HP Didn't Plan Its Best Expansions. Its Customers Pulled It There.

In 1966, HP built a computer it did not really want to sell. The HP 2116A was made for one job: to control the test-and-measurement instruments HP already shipped to engineers and labs. It was a tool for running other HP tools.¹ The general-purpose computing business that followed was, in the original plan, a byproduct - an accident HP stumbled into because the customers it already had needed something next to what they were already buying. That accident contains the whole strategy. HP's best expansions were never bold leaps into new markets. They were short steps into the room next door, with a customer already standing in it.

The official story is that HP is a visionary that kept reinventing itself - instruments, then computers, then printers, then enterprise IT. The truer story is quieter and more useful: HP grew best when it solved a problem sitting right beside its installed base, and it stumbled badly the moment it chased scale for its own sake. The HP 2116A wasn't a pivot. It was a chore that turned into a market.

The expansions that worked started where the customer already was

Look at what actually stuck, and a pattern emerges. The HP-35 scientific calculator arrived in 1972 at $395 and sold more than 300,000 units within three years - and it was that product, a handheld for the same engineers who bought HP's instruments, that accelerated HP's pivot toward computer technologies.¹² HP wasn't winning a calculator war; it was selling the same person a smaller, smarter version of the precision they already trusted HP for. Each move was an adjacency: the new product borrowed the old customer, the old reputation, and the old distribution. The risk wasn't 'will anyone buy this in a new market' - it was the far smaller 'will the people who already buy from us buy this too.'

Contrast that with the time HP tried to leap instead of step. Its first desktop PC, the HP-85 in 1980, was a commercial failure - because it was incompatible with the IBM PC standard.² HP had built a computer for a market that had already chosen its rules, and HP's installed base of instrument customers couldn't save a product aimed at general consumers who'd never heard of HP for PCs. The leap missed. Then came the move that defined HP for a generation, and it wasn't a computer at all: the LaserJet, launched in 1984, became Hewlett-Packard's single most successful product.² A printer. The thing HP's customers needed to do with their computers, sitting in the exact room next door.

First HP cut the core away, then it bought a giant

By the late 1990s HP had become two companies wearing one name: the original test-and-measurement business it was founded on, and the computing-and-printing empire that had grown up beside it. So HP did the disciplined thing first. In 1999 it carved out the test, measurement, chemical analysis, and medical businesses into Agilent Technologies, filing the separation with the SEC.⁴ Worth noting because it is routinely told wrong: the 1999 IPO sold only about 15.9% of Agilent at $30 a share. HP held onto 84.1% until it distributed the rest on June 2, 2000 - Agilent's real independence date.³ HP had shed its founding business to bet everything on computers, printers, and the enterprise.

Then it did the opposite of stepping into the next room. On September 3, 2001, HP announced an all-stock merger with Compaq valued at roughly $25 billion, with Compaq holders receiving 0.6325 of a new HP share each.⁵ This was not adjacency logic. This was scale logic - a bid to out-mass Dell in a commoditizing PC market by bolting two struggling computer makers together. The question was no longer 'will our customers buy this too.' It was 'will being bigger fix being unprofitable.' Those are not the same question, and the difference is the whole story.

A merger that won by just 45 million shares out of 1.6 billion cast

The market knew. Because it was an all-stock deal, the value moved with HP's share price - and during a brutal proxy fight, that price fell. The roughly $25 billion announced in September 2001 had decayed to about $19 billion by the time the deal closed on May 3, 2002.⁹ Same shares, same companies, billions evaporated while shareholders argued about whether the thing was worth doing at all. That is the sound of a market voting with its wallet before it votes with its proxy.

And the proxy vote was a knife's edge. It was widely described as one of the most divisive merger battles in recent memory, and it turned on borrowed muscle. Deutsche Bank's asset management division had initially voted 17 million HP client shares against the deal - then reversed those votes at the last minute after HP paid Deutsche Bank's investment banking unit $1 million, with another $1 million contingent on the deal succeeding. The SEC later fined Deutsche Bank $750,000 for failing to disclose that conflict.⁶ When a merger's survival depends on a bank flipping its clients' votes for a fee, the merger was never being pulled forward by a customer. It was being pushed through by an ambition.

The fair objection: HP survived, so didn't Compaq work?

Here's the honest counter. HP did not collapse after Compaq - it became, for years, one of the largest PC and printer companies on earth, and you could argue the scale genuinely bought it survival in a commoditizing business where smaller players got crushed. Maybe the brutal logic of PCs really did demand mass, and HP's leadership saw that clearly when its critics didn't. That objection deserves respect, and the proxy fight's narrowness cuts both ways: half of sophisticated shareholders thought it was right.

But notice how the story ends. In October 2014 HP announced it would split itself in two, and on November 1, 2015 it spun off Hewlett Packard Enterprise as an entirely separate company, distributing 100% of HPE's shares to existing holders.⁷ The empire that Compaq's scale was supposed to fortify was, thirteen years later, deliberately taken apart - separated back into a printing-and-PC company and an enterprise company, the two halves untangled from each other. A merger justified by the power of being one big thing ended with HP concluding it was worth more as two smaller ones. Scale was the argument, and scale was eventually the thing HP gave back.

HP's history is usually read as a parade of reinventions. It is better read as a single recurring test, passed and failed in public. The 2116A, the calculator, the LaserJet - HP won every time it expanded into a need its own customers already had, one room over. Compaq was the time it forgot the customer and chased the scoreboard instead, and it spent the next decade discovering that a bigger company is not the same as a better one. The adjacency was always the moat. The leap was always the loss. HP's genius was never that it kept reinventing itself - it was knowing, most of the time, exactly how far the next step should be.