Category: tech

In the mid-80’s I took my second tech job, at a company called Silicon Graphics, Inc.

SGI, before it officially contracted its name to “SGI”, was a fantastic place to grow and learn. It was intense and focused, and my modest technical know-how in Unix, TCP/IP Networking, and Computer Hardware matched well as the company grew its Unix workstation business and more customers connected computers to Internet Protocol networks. I was soon a main contact point for the most intractable problems from customers, and the Customer Support representative to new product introductions.

The focus and intensity made for arguments, some that probably would seem to outsiders as knock-down, drag-out fights.  But a friend from those days put it well when he said, “We fight with each other, but we fight to make a great product.”

If I came to engineering with a problem and hadn’t done my homework, didn’t have a good handle on what was happening and why, and hadn’t collected useful information from the customer, the criticism was withering. One networking engineer, who I consider an important early mentor, was especially tough. But once I learned to do my research and bring him useful information on real problems, he was very responsive and engaged, helping me pull in other engineers on really complex cases.

My team worked with real customers every day. We had an understanding of how real customers used these computers, and the real environments they were in — the networks, the labs, and the variety of configurations, and why a problem that some might consider a minor annoyance was actually a major disruption for another customer. But often it was difficult to produce clear data supporting our insights. By the time the data was clear, the problem was too big, or at least much more expensive to fix.

This was especially the case in new product introductions. Even down to the question, “Is this ready to ship?”, we were there, contributing to the decision and wielding a rarely used veto. These arguments were often heated, but the Product Managers and Marketing and Sales Execs came to know that our concerns should be considered, even on apparently light evidence. And we knew that a big part of our job in new product intro’s was to ensure that we never needed to use that veto.

It was an environment where you were expected to know your stuff, whether on technical topics or in business. And even if you had a reputation for knowing your stuff, you would get challenged, and you had to fight it out.

As Silicon Graphics grew, it became known as the place the best people all wanted to work. Not because of that original espresso machine that Jim Clark bought for the engineering team, but because of our reputation for excellence in every part of the company.

But as the growth started to accelerate in the early 90’s, things changed. Instead of arguments about making our products and services better, about serving customers, fixing problems and getting things done, the fights became political. Which VP was supporting that position? Who are you? Whose budget is this going to come out of? In-fighting between divisions and groups and backstabbing and office politics grew. SGI had become a “big company.”

Many great people left, but many stuck around. Even if you were unhappy, the growth made it easy to find something new and interesting to do.

But the necrosis had set in.

We were no longer focused on building great products and changing the world. We were fighting the wrong fights.

—

Building a great product (or service) is never easy. Tough decisions must be made, and among smart, motivated people, coming from differing experience and perspectives, that’s going to cause arguments. And even in the best teams, sometimes the argument will go too far, or stray into personal attacks, or involve some back-channel politicking. And learning to develop an idea and gather support for it is critical to getting anything done in a large organization.

You have to keep bringing everyone back to the point: To make a great product.