What integrating a collapsing global bank taught me about the deals PE firms actually do — and why the first 100 days after close decide the next five years.
In September 2008, Lehman Brothers filed the largest bankruptcy in United States history. Within days, Nomura Securities acquired its Asia-Pacific and Europe/Middle East operations. I was on the team that had to make that acquisition actually work — technically, operationally, across multiple regions and time zones, with trading floors that could not go dark for a single session.
Two incompatible technology estates. Data centers, trading floors, the desktop environment, identity systems — all of it had to be reconciled while the business kept running under the most intense scrutiny in modern financial history. There was no “we’ll fix it next quarter.” The market imposed the timeline, and the timeline was now.
A year earlier, I had run the onsite integration when Lehman acquired Grange Securities, a fixed-income firm in Sydney. So I came into the Nomura work already knowing something most integration plans ignore: at real scale, under real pressure, you cannot do everything — so the entire game is knowing the order of operations.
Field Note
When the mandate is “the trading floor cannot go dark, and you have days — weeks would have been a luxury,” you find out very quickly what actually matters and what was only ever noise on a slide.
Fifteen years and a few hundred mid-market engagements later, I run technology due diligence and fractional CIO work for private equity sponsors and the companies they own. The deals are smaller now — a $60M distributor, a $200M healthcare services platform, a manufacturer being bought for a buy-and-build thesis. The stakes, in absolute terms, are a fraction of a Lehman. But the discipline is identical. And most of the people advising on these integrations have never had to run one when getting the order wrong meant halting a market.
Why the first 100 days decide the outcome
In private equity, the investment thesis is written before close. The model assumes a certain speed of integration, a certain ability to bolt on the next acquisition, a certain margin of operational safety. Technology is the substrate all of that runs on — and it’s the part the diligence process understands least.
What happens in the first 100 days after close either protects that thesis or quietly begins to erode it. Not with a dramatic failure. With a slow accumulation of the wrong decisions: a system ripped out before anyone understood it, a security gap left open because no one owned it, an integration rushed to hit a date that had no business being the priority. By month nine, the “tech problem” everyone’s frustrated about was actually a sequencing problem in week three.
The order of operations that survives contact with reality
Here is the sequence I run, whether the company has 90,000 users or 90. The scale changes. The order does not.
Step 1 · Stabilize what generates revenue
Before anything else: what is one outage away from stopping the money? In a bank it was the trading floor. In a mid-market company it’s the ERP, the order system, the thing the business literally cannot invoice without. You find the single points of failure and you make them boringly reliable. Nothing else matters if revenue stops.
Step 2 · Secure what could kill the deal
The exposure that turns a good acquisition into a claim or a headline is almost always the same short list: one over-privileged admin account shared by several people, a vendor with deep access and zero oversight, and no tested incident response. None of it shows up on a security questionnaire. All of it shows up in a breach, a stalled integration, or a discount at exit. Close these before you do anything creative.
Step 3 · Map before you touch
The fastest way to break a newly acquired company is to rip out a system on day ten because it looks old. At Lehman-into-Nomura, “old and ugly but load-bearing” described half the estate — and you did not touch load-bearing walls on a hunch. You map what each system actually does, what the thesis needs it to do in twenty-four months, and where the gap is. Only then do you have the right to make changes.
Step 4 · Then, and only then, build the roadmap
By now you’ve earned the information to make good decisions instead of expensive guesses. The roadmap is where value creation lives: consolidation, the platform that enables add-ons, the automation that lifts margin. But a roadmap built in week two is fiction. A roadmap built after you’ve stabilized, secured, and mapped is a plan you can actually stand behind in front of the investment committee.
The three mistakes that quietly cost the most
Across bank-scale and mid-market work, the failures rhyme. Three come up again and again.
1. Speed mistaken for progress. Ripping things out early feels like momentum. It’s usually just expensive motion that creates the next fire. Speed without a map is the most reliable way to destroy value in the first 100 days.
2. “We have an MSP, so we’re covered.” A managed services provider is a pair of hands, not a CIO. It executes decisions; it doesn’t make them in the owner’s interest. If a portfolio company’s technology strategy is “whatever the vendor recommends,” there is no strategy — there is a vendor optimizing for the vendor.
3. The hero dependency. The most dangerous thing in a mid-market target is often one person who understands the one system everything depends on — and who may not survive the transition. I’ve watched a clean-looking deal seize up because the institutional knowledge walked out the door. You find that person in week one and you de-risk them immediately.
Operator, not advisor
There’s a reason I lead with the integration story. An advisor writes the report. An operator knows which line in the report is going to become a 2 a.m. phone call — because they’ve taken that call before, when the whole world was watching. The value isn’t the pedigree. It’s that the pedigree was earned running the thing, at a scale where there was no margin for getting the order wrong.
That’s what Vertex CIO brings into a mid-market deal: not a questionnaire, but the judgment of a team that has run production infrastructure daily and integrated it under maximum pressure. The stakes are smaller now. The discipline is exactly the same.
