Archive for September, 2008

Geschke on corporate culture

Want to predict how Adobe will act? Look at its past. Groups develop different cultures, and knowing how a group “works” gives big clues on its future.

The following quotes from Adobe co-founder Charles Geschke appeared in a lengthy interview from Knowledge@Wharton. Here I’ve pulled together some of the information on corporate culture, as in the prior posts on standards and reinvention.

The earliest impulses behind Adobe were entrepreneurial, about the need to make a real difference in the world:

Knowledge@Wharton: What prompted you and Warnock to leave Xerox?

Geschke: Xerox loved [what we had done]. They said, “We’ll make it a corporate standard.” I said, “Great!” I went to Connecticut to [Xerox] headquarters and said, “I’m here to talk about the marketing plan for rolling this out.” They said, “Oh, wait a minute. At Xerox it takes seven years to develop a product — and we can’t talk about this because other people will get the idea and beat us to market.”

I said, “Seven years! [In the] industry you’re about to go into, that’s two to three generations. By the time you bring it out it will be worthless.” [They said,] “Sorry, [it takes] seven years at Xerox.”

John and I were frustrated. It turned out that his thesis adviser at the University of Utah was on the board of the investment bank Hambrecht & Quist. Bill Hambrecht was one of the first guys pushing venture investing in high-tech companies….

There were strong lessons, early on, about the need to continually re-check assumptions:

Knowledge@Wharton: When did you enter the Japanese market?

Geschke: We licensed type technology for the Japanese market in 1986 or 1987. [It was] almost the identical situation as with Compugraphic and Linotype.

Originally the Japanese typesetting industry had only one company before World War II. The two guys who ran it got into a disagreement over the Japanese conduct of going to war. One guy was a supporter [of the war effort] and the other was not. So they split the business. One guy stayed in Tokyo and his business took off, because he was at the seat of power. And the other guy took his part of the business to Osaka. He did okay, but he was struggling.

We first went to the Tokyo branch. At that point it was run by a woman, which was quite unusual in Japanese industry. She was the widow of the fellow who had stayed in Tokyo. I could never get to the second cup of tea with her. She just didn’t want to talk. She had 90% of the business and she didn’t need anything new-fangled.

So we went to Osaka and talked to the other company and, again, because they wanted to get into a stronger position and they intuitively knew the industry was going to change, we did a deal with them. We got the license to their type library and they now have 80% to 90% of the business.

Knowledge@Wharton: As was the case with Linotype, the incumbent wasn’t interested, but the underdog was. And it completely changed the fortunes of both companies.

Geschke: I think if you did a study of business deals, it’s almost always that way.

The companies that think they’re in control of their long-term destiny don’t want to talk to a young start-up company that’s going to change the rules of the business.

Part of the core philosophy is to strive to treat people fairly, and to also be perceived as treating people fairly. And when there’s a challenge, it needs to be surmounted:

Knowledge@Wharton: Okay. What was the response to the Apple-Microsoft announcement at that Seybold Conference?

Geschke: I think if you were to talk to Adobe customers in any era, one of the things they’ll tell you is that, while they may have a disagreement about this or that, by and large they were treated fairly. They could depend upon getting great technology licensed to them on a fair and equitable basis. Customers like that. If you treat them the way you like to be treated, they sense that. That was always part of our core philosophy as a company.

At that Seybold Conference, when the announcement was made between Apple and Microsoft, the Seybold people immediately changed the schedule to put a panel on the last day to discuss all of this. Before the panel began, the moderator got up and said, “I want to take a straw vote. I want everyone to raise their hand who thinks they would rather have Apple and Microsoft take over this segment of the business.” There were a few hands raised — I’ve always believed they were Microsoft and Apple people — but, basically, no one voted against us. They all wanted Adobe to continue to be their vendor.

We brought ATM out within about 60 days and sold hundreds of thousands of units in the first quarter, which was a lot in those days. It was three years before Apple and Microsoft shipped TrueType.

A business needs to cannibalize itself, otherwise it will be cannibalized by others:

If you’re lucky enough to get a “franchise” product like Photoshop, you can’t become complacent. For example, when we brought out PhotoDeluxe we were initially the strongest competitor for Photoshop. Rather than cede the low end of the market to someone else, we took it. And that allowed us to upgrade people. With hindsight it was the exactly the right thing to do — not give someone the ability to undercut you, come in and be “good enough.”

We learned that from the LaserJet. [Although a] PostScript [printer like] the LaserWriter was a much better product, eventually the [Hewlett-Packard] LaserJet was good enough. [Although PostScript is] still a profitable business.

Early experience in executive succession showed the necesssity of growing and trusting the “corporate culture”:

We kept a pretty lean organization with a pretty shallow management structure. We frankly hadn’t done work trying to groom anybody [as a successor], because we weren’t thinking about it.

So — this was not our brightest decision — we decided we could recruit replacement talent, spend a couple of years with them, get them up to speed, and their natural abilities would take over.

We hired three or four very bright people, [people who were] very successful in their previous businesses and had the right pedigree to run a business like Adobe. We figured we would meld them into a team.

Well, each one of them thought that he should be the future CEO of the company. Rather than melding as a team, they fought like cats. It became miserable. It was impossible to have a staff meeting without it breaking out into angry disputes. It was getting ugly. Because of people trying to build turf, spending was getting ahead of revenue. And in the summer of 1998, the Japanese market went off a cliff for six months. It just stopped. I’ve never fully understood why it happened. And we found ourselves having to preannounce a very bad quarter and dealing with these people who weren’t getting along.

John and I decided that we would fire all of them on the same day.

We identified the person — Bruce Chizen — we wanted to groom from the inside…

[Corporate succession] is hard. If you want to give real value to your shareholders and build a company that has a long life, you realize pretty quickly that that is a major problem and you’ve got to start working on it. We should have been brighter and known that we couldn’t go outside; you have to do it from the inside.

I have always believed that the principles and values of the company you build is an important ingredient in its success. You can’t get someone to come in from the outside and just pick that up in a six-month training period. It doesn’t work that way. Not everybody agrees with some of those principles — they think they have to do it a different way.

How to have a group endure and be successful over time, despite different personalities, different problems, different opportunities?

Knowledge@Wharton: It seems like you’ve tried to maintain the consistency of Adobe’s corporate culture even as the company’s technology focus has evolved.

Geschke: We really didn’t start emphasizing [corporate culture] until roughly 1989.

When we began, John and I wanted to build a company that we would like to work in. Why would you build a business you didn’t want to work in?

We both had enough experience that we knew there were a lot of things we felt we shouldn’t do and had some ideas about things that we thought were important to do. But we didn’t really start verbalizing it until it became clear that, as we got bigger and we had to confront tough things, we needed to make sure that everybody understood and could help us enunciate what those principles were, what that culture was about, and make this ingrained in how the business ran…

One of the things I talk a lot about is the necessity to juggle all of the constituencies that have an interest in the business: shareholders, customers, employees, vendors, and the communities in which we operate. Those constituencies are all mildly in conflict with one another in terms of what’s best for them. Your job as a leader in a company is to find an appropriate way to juggle those conflicting interests so everybody feels like they’re getting a fair deal, without letting any one dominate the others because they’ll drag your company down.

(I like that final quote an awful lot, so I BOLDed it.)

When you’re making bets on technology, it helps to predict what might happen in the future. Adobe may some day act unfairly, that’s true. It’s also true that Microsoft may some day act benevolently, or Apple some day may act openly… Google may some day act respectfully, or the W3C may act efficiently. All these things are possible, but each group of people grows a culture which affects their future decisions.

The future isn’t always predicted by the past, but the past always predicts the future.

Geschke on corporate reinvention, betting the business

Want to predict how Adobe will act? Look at its past. Groups develop different cultures, and knowing how a group “works” gives big clues on its future. This Knowledge@Wharton interview with Adobe co-founder Charles Geschke can give you deep background on how the people of Adobe look at things.

Yesterday I pulled out a section on core history with standards… the “open” model with PostScript, PDF and SWF is very similar to the model with HTML, JavaScript and CSS. Here’s a section on how the company has reinvented itself a few times, from turnkey systems provider to PostScript licensing to shrinkwrap software for a consumer channel, then enterprise, and now the various web models.

The early company was based on the idea of providing complete digital printing systems to the types of large companies which were then investing in early computers for database and processing:

Bill Hambrecht was one of the first guys pushing venture investing in high-tech companies. We told him our plan, which was to build a complete turnkey publishing system — the computers, the printers, the typesetting equipment, everything — and sell it to the Fortune 500 so they could bring a lot of their production work in house. He liked the idea because he hated financial printers. Every time he did a prospectus he felt that he was being robbed. He agreed to invest $2.5 million over two years in two equal payments and told us we would have to quit our jobs….

But once they started meeting with potential customers and partners, a need arose which was greater than that for printing systems integration. Both Digital Equipment and Apple had computers themselves, and had printer deals, but were stuck at getting the two to work together — to be able to create and edit a document with a computer at screen resolution, and have it be understood by a printer at high resolution.

Knowledge@Wharton: This is how Adobe began as a printer software company?

Geschke: Yes. We developed the [PostScript] language after we left PARC. What really caught people’s attention was what we were doing in terms of rendering type on-the-fly from outlines [the curves used to define the characters of a typeface at any size or angle].

Popular mythology at the time said that couldn’t be done; you have to hire a bunch of monks to build bitmaps. We knew that wouldn’t fly because we really wanted [to do] arbitrary expansion and contraction [of the typefaces]. You could never have enough bitmaps or enough typefaces to satisfy the printing world. It was just infeasible. You had to do it from outlines. We eventually developed software that did that and continue to use it today.

(It’s funny… what PostScript did with curves-to-dots across early computers and printers, Flash did with vector rendering in the early web-browsers. It’s hard to remember today, but the big reason people first started noticing Flash was that it used “scalable vector graphics” for smooth drawings regardless of resolution, offering very small filesizes for large drawings of low detail.)

PostScript was cloned and resold by competitors, but the Adobe implementation finally ended up as the most important in the printing ecology it spawned. But at this point a new need emerged: the need to be able to actually design with the new publishing technology. A new type of business needed to be invested in and developed, that of software which could be sold to individuals. This was not easy:

Knowledge@Wharton: How was it to go from being a company that sold printer software to OEMs [original equipment manufacturers] to being a shrink-wrapped desktop software company with products like Illustrator and Photoshop?

Geschke: It was a “come to Jesus” moment inside the company, because all the profits were being made by PostScript. As we began to invest in not only development, which was relatively inexpensive — just a few people — but more importantly in building a sales and marketing presence in the retail channel, we were chewing up resources. It was difficult in the early days to demonstrate profitability in that business. Every time we would have a budget debate, you can imagine how that went: “I’m bringing in most of the money. I should be getting these resources.”

Knowledge@Wharton: What motivated you to stick with this second line of business?

Geschke: One of the obvious lessons in business is that you can’t continue to be a one product company and survive. We also felt we couldn’t be a one channel company and survive. We were captive of those OEMs [of PostScript printers]. If they ever decided to drop us, we [would have been] toast. We had no channel.

John’s wife was — and is — a graphic designer. John had a second job [when he went] home at night writing PostScript code to do what she wanted to do on the LaserWriter. He got a couple of engineers to come up with Illustrator. We launched Illustrator and it did very well.

We realized then that we spent a lot of money to build a retail channel and a sales force and we had to feed it. There had to be more products….

Turnkey provider, to printer software developer, to digital creative tools sold at retail. The next jump was to dealing with entire documents across all types of machines, rather than just isolated files. The Portable Document Format and Adobe Acrobat were the next round of investment, development, and ultimately worldly innovation:

Knowledge@Wharton: For a long time it cost Adobe more to develop and market Acrobat than the revenue it returned.

Geschke: Oh yeah, the antibodies inside the company were just all over it.

Knowledge@Wharton: People wanted to kill the product?

Geschke: Of course, because they said, “Look, we’re selling all this Photoshop. We’re selling all these printers. Why the hell are we investing in this thing, and giving it away?”

Knowledge@Wharton: Why did you continue to invest in Acrobat for so long? What gave you the confidence to continue to pour money into it?

Geschke: Your own instincts. You can’t analyze a market that has never existed.

This theme, of radically changing the company to fulfill what the future requires, continues today:

We instinctively knew that was where things were going. The only business book that John and I ever read, and the only chapter that I remember from it was a chapter entitled “Market Gap Analysis.” The one idea I remember was that it is easier to build a business if you find a new solution to an already perceived problem that no one has come out with before — because you instantly are [at] 100% market share.

We did that with PostScript, we did that with Illustrator, and we were going to do it with Photoshop even though we knew we were going in early. And it has worked out very well.

We did the same thing with Acrobat. We brought Acrobat out about three years before the Internet really took off, thinking that local area networking would be enough to support it, but it wasn’t.

And we’re trying to do the same thing with AIR [the Adobe Integrated Runtime development platform for cross-operating system software]: Bring it out before anyone else has the idea of the concept, get the platform established, and then shame on us if we can’t make money off it.

Knowledge@Wharton: While the technology has changed greatly over the past two decades, there appears to be a certain consistency in Adobe’s business strategy.

True. And where does this lead?

Knowledge@Wharton: What do you think is the biggest challenge Adobe is facing going forward?

Geschke: Inventing the future. We’ll never succeed unless we continue to open up new vistas.

I honestly believe that our technology and what’s happening in the market — where essentially all visual communication is going to the web — is the sweetheart point in our whole envelope of products and technologies. Shame on us if we can’t figure out a way to take advantage of that shift in the way the world is moving with the distribution of information.

A lot of what are there today — the limitations of browsers and of the web imaging standards — are things that we think we have a solution for.

There’s indeed “a certain consistency” in all this. It’s “inventing the future.”

From seeing the need for general digital publishing, to changing the business to focus on a gap in the ecology, to expanding outwards to the creation of standalone software and a retail channel, to the building of a document business, and now to the building of a universal media-application platform… each “molting” of the business required a clear commitment to the next generation of growth… betting big on that inevitability which cannot yet be proven. You can see the newest stage of this evolution today.

Adobe invents new things. Its natural disposition is to generate solutions for new problems, to anticipate needs for which there is yet no solution. Adobe then opens up the results, and makes them universally accessible, as the very key to its success. When one generation of problem is addressed, the company bets hard on solutions for the next important problem.

If you want to know how Adobe will act in its future, then look at how consistently it has acted in its past.

Geschke on practical standards, the necessities and risks

Want to predict how Adobe will act? Look at its past. Groups develop different cultures, and knowing how a group “works” gives big clues on its future.

I’ve been reading and re-reading this interview with Charles Geschke since it appeared last week. There’s a lot that’s pertinent to today’s issues.

Here they discuss how PostScript… well, read it yourself first, and I’ll add some of my own reactions at the bottom….

Knowledge@Wharton: From the beginning, you documented the specification for the PostScript language.

Geschke: Yes. We published the spec about three or four months before the first LaserWriter shipped.

Knowledge@Wharton: These days this is fairly common practice to help establish a standard platform, but it was much less common back then. Was this merely a practical necessity because people needed to write software to drive PostScript printers or was this a strategic move to establish PostScript as a standard?

Geschke: It wasn’t strategic in the sense that we understood this would become a standard way of doing business. It was the only way we could figure out to get the hardware manufacturers and the software developers and the platform vendors to collaborate. You couldn’t do independent deals with each of them because there would always be somebody left out.

If you really wanted to make it a standard — and our goal from the beginning was to have it be a universal standard — you have to publish. You just have no choice. You’re taking the risk that someone will do a better job of implementing it. We had the self confidence that we would always have the best implementation, and that has turned out to be true.

Knowledge@Wharton: A lot of PostScript clones did come along eventually.

Geschke: At one time there were over 75 of them.

Knowledge@Wharton: Did it ever worry you that they would start to chip away at your market share?

Geschke: No. But it turned out that Microsoft acquired one of them.

Knowledge@Wharton: From Bauer Enterprises.

Geschke: Right. TrueImage, they called it. Microsoft did a deal to license TrueImage [to Apple] and Apple would license TrueType to Microsoft. That was pretty scary for us.

Knowledge@Wharton: When did you first hear about the Microsoft-Apple partnership for TrueImage and TrueType?

Geschke: A few days before it was announced….

(More on the TrueImage/TrueType story: [1], [2].)

The “PostScript Red Book” was the language specification, and the “Blue Book” was a language tutorial and cookbook. Adobe decided how PostScript evolved, true — it was corporate governance, not consortium governance (and you could argue that a company listens to its customers better than a consortium does) — but PostScript was free, open, documented, and internally competitive back in the 1980s.

PDF did the same thing in the early 90s — Adobe was open with “how PDFs act”, even though it didn’t give away its recipe for “how to make it act that way”. In the late 90s Macromedia did the same thing with SWF, saying “here’s how a SWF file should act”.

Like HTML, the spec told you how a file should behave, and didn’t need to include sourcecode to a particular application.

If you really wanted to make it a standard — and our goal from the beginning was to have it be a universal standard — you have to publish. You just have no choice. You’re taking the risk that someone will do a better job of implementing it. We had the self confidence that we would always have the best implementation, and that has turned out to be true.

The SWF file format was first published in 1998, and has been updated with every release. In the early days it also included runtime sourcecode and SDKs, and this was the basis for Oliver Debon’s wonderful early work, which was later recycled into SWFDec, Gnash, and a number of other clones (I’d cite links, but The Web suffers linkrot.) HTML, JavaScript, CSS, none of those file formats include rendering engine sourcecode. PostScript and PDF predated these, and Flash followed the same approach later. All focused on published file formats, with the big difference between governance by a consortium, or governance by a company.

That’s history. The important thing for today? This mindset permeates Adobe. This is the corporate memory. This is what drove the evolution of the group, its beliefs, its understandings.

To achieve something new and useful, you have to have allies. You have to be clear in what you’re trying to do, and commit to it with a public specification, a promise of how functionality “should” behave. There’s the risk another group will undercut you, but it’s a risk which must be accepted if you hope to achieve progress.

From what I’ve seen since the Macromedia acquisition, this awareness on open formats to encourage consensus upon innovative technology runs throughout the company even today. And, from what Adobe’s founder tells above, it was there from the start, before The Web, even before The Internet publicly opened up.

Read the interview again. To bring about a new technology ecology, Adobe continually commits to a public specification, drawing in both partners and competitors, and trusts to innovate quickly within that new, growing area.

If you want to understand Adobe, it’s essential to understand this cultural memory of opening up, becoming vulnerable, yet innovating quickly.

Sure, join the dinner party! (but it’s potluck)

Phrases like “the execution of Adobe”, “makes Flash an absolute toy”, “Flash Killer” and the like annoy me. That’s why I pay attention to what happens later, when the press releases become deliverables.

When Silverlight announced the deal, lots of people said Flash and Silverlight would then be comparable. But long after the press release, the eventual website was still Flash, and the common forum advice was likely why we didn’t hear much later about what this campaign did for adoption.

Then deal and were prophesied to close the desktop gap with Flash, rendering the technologies equivalent. I’ve heard quotes like 50 million unique visitors for the first and 30 for the second, but no word on how many of those succeeded in watching video, rates of successful installation and so on. It has been mysteriously quiet.

Until today. Microsoft announced that H.264 video will be “in a future version of Silverlight”, and also let slip some stats on “How did Silverlight do at the Olympics?”

“On the Silverlight-enhanced NBC Olympics site, the average viewing time was over 27 minutes, as opposed to an average of just three minutes on some Flash-powered sites broadcasting Olympics coverage elsewhere. We think this indicates Silverlight provides a more compelling, engaging and rich media experience for viewers.”

This is the press release, the first after the event to put the whole Silverlight/Olympics storyline into final context. If there was good news, this is when we’d hear it trumpeted. If there was neutral or poor news, this is when we’d hear it managed. And the most important statistic mustered? An observation on the average size of YouTube files.

It seems to say a lot, but somehow, I’m not sure it will change the debate…. 😉

Look, it’s great that Microsoft is trying to refactor WMP and CLR to fit in webpages, like Flash. It’s great that Firefox is adding Ogg Theora, and that Google Chrome is trying to build a generic HTML-driven application shell. If Ajax gurus can make HTML apps smoother, then that’s a big help to the world. Java FX moving towards the Flash vision? I’m all for it.

I could do without the negative stereotypes, but having everyone agree on this “experience matters” approach is a good thing.

The more options, the better.

But… you know… not all of these technologies work equally well in the real world. Most websites can’t afford to turn audience away. You hear such objections to Shockwave, even though 60% of computers already run it.

The “software wars” are fun to debate and evangelize online, but all that extra rhetoric makes it harder for people to learn about and actually use these technologies in the world. Confusion costs.

Even newspapers frequently lump Flash and Silverlight together. I guess the brandnames fall into the same basket, for the reporter. But for readers, one has a near-prohibitive cost to use. How much of your audience can you afford to turn away? The reporter isn’t doing the reader any favors by suggesting either technology can be used equally.

Same with video. What types of sites can afford to use Ogg Theora? Commercial video sites won’t change their existing On2 VP6, Windows Media or H.264 production workflows to move to a codec considered outdated in 2001. I can see Wikipedia using On2 VP3, but how many other sites can afford it?

A fast scripting engine, eating platform-neutral files to run fast OS-native code for datagrids and 3D and games? How much of your audience must arrive in that brand of JavaScript runtime before you can afford to use it?

The more options, the better, true. But that doesn’t mean they’re all equal. The namecalling and fake drama can only distract, not hide this.

Flash just works. Back in June, 85% of consumers already supported H.264 on their computers. Forrester says 97% in enterprise have a JIT. It’s all open and free to use, right there on the world’s desktops.

You can actually use it. That’s what distinguishes Flash. How can there be a debate?

This is not an exclusive dinner party, this little RIA-in-the-browser soiree. You’re welcome to come, and it’s flattering that you’d like join in. Glad to have you here! But it’s a potluck dinner, so you really might want to bring something, other than announcements of how next time you’ll bring something, and saying (between mouthfuls) how much better it will taste than this evening’s poor proprietary fare…. 😉

Unpublished comments

I’m reposting some comments here which, for whatever reasons, never made it into the original weblogs. Mike Krisher had difficulty updating Creative Suite; “Hey I’m Ben” had a Linux-aggrieved-by-Adobe rant with many supportive comments published; Chris Nicol had a post on Silverlight economics.

Continue reading…