📚 Sweating Bullets: Notes about Inventing PowerPoint

Finished reading Sweating Bullets: Notes about Inventing PowerPoint by Robert Gaskins. I’m a big fan of dissections of all things pop, and although the topic of this book isn’t quite the same as a music documentary or a look back at 1980s UK TV programmes, the backstory of PowerPoint is in the same wheelhouse for me. It’s played a big role in my life and feels like a part of mainstream culture. The book scratched exactly the same itch.

The author was hired by Forethought, Inc. in 1984 as Vice President of Product Development, a year after the company was founded. The company had investors but little direction, and needed a reboot. Gaskins brought into the company his idea of developing a tool for creating both overhead transparencies and 35mm slides for the nascent GUI-based desktop computer platforms, Macintosh, Windows and OS/2, some of which hadn’t been released yet. His view was that commodity personal computers would get better over time and that these GUI-based platforms would be the future. Throughout the book, it becomes clear that this belief and his strong views on what platforms to develop for (and not develop for) and in what order were core components of the product’s success. PowerPoint for Windows only started to outsell its Macintosh counterpart in 1991; Windows took so long to gain traction that early versions of Excel and Word were shipping with Windows runtimes included as there was no expectation that the consumer would have Windows already installed on their computer.

Gaskins’ father knew many people in the audio-visual industry who were able to give him insight into how many overheads and 35mm slides were created every year. This data wasn’t widely known or available to people in the technology industry, but it helped to make the business case for the software product which eventually became PowerPoint.

In his previous roles, he travelled around the world and received pitch decks from many vendors. He collected the handout versions of the slides, building a library that served as input into the functionality that the software required:

They were made in all kinds of ways—handwritten, on a typewriter, drawn on an old pen plotter, or made on many kinds of large computers. But I was struck by how they were very much the same all over the world. The basic style was essentially uniform everywhere, indicating that there was an international style for overheads. This meant that a single application to make presentation visuals would be saleable worldwide.

Overheads and 35mm slides were originally two very different paradigms. Overheads were usually created by the person presenting them, sometimes in the meeting itself. (I sat in many classrooms and lecture theatres where the person leading the room wrote directly on a scroll of transparent film as they went.) It was a lights-on approach that invited discussion and narrative. In contrast, 35mm slides were usually professionally created, typically by a company such as Genigraphics, and would be presented in a darkened room. I started work in 1999 and never saw an overhead projector in a corporate setting. I also never saw a 35mm slide presentation; the transition had already happened. Through PowerPoint, both of these formats collapsed into the same thing and ultimately became something new — ‘slides’ delivered through video screens. Arguably this has directly led to some of the ‘Death by PowerPoint’ narrative, where some people deliver their presentations with slide after slide filled with tons of information.

The application, initially developed for Macintosh to create black-and-white slides that could be printed and transferred to transparencies for overhead projection, was originally called ‘Presenter’. The name ‘PowerPoint’ came to Gaskins while he was in the shower one morning, the place where many good ideas are born, about three months before the product was released to the public.

Although PowerPoint and applications like it, such as Google Slides and Apple’s Keynote, are ubiquitous now, during the late 1980s the idea wasn’t obvious to everyone. Gaskins had to “pitch the idea repeatedly, hundreds of times, without getting much response, and to keep refining the idea in the face of prolonged skepticism.” He even had to lobby Apple to let the application run in full-screen mode, as it was against their own design guidelines at the time.

One of the most fascinating insights was how you don’t (or, perhaps more correctly, he didn’t) need to closely guard your plans. He went from trying to keep it a secret to telling everyone, and didn’t need to worry about someone else developing the idea too. I wonder if this is still true today, given how quickly information travels in the Internet age.

In the early months of working on PowerPoint, I felt very protective about the idea; it seemed so obviously great that anyone who heard of it would copy it immediately. Over the three years of pitching the idea hundreds of times, I more or less came to the opposite conclusion: that no one would want to copy it, and I would talk about it to almost anyone for a business reason.

In those days, I first heard a line attributed to Howard Aiken (designer of IBM’s Harvard Mark I computer during World War II):

“Don’t worry about people stealing your ideas. If your ideas are any good, you’ll have to ram them down people’s throats.”

[…]

The ideas that everyone recognizes immediately as worth stealing are going to be just barely in front of the current commercial frontier and can’t be successfully stolen quickly enough, while an idea just a little further away in time will seem to have objections that make it appear a bad idea to most people.

On the journey to releasing PowerPoint 1.0, Forethought went through very tough times. For three years in a row, the management teams only allowed staff a week off over Christmas once they had calculated that they had enough cash to get through the period as well as to pay minimal severances if they had to liquidate in the New Year.

Gaskins looks back to a time when the final ‘golden master’ build of the software was critical. There were no automatic updates via the Internet, so it had to be as complete and bug-free as possible in order to avoid costly recalls and reissuing of the application on a new set of disks, something that could put a small company out of business. The whole concept of creating a ‘minimum viable product’ and getting software into the hands of users as quickly as possible in order to get feedback just wasn’t a thing. It’s the total opposite of the evergreen nature of application software today. As I sit and type this, my iPhone has three apps that are ready to install an update and 15 that have already updated today. If any of these updates introduces a bug, it can simply be patched with another update, probably without me even knowing about it. The process of creating the ‘golden master’ sounds a lot like pressing a vinyl record:

When the creation of precisely one perfect disk was verified, this became the “Golden Master.” Quality Assurance would hand-deliver this master to manufacturing. At the manufacturing plant, the single master received would be used only on specialized duplicating machines to make further masters, called “silvers”; the Golden Master was then locked up in a secure cabinet. The silver masters were used on duplicating machines to manufacture product disks, which were verified against silvers. If more silvers were needed, they were made from the unique Golden Master. Golden Masters were never de-accessioned, but remained permanently in secure storage, so later questions about what was on them could always be answered with assurance. The whole Golden Master process helped to prevent errors, and also made it easy to assign responsibility for any mistakes discovered in products.

But they added some ceremonial steps, as quoted from Dennis Austin, the original architect and developer of the software:

“We first recognized that two people were required to do the work. One would follow the steps on paper and the other carry out the instructions. The two would monitor one another’s actions to be sure no one slipped up. (The two people were generally Bob and I.) We donned special robes to signify that we were engaged in a solemn rite. (Our robes were actually some old purple T-shirts from Bell-Northern Research, but it’s the idea that’s important.) Step by step we soberly followed the instructions to guarantee the perfection of our master.”

Gaskins adds:

The precautions we took were extreme, but so were the risks. In fact, we never made a mistake in preparing a Golden Master. In later years, the same ceremonial rituals (including wearing special costumes) were taught to everyone in the GBU [Graphics Business Unit] who made masters, eventually usually a pair consisting of a Program Manager and a Quality Assurance Engineer, and our perfect record continued.

There are some lovely technical details about how the software worked. The computers of the mid-1980s didn’t yet have true multitasking capability. The ‘slide sorter’ feature, which allowed you to interact with small thumbnail representations of the slides in the deck, created them in the background when PowerPoint was idle, with the routine constantly checking for user input. If it detected any, the process would be abandoned.

In the days before version control systems such as Git, working with multiple developers on a project was non-trivial. After the company hired Tom Rudkin to join Dennis Austin in development, the team of two bought a wall rack that was typically used to hold punch cards for employees clocking in and out. Each card represented a file in the code base; if you wanted to modify it, you would take the card from the rack to your office, and then add details of your changes to the card when you were ready to upload the updated file to the server.

Gaskins covers the hand-porting of clip art from the Genigraphics library into files compatible with Windows. They thought that this would be a useful income stream for Genigraphics; it cost Microsoft USD 250,000, but took place at a time when everyone started producing clip art and end-user prices were racing to the bottom. The main legacy was the preservation of this artwork that might otherwise have been lost to obscurity as the use of PowerPoint and the rise of self-produced video slide projections eventually forced Genigraphics out of business.

Forethought represented a first for both Apple and Microsoft — the first company that Apple’s Strategic Investment Group took a stake in, and the first company that Microsoft bought outright, for USD 14m in 1987, a number that feels like a rounding error today. At the time of the Microsoft sale, they were simultaneously considering an IPO of the company, but given the stock market crash of Black Monday and subsequent poor environment in which to go public, they made the right choice.

The deal to join Microsoft didn’t include the need for Forethought staff to relocate to Redmond, Washington, where Microsoft’s headquarters are based. Instead, the team were allowed to stay in Silicon Valley. The geographic distance probably also played a part in them keeping a logical and procedural distance. They were allowed to operate somewhat autonomously, including not being forced to use the internal Microsoft software development tools. I love the insight about how Gaskins assumed responsibility for things:

Early in September, we had a couple of very senior Human Relations people come down from Redmond to explain recruiting to us and answer any other HR questions in an all-hands meeting. Someone from the GBU asked about signing authorities, and one of the Microsoft HR people explained to our group that “BillG is the ultimate authority for everything in Redmond, and in the same way, BobGa [me—Microsoft people often talked in email aliases] is the ultimate authority for everything in Silicon Valley.” When I heard that, I realized that I should take exactly that attitude, about everything, until I was challenged on something. I already understood that how much authority you have in any business largely depends on how much authority you assume you have. I had started out by being a bit tentative, while finding out how Microsoft worked, but no one in Redmond seemed to be trying to tell me what to do—at all. I should start immediately to assume all the authority I could, however unreasonable it might seem.

He also has some nuggets of wisdom about hiring. Bill Gates encouraged him to never hire a less-than-great person, no matter how desperate the team is to fill the role — this is something I’ve tried to practise myself for many years. Interestingly, Gates also said to never fail to hire a great person, whether there is budgeted headcount or not; this one I would struggle much more with in the companies I have worked for. Gaskins also notes that he felt strongly that he should never hire anyone into a role that had no significant internal career path; these jobs should be outsourced to companies that did have career paths for people in these roles.

The book covers how Microsoft changed between the Forethought acquisition in 1987 and Gaskins’ departure in 1993. During this time, the company increased in size from 1,200 to 12,000 staff. Once it was released for Windows, PowerPoint was soon bundled with Excel and Word to create the first version of Microsoft Office. However, it was credited with only 12% of the revenue from the bundle, with the two others receiving 44% each. Microsoft said this split was based on historical sales, but Gaskins believes it was due to the larger sizes of their teams and the lobbying by their team leaders, made easier through their proximity to decision-makers in Redmond.

Gaskins writes wistfully about how meetings and presentations have changed during his lifetime. Giving a pitch in the mid-1980s involved creating slides, but they would typically be accompanied by a written document and spreadsheets that gave the detail. By the end of the first decade of the 2000s, people like Marc Andreessen and investment firms such as Sequoia Capital were asking for no more than 20 slides for a pitch, without any accompanying material.

There is some repetition in the book, but I didn’t get bored, particularly where the author covers the differences between types of presentation at the time that PowerPoint was released.

Gaskins left Microsoft in 1993 after PowerPoint 3.0 was released. He considered the product to be “finished”. I know what he means, but as a long-time PowerPoint user, I can definitely say that some small improvements since then have made a massive difference. For example, Smart Guides that appear when dragging objects, allowing you to quickly align the components of a slide, eliminated many hours of nudging and distributing objects into place. The innovative move to Office Open XML formats such as .pptx and .pptm means that I can’t remember the last time one of my PowerPoint files got into an unrecoverable pickle.

If computing culture is your thing, this is an excellent read. I really enjoyed it.