When vertical integration makes sense, and when it does not

Owning the stage above or below you in the chain is one of the most expensive choices a business can make, and one of the hardest to undo. It pays in a small number of situations. Most of the reasons given for it are not on that list.

Vertical integration means doing for yourself what a supplier or a customer used to do for you. The miller buys the farm, the brand opens its own shops, the manufacturer builds the component plant. The case for it is usually put in one of three words: control, security or margin. None of those is a reason on its own.

The four reasons that hold up

The test we use has been in the strategy kit since the early nineties and has not needed much updating. There are four reasons that justify owning the adjacent stage.

The market between the stages has failed. A market fails when there are only a handful of buyers and sellers, when the assets involved are specialised and long-lived, and when the two sides trade often. Each side then knows the other cannot walk away, and every negotiation becomes a hostage exchange. A contract long enough to cover every case cannot be written. Molten metal moving from one plant to the next is the classic case. A bespoke component with one supplier and one customer is the everyday one.

The adjacent stage earns better returns than yours. If the business above or below you in the chain sits in a structurally more attractive industry, moving into it can be a sound where-to-play choice. The question is then the same as for any market entry: can you win there, not only be there.

Integration raises a barrier or lets you price differently. Owning distribution can shut competitors out of a channel. Owning the input can let you serve segments at different prices without the arbitrage a middleman would run. These are real reasons, and they are also the ones regulators watch.

The market is very young, or dying. In a new market nobody yet supplies the stage you need, so you build it to get the market going. In a declining one the independents leave and you take on the stage to keep your own business alive. Both are temporary positions, and a sound plan says when you will hand the stage back.

The reasons that do not

Three arguments come up in almost every integration proposal, and each fails on inspection.

"It will smooth our volatility." Owning a supplier whose prices swing does not remove the swing. It moves the swing onto your balance sheet and adds a plant to run. If the input market is genuinely competitive, a hedge or a contract does the same job for a fraction of the capital.

"It secures supply." Only if the market has actually failed. Where there are many suppliers, supply is already secure, and the ownership premium buys nothing but a fixed cost.

"We capture more of the value added." Margin in an adjacent stage belongs to whoever has the advantage in that stage. Buying the stage buys the revenue, not the advantage. A business that was an average supplier to others becomes an average supplier to itself, and one that cannot be sacked.

The middle options

The choice is not make or buy. Between the two sit long-term contracts, joint ventures, minority stakes, franchising and preferred-supplier agreements. One of them is usually the right answer. They buy most of the coordination for a fraction of the capital, and they can be unwound.

A steel maker we advised asked whether to build its own cold-rolling capacity to feed its coated-products plants across several Asian countries. The question was answered country by country. In most of them the answer was no: the open market worked, and a supply agreement backed by a credible threat of building was enough. In one country the market had few sellers and had failed, and there the investment case was real.

The reverse also happens. Soft-drink majors have bought their bottlers and then sold them back to franchisees, more than once, as the balance between control and capital shifted. Car makers spun out the component divisions they had spent decades building. Neither is a failure of the framework. Integration is a position, not a permanent identity.

A test you can run in an afternoon

Take the stage you are thinking of owning and answer four questions.

1. How many credible suppliers or customers are there for this stage, and how many of them could serve you next year?
2. Are the assets involved specific to this relationship, and how long do they last?
3. Which of the four good reasons applies?
4. What is the cheapest arrangement short of ownership that gets the same result?

If the answers are "many", "not very", "none, but it feels safer" and "we have not looked", the proposal is about comfort, not strategy. If the answers are "two", "very", "the market has failed" and "we tried a contract and it broke", you have a case.

Does your strategy make real choices? · Where to play: how to diagnose markets · The stages of excellence in strategy