Why Most Tech Acquisitions Destroy the Thing They Paid For
Over the years I had the opportunity to be part and/or watch several acquisitions; from both sides of the table. Getting acquired is considered a huge career milestone for founders and executives alike.
Even in the best scenarios and with the best intentions, most acquisitions destroy the value they paid for. Outside of my personal experience, there hundreds of examples of acquisitions that went sideways. For example:
Yahoo acquired Tumblr for $1.1 billion in 2013. CEO Marissa Mayer promised to "not screw it up." Within two years, Yahoo had written down $230 million in one quarter and another $482 million the next---over half the purchase price erased in six months. By 2019, Tumblr sold to Automattic for $3 million. A 99.7% destruction of value.
Salesforce paid $27.7 billion for Slack in 2021. Stewart Butterfield, the founding CEO, made it eighteen months. Cal Henderson, the CTO and last remaining cofounder, made it thirty. By early 2024 every cofounder was gone and Slack was being run out of Salesforce. A leaked all-hands surfaced employees calling it a "strong culture clash." Butterfield was less diplomatic on his way out: "There's no incorporation of the Slack culture into the Salesforce culture. And unless there is some element of that, then it's not integration in any sense. It's just the elimination."
Broadcom closed its $69 billion acquisition of VMware in November 2023 and laid off 401 employees within the first week. Over 2,800 followed. VMware's president left; senior staff who had watched Broadcom's prior playbook at CA Technologies retired preemptively rather than wait. The goal was straightforward---more than double VMware's EBITDA to $8.5 billion within three years, primarily through cost-cutting. Customers got a forced transition from perpetual licenses to annual subscriptions; employees got return-to-office mandates within sixty miles of a Broadcom office. Financial extraction, not integration.
Google bought Nest for $3.2 billion in 2014. Founder Tony Fadell left after two and a half years of internal fighting and product stagnation. Alphabet reportedly tried to sell Nest in 2016---two years after paying $3.2 billion for it.
HP acquired Autonomy for $11.1 billion in 2011 and took an $8.8 billion writedown. HP blamed accounting fraud at Autonomy and years of litigation followed, but a UK court found that 80% of HPE's losses had nothing to do with the alleged misconduct. HP's own mismanagement of the integration was the primary cause. They destroyed the value, then blamed the seller for it.
All different industries, acquiring companies, and yet the same unfortunate outcome. Roger Martin, writing in HBR, called M&A "a mug's game, in which some 70% to 90% of acquisitions are abysmal failures." McKinsey found that 61% of acquisition programs didn't earn back their cost of capital; Bain pegged it at 70% failing to increase shareholder value.
Most acquisitions destroy more than they create.
What You're Buying
In order for us to understand why most acquisitions fail, we need to understand what is being bought in the first place. Strip away the self-congratulatory press release language and every acquisition is buying some combination of three things: the product, the team, or the book of business. Most acquirers tell themselves they're buying all three. Most integration playbooks only know how to keep one of them — and sometimes they can't even do that.
Product is the easiest to keep, and the least valuable on its own
The codebase, the IP, the stack: those are snapshots. You can audit them, document them, port them to your own infrastructure. They behave the way assets are supposed to behave on an acquirer's balance sheet, but fail to account for the fact that they are often built by a team that is no longer there.
A product without the team that built it ages exactly as fast as the engineers you have left to maintain it — and after an acquisition, the engineering bench is usually thinner than the deal model assumed it would be. This becomes even more pronounced when the product required industry or specialized knowledge that is difficult to build or replace; for example products that operate in highly regulated industries or require specialized hardware or software.
The product is the part of the deal you most reliably take possession of and the part that decays fastest without the team and the culture that built it. This also the number one mistake acquirers make.
The team is the hardest to keep, and where most of the value actually lives
Let's look at the numbers around retention, once the aquisition closes MIT Sloan ran U.S. Census Bureau data on 230,000 acquired startup workers over twenty years — 33% of them walked within the first year, compared to 12% for regular hires. Almost triple the churn, on a sample size big enough that it isn't really arguable. EY pegs it worse: 47% gone in year one, three-quarters gone by year three. Founders specifically: 36% leave within six months, 75% are gone by eighteen months, fewer than a quarter remain at three years.
It has very little to do with comp or titles. MIT Sloan researchers noted that "a larger, more established firm has varying levels of bureaucracy and a formal corporate culture" while startups attract workers who prefer risk-taking and autonomous environments. More than half of founders who leave do so because they no longer fit in — not performance pressure, not money. They just don't recognize the place anymore. The qualities that made the team valuable in the first place — autonomy, speed, the willingness to break process to ship — are the same qualities that make them walk out when those conditions disappear.
You paid a premium for a team built on independence and then your integration playbook removed the independence, the insights, and the culture that built the company worth acquiring in the first place.
Retention doesn't mean engagement. Earn-outs and retention bonuses turn into golden handcuffs — people physically present but mentally counting down to a vesting cliff. That's attendance with a deadline, not retention.
The book of business is the most fragile line item in the deal
Customers weren't loyal to a logo. They were loyal to specific people, specific terms, specific responsiveness. Change the terms and a meaningful chunk of the book walks. Broadcom learned this in real time when it forced VMware customers off perpetual licenses and onto annual subscriptions — the customer-revolt headlines started inside the first year. Channel partners and distributors run the same dynamic in the background; they negotiated their margins and SLAs with one company, and now they're being asked to renegotiate with another, usually on worse terms.
Most deal models assume something close to 100% revenue retention in year one. The actual range, in my experience and the data, is closer to 70-85% — and that's before the acquirer starts changing pricing or sunsetting the features that didn't fit the parent company's "vision". The book of business is the easiest line to model and the hardest to actually retain at the modeled value.
What acquirers take for granted
Across every deal in the intro — Tumblr, Slack, VMware, Nest, Autonomy — the acquirers made some version of the same set of unforced errors. The specifics vary; the assumptions don't.
They assumed tacit knowledge would transfer. It doesn't. The codebase ships; the institutional memory walks out with the engineers. Most of it was never written down because it didn't need to be when the people who knew it sat two desks over.
They assumed trust networks survive a reorg. They don't. Informal relationships, internal shortcuts, the "go ask Jeff" reflex — none of that lives in the org chart. Reorganize the boxes and you've severed the connections that were getting work done.
They assumed customer loyalty was portable. It isn't. The brand transfers; the relationship doesn't, especially when the support engineers who built it have been folded into a centralized helpdesk three time zones away. This can be even worse when dealing with an international acquirer, who may not understand the local market and the local customers.
They assumed decision velocity was a property of the team. It isn't — it's a property of the team plus the org around it. Even if every person stays, decisions now route through approval cycles, legal review, brand guidelines. The product slows down with the same engineers in their seats.
They assumed the absences were accidental. Most weren't. What the startup deliberately didn't build — the features it refused to ship, the customer segments it refused to serve, the meetings it refused to hold — was knowledge acquired through years of trial and error. The acquirer thinking that knows better may bring those things back on the table, and walk right into the landmines the startup avoided.
A honest version of an acquisition memo would say: we are buying a product we think we know how to integrate, a team we will probably lose, and a book of business that will partially churn the moment we touch the things that made sign in the first place. That is the deal. The shareholder pitch never reads that way, and the gap between the two is roughly the size of the writedown that shows up two years later.
Culture Isn't a Soft Metric
Early stages of the acquisition the acquiring team will often try to reassure the team that they will be able to maintain their autonomy and culture. They will promise to keep the team together, to maintain the culture, to allow the team to continue to operate as they did before the acquisition. That both companies cultures are aligned and nothing will change.
More often than not, the acquirer will not be able to keep their promises.
The academic literature calls it the coordination-autonomy dilemma. Acquirers have to integrate the company in order to be able to operate; they also have to preserve organizational autonomy to avoid destroying the conditions that made the company worth acquiring in the first place.
The two pull in opposite directions, and the structural incentives inside any large organization all push toward integration. Finance wants margin expansion, IT wants a single stack, HR wants one comp band, legal wants one set of contracts. Autonomy doesn't have a department to fight for it, so it loses every quarterly review.
There is plenty of data to support this:
PwC found that 65% of acquirers say cultural issues hindered value creation in their most recent deal.
Deloitte attributes 30% of failed integrations to cultural differences.
FranklinCovey estimates up to 30% of M&A value loss comes from unresolved cultural friction.
This slowly chips away at the team's ability to make decisions and ship product; burns them out and slowly kills their motivation and their ability to ship product.
I've been there myselft, the deal closes with promises — preserve the culture, maintain autonomy, keep the team intact. Then the benefits start getting trimmed inside the first quarter. The brand goes quiet a few months in. Titles get changed without anyone asking, and local leadership starts losing arguments to a head office that doesn't really know the market or the product or the people. The acquirer's processes get rolled in wholesale, not because they're better, but because standardization is easier than understanding.
The best people leave first. They always do; they have options and they use them. What's left after a couple of years shares a name with the original company and not much else.
The promise is always integration but the reality is always absorption. It happens whether the acquirer means well or not, because the structural incentives all push toward control — and control suffocates whatever made the acquisition worth doing in the first place.
The 18-Month Cliff
By month eighteen, the founders have usually left, most of the early hires went with them, and whatever product velocity made the company worth acquiring is gone. The writedown that follows is just the accounting catching up, and it usually doesn't land for another year or two on top of that. Deal models price for year-one revenue synergy; they were never built to capture the compounding mechanics that decide whether the acquired company is still recognisable twenty-four months in.
The people who leave aren't just lost headcount either. Almost a third of founders who leave shortly after an acquisition launch a new startup within six months. They take their context with them — not just the technical knowledge, but the customer relationships, the design judgement, and the hard-won understanding of why certain bets paid off and others didn't. A meaningful share end up competing directly with the company that just paid to acquire them. The purchase price effectively underwrote the next competitor's seed round.
Warren Buffett saw the broader pattern decades ago: "While deals often fail in practice, they never fail in projections---if the CEO is visibly panting over a prospective acquisition, subordinates and consultants will supply the requisite projections to rationalize any price." None of this damage was in the projections to begin with, the post-mortem won't capture it either, and the next deal will use the same model with a different logo on the cover.
What I'm Doing
Having gone through a couple acquisitions, I'm speaking from experience, not just the data or third party articles, here is what I'm doing to avoid the same mistakes:
Building culture that doesn't depend on a single owner. Culture is documented, practiced, and distributed---not held in one founder's head. If we were ever acquired, the culture should be legible enough that an acquirer could understand what they're buying and why it works. Most don't bother. Making it explicit at least gives them the chance.
Retention through meaning, not handcuffs. Golden handcuffs keep people physically present; they don't keep them invested. Retention is built on autonomy, ownership of outcomes, and career development---the things that make engineers stay because they want to, not because they're waiting out a vesting schedule.
Knowledge documentation as insurance. Institutional knowledge that only lives in people's heads is the first thing to go in any transition — acquisition, leadership change, plain old attrition. So we write down the engineering decisions, the architecture trade-offs, and the reasoning behind systems while the work is happening, not six months later when nobody remembers why we picked option B over option A.
Honest conversations about uncertainty. When somebody on the team asks about acquisition risk, they get an honest answer. People can handle not knowing — what they can't handle is being told everything's fine while the ground shifts under them. Deloitte's number is 3.5x better retention when communication is "effective," which in my experience mostly means not lying.
The bet is simple enough: companies that build cultures worth preserving either won't need to be acquired or will survive being acquired better than the ones that don't bother. Although there's no guarantee in any of that. But the alternative is the 70% failure rate, and at some point you have to try to even the odds.