Hacker News Reader

Comment by quaunaut | Hacker News Reader

Well, folks are already off to the races reading the headline and not the core of the story.

But, this is something I've thought of often over the last several years. If perhaps, a better angle would be to build a business around Unix-style philosophy, of "Do one thing, and do it well." Then, you simply start over with another brand new, small thing, and grow it from there, while lending it the HR resources of the core.

The closest company to trying something like that, that I know of, is Google. But even there it seems they've hit some stumbling blocks.

View on HN

When I worked for Microsoft, this part: "Then, you simply start over with another brand new, small thing, and grow it from there, while lending it the HR resources of the core."

...never worked. Because the core businesses (Windows and Office) were $10B products. Investing a little bit to get a 0.1% move in revenue there, brought in $10M more. It was bloody difficult to persuade anyone that the $ should be invested in a brand new startup idea, rather than re-investing it in the core, growing business. And who can blame them?

There were two businesses that mattered, and everything else was a distraction. Actually after a while, the combination of Windows Server and SQL Server became a business that mattered, too.

In those days, everything else was a weed that might/could grow, if it could shoot up and grab some attention and look pretty, before the groundskeeper just snipped it off at the base.

If you look at what succeeded at Microsoft since then, it has not been "investing in some some small thing." It has been MASSIVE investments in things, like Azure and Xbox. Of course, MSFT also made MASSIVE investments in plenty of things that did not work out.

But the point is, proposing to start with a brand new small thing, was a sure way to be asked to get with the program, or leave the company.

Well, small things are very competitive, everyone can do them and do them better than a large corporation optimized for innovation-killing efficiency. I guess it's only natural for a corporation to go big and do massive investments in big things, because there is almost no one on that level to compete on innovations, no matter how poor they are.

But a lot of that would depend upon support structure or company architecture, right?

After all, if the idea of starting something new gets you asked to leave the company, the company itself is probably hostile to new proposals, so it's fundamentally against what I'm talking about.

I don't disregard what you're saying- but it's a huge mentality shift, to be able to do things like this.

I also think it'd be beneficial to run such an organization in a smaller side-company, similar to what Alphabet does, than merely as a new department inside of a huge monolith like a Microsoft. Otherwise, you risk just this sort of thing, and being able to consistently maintain a culture of creation in something meant to last a century or longer is incredibly difficult.

It's quite unlikely you've heard of Eckart Wintzen [1]. He was leading one of the fastest growing software engineering companies in the 70's and 80's called BSO using a 'cell management', where sub-companies have a set maximum size of 50 and need to split afterwards. I worked for a company called Topicus [2] in the Netherlands which followed a similar philosophy for a while.

[1] http://www.extent.nl/about-us/people/eckart/ [2] http://www.topicus.nl

Similar models exist in big business under the name of divisional organizations and matrix-style management. The latter is nowhere near as cool as it could be with some minor changes involving VR and AI's. Yet, it worked in real-world businesses well enough.

W.L. Gore and associates does something like this:

https://en.wikipedia.org/wiki/W._L._Gore_and_Associates

(The company is famous for Gore-Tex. The structure is famous in part because Malcolm Gladwell used it as an example in his discussion of Dunbar's number in The Tipping Point. But I guess it was studied lots before then.)

> while lending it the HR resources of the core.

"HR resources" is redundant.

> HR resources" is redundant.

No, its not. It does expand out to have a repeated word, but the reference is different (the HR resources -- Human Resources resources -- are resources dedicated to the Human Resources function.)

? Why is it redundant? For the love of God, why is HR redundant?

Let alone the other answer ("HR" == "other engineers"), which is brilliant, the HR department is the greatest weapon a large firm can deploy. Do you want insurance in case your kids get sick? Do you want days off to relax after a month of hard work? Do you want to ensure your boss ain't making inappropriate advances?

Why on earth would a smart people want to be burdened with all of those works? Why on earth would a smart people hitch non-core work onto main work hour? Why on earth would a smart person refuse to delegate?

"HR resources" is redundant. Such arrogance.

Reply on news.ycombinator.com