Friday, January 25, 2013

Thoughts on "What Happened"

Forceful reminders come back in early morning hours.  I'm doing some work around the early deployment of routers (think Cisco, Wellfleet), and found a passage in Ed Paulson's 2001 book, Inside Cisco, that resonated re HP.  Chambers developed a very specific M / A strategy, essentially based (he was reported to say) on the HP model -- buy small, keep all the people, assimilate into YOUR culture.  Specifically, DO NOT ATTEMPT a merger of large companies, there is no merge of "EQUALS".   And, aside from Sanborn, the first thirty-three acquisitions HP made were all small technology plays compared to HP's size at the time.  The first one, Moseley recorders, which started the phenomenal printing and imaging world for HP, was a $400,000 down payment on a $2 million company.  When Moseley was bought out in 1963, HP was sixty times larger.

Paulson's book spends much time on the merits and issues of M / A, which at Cisco is called A / D.  The notion is that Cisco did little R for the traditional R / D, instead acquiring key technology and technologists via small companies that had proven new ideas.   This approach got a lot of traction, and in fact became known widely as "the Cisco Way", titled "A / D" in place of "R / D".

The words I particularly liked are on pp. 138-39:  "A merger of equals is not a good starting point for a successful acquisition..... A large organization will have its own set of integrated cultural ideas, standards, and rituals, which will likely differ from those of (yours).   If the merged organization is to appear as one after the merger, then it is important that the cultures somehow blend into a hybrid.....  If this does not work out smoothly, the infighting, squabbles, and daily distractions with a poor cultural merging can make the combined entity far less valuable than the total of the two separate companies, ... in essence, eroding the values of the two previously healthy companies (particularly true with respect to the interactions between managers of the two companies)...."

Sound familiar?  From our book, The HP Phenomenon, p. 465, there is a heading "Merging Two Cultures" which describes the extraordinary effort led by Webb McKinney to blend the Compaq group with HP in 2002.  The difficulty in part was that Compaq itself was a heady brew of three disparate cultures -- Compaq's PC lines, DEC's mini lines, and Tandem's Non-Stop lines; each had 'quarrels' with the respective HP units, which themselves had come through a series of cultural wars for reasons having to do with HP's own evolution (think Fort Collins and its 'war' with acquired Apollo; and the classic Cupertino/Ft.Collins battles).  Throw in the multiple contenders for PC hegemony at HP, and the merging seems daunting in retrospect, as well as from the savage civil war that broke out between oldtimers and the new Carly-led team.  This is years before the challenge from Hurd's EDS 'merger'.

On p. 501 in our book, it is noted that a long-term cultural company -- HP -- divided 125,000 folk, sending 52,000 to spinout Agilent, and three years later, adding 65,000 Compaq folk, and then five years later adding 140,000 EDS folk.   So in nine years, at most there were 60,000 "HP Way" people in a company of 310,000 bearing heavy doses of The Compaq Way, The DEC Way, and the EDS Way.  And the odds were almost nil that these were aligned with Dave and Bill's view of dignity and personal contribution at all levels of the company.

Leaders as solid as Dave and Bill would have found this task extremely hard; Fiorina and Hurd certainly didn't bring the Bill 'n Dave HP Way views and leadership approach to the challenge.

Make sense?  Large scale mergers seldom work.

Webb McKinney's forthcoming book will hopefully shed some insider perspective on the huge effort he led to integrate Compaq into HP; it'd be great to get Anne Livermore's views on the EDS merge.




3 comments:

Anonymous said...

Chuck --

While I can't comment on Services or PCs, from my 'in the trenches' view of the Compaq acquisition with the DEC parts of Compaq went pretty smoothly.

Digital had a very similar culture to that of HPs...they had a different name for it of course but is was built around the ideas of engineering excellence and consensus. The DEC folks got us in HP-UX networking and we got them as well. I'm sure there were bumps but not anything like on the PC side of thing. The DEC folks respected our point of view and we theirs -- we had many more strengths in networking than they and they viewed partnering with us as a chance to build a better whole from two different parts. DEC had...HP had IPv6 and they had TruCluster...each party had something of equal value they could bring to the table.

From the (hearsay) reports on the PC side of things from friends of mine who worked there at the time, the situation was all-out war...on top of that, Carly had given the lead in PCs and PC servers to the Compaq people so the HP folks were just the implementation lackeys for other people who were actually running the show. Most of us in UNIX networking were pretty glad we didn't work there (not that we would have wanted to work there to begin with).

BTW, after completely integrating TruCluster into the HP-UX kernel and networking stacks, Carly decided that MC/Serviceguard was going to be the clustering solution of choice. Most of us were aghast...in our collective opinions, TruCluster was so much more superior. I guess the MC/Serviceguard people had better lobbyists with the higher ups. We were given 1 month to back out 18 months worth of work from the kernel and networking source trees.

Steve Witten

Anonymous said...

Chuck --

While I can't comment on Services or PCs, from my 'in the trenches' view of the Compaq acquisition with the DEC parts of Compaq went pretty smoothly.

Digital had a very similar culture to that of HPs...they had a different name for it of course but is was built around the ideas of engineering excellence and consensus. The DEC folks got us in HP-UX networking and we got them as well. I'm sure there were bumps but not anything like on the PC side of thing. The DEC folks respected our point of view and we theirs -- we had many more strengths in networking than they and they viewed partnering with us as a chance to build a better whole from two different parts. DEC had...HP had IPv6 and they had TruCluster...each party had something of equal value they could bring to the table.

From the (hearsay) reports on the PC side of things from friends of mine who worked there at the time, the situation was all-out war...on top of that, Carly had given the lead in PCs and PC servers to the Compaq people so the HP folks were just the implementation lackeys for other people who were actually running the show. Most of us in UNIX networking were pretty glad we didn't work there (not that we would have wanted to work there to begin with).

BTW, after completely integrating TruCluster into the HP-UX kernel and networking stacks, Carly decided that MC/Serviceguard was going to be the clustering solution of choice. Most of us were aghast...in our collective opinions, TruCluster was so much more superior. I guess the MC/Serviceguard people had better lobbyists with the higher ups. We were given 1 month to back out 18 months worth of work from the kernel and networking source trees.

Steve Witten

jjones said...

No matter how you dress it up, HP will struggle for years with the albatross of EDS. It is not enough that morale at HP is bad, but it is awful at EDS. In the I T services sector, you need good morale, inspired leadership, quality and price. I think it accurate to say that EDS lost these qualities years ago, as did HP. Sort of a marriage of two very ugly people doesnt make either one of them better looking. Most of the staff are busy looking for other jobs and doing just enough to hang on. I dont believe that the new leadership is going to do much more than get hefty parachutes when they leave. The focus on staff and customers left this company many years ago, that is why the leadership of EDS bailed out with pockets full of HP loot.