. If your divisions are correlated, you have not diversified risk, you have concentrated it while telling yourself a comforting story.

Test it with history, not intuition. Take your worst two years in the last twenty and ask what each division did in those years. If they all went down together, the portfolio is a single bet wearing several coats.

When diversification genuinely works

The groups that hold together share one trait: an owner who adds something specific and repeatable.

Danaher spent decades acquiring industrial and life-science businesses and improving their margins with a single operating system applied to every one. The value was not the diversity, it was the method. LVMH runs dozens of luxury houses that share a distribution machine, a retail footprint and a category expertise, which is related diversification, not random assembly. Berkshire's edge is a permanent, low-cost source of capital from insurance float plus an owner who does not interfere. In each case the parent is demonstrably the better owner.

The groups that fall apart are the ones where the parent was, in the end, only an accountant.

Key takeaways

• Markets discount diversified groups by default, so the burden of proof is on the owner to show the parts are worth more together than apart.
• The only durable reason to own a business in another industry is that you are the better owner of it, and you should be able to say why in one sentence.
• A synergy that cannot be described as a specific transaction between two divisions is not a synergy.
• Senior attention, not capital, is the binding constraint in a diversified group, and a new vertical always consumes more of it than the plan assumes.
• A new business must beat reinvestment in your existing businesses by a clear margin, not merely beat your cost of capital.
• Diversification only reduces risk if your divisions fail in different years, which you should verify against your own history rather than assume.
• If the parent adds nothing beyond capital and reporting, the discount the market applies is not a mistake. It is an accurate assessment.

Frequently asked questions

What is the conglomerate discount?

It is the gap that appears when the stock market values a diversified group at less than the combined value of its individual businesses. Investors apply it when they suspect the parent company adds cost and complexity without adding operational value, or when the group is too opaque to analyse division by division. McKinsey's view is that the discount usually reflects underperformance or poor communication rather than diversification itself, which means it is often fixable without a breakup.

How do you calculate the conglomerate discount?

You value each division separately using multiples from listed pure-play competitors in that industry, add them together, subtract net debt and central costs, and compare that sum-of-the-parts figure with the group's actual market capitalisation. If the market value is lower, the difference is the discount, usually expressed as a percentage of the sum of the parts. Private companies can run the same exercise using transaction multiples to see whether their structure is helping or hurting.

What are the four main types of business diversification?

Horizontal diversification adds new products for your existing customers. Vertical diversification moves you up or down your own supply chain, into your suppliers' business or your distributors'. Concentric or related diversification enters a new market that shares technology, skills or channels with what you already do. Conglomerate or unrelated diversification enters an industry with no operational link at all, and it is the type that carries the highest failure rate and attracts the largest market discount.

What are the reasons for diversifying a business?

The legitimate ones are reducing dependence on a single market or customer, using spare capacity in an asset or a team, entering a structurally faster-growing category, and buying a business you can genuinely improve. The illegitimate ones, which are far more common, are boredom with the core business, a wish to look bigger, and the belief that a strong balance sheet in one industry automatically transfers competence to another. It does not.

What did Warren Buffett say about diversification?

Buffett has repeatedly described broad diversification as, in his words, "protection against ignorance", arguing it makes little sense for an investor who genuinely understands the businesses they own. The distinction matters for operators: he is not attacking concentration in things you know, he is attacking the habit of spreading capital thinly to avoid having to form a judgement. For a company owner, the equivalent discipline is refusing to enter an industry you cannot explain.

Before you approve the next one

Take the last business your group added and run it through the five tests honestly. If it fails two of them, you already know where next year's management problem is coming from, and you have time to fix it.

See how the eight divisions of Manzanos Enterprises fit together, and for the two pieces that sit either side of this decision, read capital allocation, the CEO's most important job and build or buy, how to choose between organic and inorganic growth.