THREE EXITS · BUSINESS BREAKDOWN

Marmite Was Not Sold Because It Failed

Unilever's board sold a business generating 22.6% margins because Nelson Peltz's framework made the asset non-core and Hein Schumacher's pay depended on growth it couldn't deliver.

$44.8bn: Enterprise Value

13.8x EBITDA / 3.6x Sales

$15.7bn: Cash Proceeds

Post-adjustments

65%: Equity Retained

~$29bn at VWAP $57.84

$600m: Synergy Target

Run-rate, Year 3

SITUATION

A profitable division gets reclassified as a liability

On 31 March 2026, Unilever announced the merger of its Foods division with McCormick & Company. Enterprise value: $44.8bn. Cash proceeds: $15.7bn. Equity retained: 65%. Synergy target: $600m by Year 3.

Marmite is the headline. It is a £200m bundle inside a $20bn transaction. The real story is not about yeast extract.

Unilever called it “another decisive step to reshape Unilever into a simpler, sharper, higher growth company” and said it would position the group as a “leading pureplay HPC (health personal care) business.” That is the official framing. The unofficial framing is simpler: Foods was profitable, it just wasn’t the right kind of profitable.

The Foods division was not underperforming. In FY2025 it delivered a 22.6% operating margin — above Unilever's group average of 20%. Underlying sales growth was 2.5%. The business generated margin. It generated cash.

It did not generate the number Unilever needed: 4–6% underlying sales growth. That is a different thing. The gap between those two numbers is where this case study lives.

TIMELINE

Q4 2022

Nelson Peltz / Trian Fund acquires significant stake. Power Brands framework adopted unanimously: 30 brands, 4–6% sales growth required, premium pricing and emerging-market scalability. Anything outside the framework managed for cash or exited.

May 2023

Hein Schumacher appointed CEO. Compensation indexed to sales growth and HPC margin expansion. The food division is already a drag on the scorecard.

Oct 2024

£40m invested in Marmite’s Burton-on-Trent facilities. Capacity doubled, 160 jobs created. This was not a commitment to Marmite's future, it was sale preparation. Condition determines exit price.

Nov 2025

Foods review signalled publicly. UK legacy bundle of Marmite, Colman's, Bovril (~£200m revenue) all confirmed non-core. The sale is now in motion.

Q1 2026

Foods sales growth: 1.6%, volume -1.1%. The numbers are deteriorating. The timing of McCormick purchase enquiry is good, before the sales numbers get worse.

31 Mar 2026

Deal announced. McCormick came inbound, Unilever did not advertise the asset. Close targeted mid-2027, subject to FTC and EU approval. $420m break fee if regulators block it.

MECHANISM

How an investors framework becomes a disposal order

Four steps. One outcome.

Step one:

Trian buys in, Q4 2022. The Power Brands framework is adopted unanimously. The board didn’t debate it. Schumacher later said Peltz’s views were “very much in line” with Unilever’s existing strategy.

Unanimity wasn’t groupthink. It was a rational response to knowing what Peltz was there to do. The framework set a hard filter: 4–6% underlying sales growth, premium pricing, emerging-market scalability, concentration in the top 30 brands representing over 70% of sales.

Marmite is stubbornly British, ~90% UK and European, with no credible path to faster growth. It doesn’t clear the filter on any dimension.

Step two:

Schumacher is hired in May 2023. His compensation was structured to reward sales growth and margin expansion, the same metrics the framework demands.

The contrast is measurable: post-separation, the remaining HPC portfolio had delivered 5.4% underlying sales growth over three years, a 48% gross margin and a 19% underlying operating margin.

Foods delivered 22.6% operating margin but 1.6% underlining sales growth. Every quarter it underperformed on growth, Schumacher’s scorecard got worse. This is not a character judgement. It is how pay structures work.

Step three:

the market reprices Unilever for holding the wrong mix. Comparable health and beauty businesses trade at 16–18x annual profits. Food businesses trade at 10–12x.

Foods was ~26% of Unilever's revenue. A lower-valued division making up that much of the group suppresses the blended multiple by 1.5–2.0 turns.

At Unilever's scale, each lost turn costs approximately $11bn in market capitalisation. Every quarter Foods stayed in the group, that cost continued.

Step four:

the exit becomes structural, not strategic. The framework doesn't leave room for Marmite to fix itself, the category doesn't support 4–6% growth regardless of execution. The CEO's pay actively penalises keeping it. The market discount compounds quarterly.

By 2024, the only remaining question was price and timing.

That's the mechanism. The disposal wasn't a decision made in 2026. It was a decision made in 2022. March 2026 was the delivery date.

The £40m tells you when the decision was made.

In October 2024, Unilever doubled capacity at the Burton-on-Trent factory and created 160 jobs.

Six months later, it confirmed the asset non-core.

This was reported as a contradiction. It isn't. You don't invest in an asset's future when you intend to sell it. You invest in its present condition because condition determines exit price. The factory upgrade was sale preparation. The workers in Burton-on-Trent would likely see it differently.

The decision was made in 2022, confirmed in 2024.

The valuation gap is measurable.

Foods at 10–12x, HPC at 16–18x: Unilever was carrying a valuation gap that grew every quarter.

Selling at 13.8x , full multiple, no distress discount and before the numbers deteriorated further was the only way to avoid advertising a sale and accepting a 15–25% discount from buyers who know you need to sell. McCormick came inbound. Unilever sold at full price.

P&G made the same call in 2014: cut ~100 brands, keept 65.

Duracell is the direct equivalent — $2bn in revenue, strong margins, sold to Berkshire Hathaway for ~$4.7bn at roughly 11–12x EBITDA.

Not because it was failing, but because P&G's cost of capital and growth targets made a stable, slow-growing battery brand the wrong thing to own.

Berkshire's cost of capital made it worth owning. The brand didn't change. The owner changed. P&G's share price rose more than 30% in the three years that followed, but it took 18 months of earnings dilution before the re-rating arrived.

Unilever is at that same 18-month window now.

DECISION LOG

Three gates, three names, one predetermined outcome

Gate 1: Q4 2022

Board / Peltz (Trian)

Adopt Power Brands framework. Set 4–6% sales growth as the qualifying threshold. Foods managed for cash.

The disposal is now structural. Marmite has no path to meet the threshold. What remains is when and for how much.

Gate 2: Oct 2024

Schumacher (CEO)

£40m capex at Marmites Burton-on-Trent facilities. Capacity doubled, 160 jobs added. Asset prepared for sale.

Exit price supported. The investment is preparation, not commitment. Six months later the asset is publicly non-core.

Gate 3: Mar 2026

Foley (McCormick CEO)

McCormick approaches inbound before Foods numbers deteriorate further. Unilever accepts at 13.8x EBITDA.

Full multiple achieved. No distress discount. $420m break fee means neither side expects this to fall apart.

Why unsolicited matters. When you advertise an asset for sale, buyers know you need to sell and price accordingly, typically 15–25% below an unsolicited offer.

McCormick came in before the Q1 2026 numbers had visibly deteriorated. Unilever sold at full price.

THE PARALLEL

The same asset, a different owner, a different outcome

McCormick is buying what Unilever is selling for the same reason Berkshire bought Duracell from P&G: a lower cost of capital makes slow, reliable cash flows worth owning. Berkshire's hurdle rate is approximately 8–10%.

Unilever's implied equity cost of capital, under investor pressure to deliver 4–6% sales growth, is closer to 12–14%.

At Berkshire's rate, Duracell's margins justified the multiple. At P&G's rate, they didn't. Nothing about the battery changed. The owner changed.

The Marmite version is the same mechanism. McCormick runs a global flavour and condiment business where a 2.5%-growth, high-margin UK brand is an asset, not a drag.

Inside Unilever's growth framework, Marmite suppressed the multiple. Inside McCormick's portfolio, it's accretive. The brand is unchanged. The capital structure around it is not.

This pattern is repeatable and applies to any founder carrying a profitable, slow-growing division inside a faster-growing business. The question is never whether the asset is good. The question is whether it belongs in your portfolio at your cost of capital.

FRAMEWORK

When to exit a profitable asset: three questions with pass/fail conditions

Question: If this division disappeared, would your growth numbers still hold?

Yes — it's a drag on the headline number, not a contributor.

No — you're not removing a drag, you're removing a buffer. Know which before you sell.

Question: Who set the growth target — and can this division ever hit it?

The target is achievable by the division within the category's natural ceiling.

The target was set by people whose returns depend on it and the category structurally can't deliver it. The exit is a timing exercise, not a decision.

Question: Why are you still investing in this business?

Strategic investment with a 3-year horizon and a credible path to the target.

Inside 18 months of exit, capital has one job: raise the exit price. If you're still investing, ask whether you're preparing or pretending.

One structural rule:

if the buyer can do something with the asset that you can’t, lower cost of capital, better distribution, category fit, don’t sell it outright. Keep a stake.

The ownership structure here is precise: Unilever shareholders receive 55.1% of the combined entity, McCormick shareholders receive 35.0%, and Unilever itself retains 9.9% on a fully diluted basis.

They are converting direct operating exposure into a minority participation in future upside. They get the synergy benefit without having to operate a brand with a structural growth ceiling. It only works if you sell before the numbers deteriorate. Once they deteriorate, the buyer knows it too.

VERDICT

Mis-owned, not mismanaged

Marmite was not mismanaged. It was mis-owned.

Unilever's cost of capital, growth targets and investor-driven framework made a profitable, well-run business structurally non-core. The board worked that out in 2022 and closed the deal in 2026. The four years in between were preparation, not indecision.

Unilever sold a known outcome for a better but uncertain one. $15.7bn in cash. 65% of McCormick's combined entity. A $600m synergy target they don't have to deliver, they just have to own 65% of the entity that does. The trade is rational. It is not guaranteed.

The trade breaks under three conditions:

HPC growth slips from 5% toward 3%;

McCormick fails to integrate a business twice the size of anything it has bought before;

and nothing replaces the earnings stability the Foods division provided.

Whether this was the right call will be clear in approximately two years. Ask again in 2028. The people who said it was rational have a financial interest in being right.

BY THE NUMBERS:

Foods Sales Growth at exit

1.6% (Q1 2026, volume -1.1%)

Foods operating margin

22.6% — above group average of 20%

HPC growth target

4–6% (Power Brands framework)

HPC actual (2025)

Beauty & Wellbeing 4.3%; Personal Care 4.7%

Exit multiple achieved

13.8x EBITDA / 3.6x Sales — no distress discount

McCormick synergy target

$600m run-rate by Year 3

Break fee

$420m if FTC / EU regulators block the deal

Equity retained

65% of combined entity (~$29bn at VWAP $57.84)

P&G / Duracell parallel

~11–12x EBITDA; 30%+ re-rating over 3yrs post-disposal

Post-separation HPC profile

5.4% USG (3yr); 48% gross margin; 19% underlying operating margin

Ownership split (fully diluted)

Unilever shareholders 55.1% / McCormick shareholders 35.0% / Unilever retained 9.9%

THREE EXITS · threeexits.com