You Can't Integrate Your Way Out of a Bad Purchase: What 79 Acquisitions Show

The integration worked. The deal didn't.

In 2019, McDonald's bought Dynamic Yield, a personalization start-up, and did what the integration playbook asks. It moved fast. The technology was running in its U.S. drive-thrus "during a quick, six month ramp up". Then McDonald's sold the business to Mastercard.

The integration worked. The deal didn't. That pattern turns up again and again, so we went looking for it. We took 79 acquisitions whose outcome is known and read what was published about how each one was integrated: company announcements, filings, press coverage and case studies. Of the 79, 27 were later sold off, unwound or written down. Then we asked a simple question. Did the way a deal was integrated tell you which ones?

Barely. What the buyer bought told you far more.

Every deal, every outcome and every source is public at stratrix.com/data/acquisition-integration, with a CSV to download. Check our work.

Absorb it, leave it alone, or meet in the middle: a third went wrong either way

Absorb the business. Keep it separate. Combine the back office and leave the brand alone. Each went wrong about a third of the time.

Keeping the acquired company's leaders didn't save deals. Losing them didn't doom them. Cost synergies, the part of the deal case everyone calls safe, failed about as often as revenue synergies. And speed? Too few deals report it to say: two of nine fast integrations went wrong, and three of eight slow ones.

Some of the best-run integrations in the record belong to deals that didn't last. HP and Compaq planned so carefully that "on the day the merger was approved, the new company was ready to go". Years later, HP split the merged company in two. Lululemon did the opposite with Mirror and left it alone: "MIRROR will operate as a standalone company within lululemon". Three years later, Lululemon said it would stop selling the device.

Meticulous or hands-off, both ended the same way.

Why culture is the wrong thing to measure after the fact

Ask why mergers fail and someone will say culture. Maybe. But try to test it and something odd shows up. Almost every report that calls two cultures similar was written by the buyer, at the announcement. WPP on Young & Rubicam, in 2000: "The two organizations are highly compatible." Nestlé on Ralston Purina: "We have seized a unique opportunity to merge two organizations that share a similar outlook and similar cultures."

Almost every report of a clash was written after the deal went wrong.

“AOL felt like a family, congenial and working toward a common good, according to employees. Working at Oath with Yahoo felt mechanical and cutthroat.”— Business Insider, on Verizon's combination of AOL and Yahoo

Compare the two and you are measuring when a story was written, not what happened inside the companies. WPP later wrote down much of what it bought with Young & Rubicam. Verizon wrote down AOL and Yahoo and sold them. Culture may well have mattered in both. But nobody describes a clash before closing, and nobody praises harmony after a failure, so the record can't tell you.

What did separate the outcomes: the purchase

Our Decision Records on 88 acquisitions of $1 billion or more, the record behind the Distance Test, sort deals by what was bought. The gap is hard to miss.

Buy a business you don't run, and 7 of 14 deals were later sold off or written down. Buy the business next door, one that shares your customers, technology or channels, and it was 4 of 32. Buy growth, and 7 of 15 went wrong. Buy a capability you lack, and it was 2 of 27.

Half against one in eight. Nearly half against about one in fourteen. No integration plan closes a gap like that.

Before you hire integration help

Good integration advisers earn their fees on execution: sequencing, systems, keeping the right people, the first hundred days. Hire them for that. But before you sign, ask one more question: what would they have said about the purchase itself? Then hold the answer up against the record.

Questions