[shepard]

I predict we will see more of this.

There was a reason for conglomerates in 1960-70s: shortage of capital meant that cross-subsidizing one division by another was the only viable way to finance growth. These days are long over.

Nowadays, it makes much more sense to operate unrelated businesses as completely independent companies, with different set of managers and investors.

[pinkyand]

Not sure. A lot of the success of Samsung in the smartphone market is due to vertical integration. They save around 10-15% of bill-of-materials cost reduction due to this, they sometimes have unique components(like being first with HD-oled display), they have more control over manufacturers(as a condition for using a Qualcomm processor, they wanted the processor to be made at a Samsung fab), They are less exposed to tight supply in components like ram/flash, And who knows what other benefits they have due to this.

Those advantages are a big reason to why they have better access to carriers - meaning much better marketing - and the reason behind their ability to push volumes, create brands, and manufacture so many variants to compete better than others.

[icambron]

Vertical integration makes good sense, but I don't think that's the situation HP is in. This is more of a horizontal split up. It does mean they can apply less collective pressure on component providers, but only to the extent that those business have overlapping needs. So one way to clarify the parent's point is that businesses should split up when they sell to different markets. I think the previous hesitation to take that step was the degree to which HP actually did have some (not hugely convincing) sales synergy between its server and printer businesses.

[jotm]

HP is so far behind Samsung in the consumer market that it hurts thinking about it :-). They failed at smartphones and tablets, their computers are kwap, the printers are absolutely awful. Their strength is (was?) enterprise and business hardware from what I can tell.

[koko775]

A relative of mine with credible information tells me HP is apparently losing a bunch of ground in the server market because of low build quality. HP's trying to compete on prices by slashing them, but their deployments are just that bad.

[enjoy-your-stay]

It's been a real shame to watch the decline of HP.

In the early 90's when I was just starting out in IT, their brand was very highly regarded, especially their laser printers which phenomenal workhorses.

[ghaff]

They also made a huge bet on Itanium which obviously didn't pay off. HP, along with Intel, has arguably recovered from that strategic decision, but a huge amount of money and corporate energy went into running that play. If that money and energy had gone into other things, HP might well be in a better overall place today.

[q2]

In one of the companies I worked in the past, we found that for a similar server configuration, Dell provides lower quote than HP and others almost always. So I doubt, HP can compete on prices with Dell in servers and support.

[jotm]

Ouch. They used to offer some awesome workstations a few years back, but the restrictions on peripherals were insane - I mean, what idiot decided to implement whitelists on processors and PCIE graphics adapters?

[shepard]

Indeed, you are bringing very interesting points:

- vertical integration (and tighter control over quality / schedules) for technology companies, and

- size/scale as an advantage in Marketplace.

We have seeing both of these in play, not just with Samsung: Apple is designing their own CPUs, and you can argue they leverage their size to help iTunes and Apple Pay businesses.

Still, there are counterexamples: NVIDIA (and many other successful companies) are fabless, and AMD divested their fabs (by splitting into AMD and Global Foundries).

What HP seem to be spinning off is their consulting arm and the enterprise hardware businesses, where vertical integration/size doesn't matter that much.

[shmerl]

It can have its downsides. For example Nvidia didn't manage to push 16nm chips into their latest Maxwell cards because plants that they use for manufacturing were already booked ahead. Intel on the other hand use their own plants and can decide things more flexibly.

[sjg007]

If anything at scale you should own the whole chain.

[giovannibajo1]

Not sure the counter example of Apple stands; they have a specific goal: be in full control of the whole technological stack for their products. They don't design processors and sell them like Qualcomm; they just do whatever is necessary for their product line.

A conglomerate like HP and Samsung often has little to no cross-pollination between divisions. Obviously the semiconductor part of Samsung can benefit the whole business, but the intersection between TV and dishwasher is probably zero, even at marketing level (they have totally different marketing and sales departments).