Fog Creek Software
Discussion Board




"...no software company can succeed..."

"...no software company can succeed unless there is a programmer at the helm."

Bravo Joel.

  --Josh

P.S. - That one post included a great, well-written, and insightful foreward, an introduction to what looks to be an interesting book that I'll definately read, and introduced me to Doft-Letter, which seems to have some interesting stuff as well. Lot's of col stuff for one post :).

JWA
Friday, August 01, 2003

Soft*Letter, you mean. It's great stuff but a subscription is around $300 a year. Probably worth it if you're a software company executive of any type.

Joel Spolsky
Friday, August 01, 2003

Yep, that fast typing getting me in typo-land again :).

JWA
Friday, August 01, 2003

"In every high tech company I've known, there's a war going on, between the geeks and the suits."

So, who wears the suit at FogCreek?  Do Joel & Michael take turns?

:-)

Jason
Friday, August 01, 2003

I think the key is that the fellow running the company must understand BOTH technology and business. It's that critical intersection that's key.


And IMHO it's easier for a programmer to learn business and marketing than it is for an MBA to learn programming.

Also, IME (in my experience), most businesses don't even do things the right way so learning on your own can be much more effective.  Thank *god* our competitors are among the businesses that do things the WRONG way. 

Entrepreneur
Friday, August 01, 2003

Something that seems to be working for us:
1) A businessman who knows the technology, but also has a deep respect for P/L, deadlines, and customer satisfaction at the helm
2) A geek who understand business needs, but knows the tech inside and out as CIO/VP-type

AND (vital) that they both respect each other's opinions.

It's kinda like web development - my mantra is that it takes two people to build a website - an artist and a coder. If you can find someone that has the talent to do both, then buy them a Mercedes and pay them half your revenues. Barring that, it's much easier to recognize that the two talents are generally separate, and it's easier to find two experts than one who can do both.

Ditto here - there are people who can do business and tech, but they're very, very rare (and how is that person really gonna keep up with the tech while they're running the business?). Better to bite the bullet and recognize that you really need two people...

Philo

Philo
Friday, August 01, 2003

Unfortunately finding one person (or two or can work together) that combine good business skills with programming/tech skills is so rare, that most companies are not going to be able to find them.

I used to fret about what my company could do to be better, then I realised that many of the most successful companies (HP and Microsoft were my examples) had technical leaders, and I came to be at peace with the idea that most tech companies would be mediocre in this respect :)

Aaron Lawrence
Friday, August 01, 2003

Great Topic, very articulately put!

Bravo! Joel!

One thing I would add is that truly growing companies would put emphasis on Sales/Marketing or on Technology in roughly about 2 to 3 year cycles. The great companies realize this.

Leaving a technology leader who does not know business is eventually fatal to the company no matter how fast you see it growing - DEC was a great example - I was there at its peak and when being good technically wasn't enough.

Oracle is a great example of leaving marketing/sales too long in charge without taking corrective action. If you read Oracle's history their plan until V6 or V7 was "Sell-plan-make". However when there were too many complaints of quality Larry Ellison had the foresight to leave a good technology guy who did major repair and got the product to a better shape.

Seems like a good tech/business leader in charge with the power and foresight to alternate the technology/sales cycles every two to three years may build great companies while others can build good companies, sell out and sip drinks with umbrellas in them in the Carribean!

My 2 cents!

Nari Kannan
Friday, August 01, 2003

You forgot that great companies make sacrifices and try new territories. Noticed how many of Microsoft's products never pan out? Yeah, well, they tried 'em and some worked and some didn't.

That's real capitalism for you-- having the balls to go ONE step further than everyone else and MAKE your own markets.

Mickey Petersen
Friday, August 01, 2003

I agree with Joel that sucessful software companies have a good mix of technical/business leadership (whether it is all rolled into one person, or a two-person team).  Problems start to crop up when:

1. The company does not keep up with technology trends
2. The company starts adapting the business practices of its customers
3. Management becomes so afraid of change that it stops moving the business forward, letting opportunities pass by.
4. Non-technical/Technical employees stop communicating effectively.

Phil
Friday, August 01, 2003

why is it so hard to find someone who started as a developer and moved up to management?  Wouldn't it be safe to say that a LOT of managers started out as *something*, wether it was engineering, sales, analytical, manufacturing, etc.  Why the big difference with software?

vince
Friday, August 01, 2003

Other professions have their governing bodies that would prevent you from advancing unless you are a credentialed member of the profession.  You won't be calling the shots in an architecture or accounting firm unless you are a qualfied architect or accountant yourself.

Software has no such governing body, and technical people don't like to become management, so you often have non-technical managers being placed in positions where they take the big decisions, often to the detriment of the programmers and the firm as a whole.

T. Norman
Saturday, August 02, 2003

I've a certain sympathy with the idea that the driving force of a technology company should have more than a passing understanding of the technology they're pushing.

However, driving through mass market product and getting that distribution right is a little more complicated than whether you do or do not understand what a rich edit control is and whether you should do one.  I've also known and worked with a bunch of technology companies that did have technology leaders at their head and that failed.

Digital Research was still run in the early eighties by Gary Kildall as uber-techno-CEO as you could imagine (I'll leave aside the somewhat Microsoftian analysis of why DRI failed alone, other than to point out that the more than close similarity of code between MS DOS and CP/M was part of MS's OEM monopoly of the same period).

Victor 9000/Tandon Chuck Peddle and Jugi Tandon, both technologists, both patent holders of significant basic technology and both separately and together created innovative software and hardware products.

They didn't fail because they had competent technical managers they failed for a whole bunch of reasons relating to the market, capitalisation, cash flow (most commonly cash flow), and distribution of product.

Having technology leadership at the pushing point of a company can also be detrimental to its growth.  In the early 90's Noorda was trying to grow Novell as an operating system company and a personal software company.  Noorda is a salesman.  The core flagship product wasn't developed to compete directly with NT networking because the Chief Architect couldn't bear to lose his 100 byte real mode kernel and wasn't interested in network services.

Novell dropped the ball and started working for entirely conflicting purposes mostly because the technologist in charge didn't seem to understand the market and how it was changing.

Are there successful technology companies that don't have technologists either leading them or pushing them? 

No, it doesn't seem as if there are.  IBM might be the exception that proves the rule.

Are there any successful technology companies that don't have salesmen or marketeers either leading or pushing them? 

I think the answer to that is no as well.  The best companies I've worked with (which would include DRI), have paired technology and marketing, sometimes they combine in the same person.  I wouldn't say that Bill Gates was the great salesman that made MS what it is, every company needs a Monkey Boy. 

I think if there is any lesson in any of this its that having the marketing and technology thrust bound up in one person is more likely to push the organisation singlemindedly over the cliff.

Simon Lucy
Saturday, August 02, 2003

>>I think if there is any lesson in any of this its that having the marketing and technology thrust bound up in one person is more likely to push the organisation singlemindedly over the cliff.

I disagree. 

It's not about whether it's a technical or marketing person at the helm, it's about whether the person at the top truly understands (and respects) the business and the market.

Yanwoo
Sunday, August 03, 2003

Ummm how is that different?

Simon Lucy
Monday, August 04, 2003

Seems to me that complete lack of understanding of even the most basic business principles is as widespread amongst programmers as the lack of basic software skills is amongst your average collection software business tie rack in our industry. I do agree with Joel that it is this rare combination of deep tech that can grok a market that makes for a winner.

Just me (Sir to you)
Monday, August 04, 2003

I think you guys are mis-reading Joel. He's not saying that any old programmer will make a good CEO. He's saying that the CEO of a pure software company almost has to be a former developer with a good understanding of not only the technical issues, but also the business issues.

In short, the best CEO for a pure software company is a CEO who is both an excellent engineer and an excellent business person. That's VERY rare.

Brad Wilson (dotnetguy.techieswithcats.com)
Monday, August 04, 2003

It doesn't look like reading comprehension is a common skill in programmers either.

Brad nailed it, but what Joel said wasn't that hard to follow. Combine someone that has good technical skills and business prowess and you've got someone that can be a great leader to a technology company.

Go Figure.
Monday, August 04, 2003

+++I think you guys are mis-reading Joel. He's not saying that any old programmer will make a good CEO. He's saying that the CEO of a pure software company almost has to be a former developer with a good understanding of not only the technical issues, but also the business issues.

In short, the best CEO for a pure software company is a CEO who is both an excellent engineer and an excellent business person. That's VERY rare. +++

Actually, I believe the best software companies seem to have a "twin star" structure.  A good technologist and a good marketing/sales type.

For instance, MS and SAP come to mind.

Adobe as well.

Companies that don't do well seem to have "Black Holes" at their centers, people who eventually suck the company into the maw of their ego and compress the firm into an infinite point, unable to communicate with the rest of the universe.

(OK, that's a great line, if I do say so myself.)

rick

rick chapman
Monday, August 04, 2003

*  Recent Topics

*  Fog Creek Home