There’s No Soul in the New Machine

1 April 2023 19:19 GMT+0100 

They would be the first.

On the bay side of Highway 101, past the six towers of Oracle’s Emerald City, was an office park of twenty new buildings: all two stories, all white, wrapped with long bands of green windows. So similar were the buildings that visitors had to pay attention to the address or look for a company’s sign to ensure they had found the right building.

The offices of one company occupied an entire bottom floor. In one corner was a glass walled conference room; a software engineer and a marketing manager sat side by side at one end of the long rectangular table. Around them were documents, coffee cups, and partially eaten bagels. Both were looking at the marketing manager’s laptop screen.

“Explain to me again what you want to do,” said the engineer. He looked at the screen as he spoke.

The marketing manager paused; the engineer should already know the project: the marketing manager had distributed a new marketing requirements document (MRD) over a week ago. Three weeks before, a high level plan had been approved by the vice-president of marketing, enthusiastically, and the vice-president of engineering, reluctantly. There had been more than enough time for the engineer to read the MRD; asking for further information meant the engineer didn’t like the feature. Software engineers never accepted anything from marketing without first asking many questions or criticizing the plan or outright mocking the idea.

The engineer noticed the marketing manager’s expression. “Well, hold on. I think I understand what you want to do. It’s why you want to do it that I don’t understand. Walk me through it again so I have a better sense of the overall context.” For the software engineer the romance of being in a start-up was now giving way to an unexpected reality. He hoped he had misunderstood what marketing proposed—it was simply too obnoxious to think of doing.

Seven months ago the company had released a free software utility, which memorized a user’s website passwords and filled out order forms: name, password, mailing address, billing address, credit card numbers. The utility simplified logging into websites and streamlined the order process by automatically filling in fields. For the first few months after its release the growth of the user base had been slow, but now there were over three million active users. The program was simple and useful, people loved it so much that several thousand had emailed the company: awesome or thanks so much or really makes things easier and similar. Now the word had come down from the board of directors, three venture capitalists and the chief executive officer, that it was time to start making money.

The marketing manager clicked back to the first slide in his presentation. “Okay, well it’s all in the MRD, but let’s start over. Here it is again. Our program is running, you can see the icon on the bottom right in the tray. The user is surfing the web, clicking, around, then comes to this product page.” The screen shot was of the Blockbuster web site, showing a Stephen King movie to rent.

“Now, the user is on the this page, and we’ve…what’s the term? You hook the browser? Or hooked Internet Explorer?”

The software engineer nodded.

“Right, anyway, whatever you call it, we know where the user is. We know the web page. So, next, we do a database look up to see if we have an appropriate ad to show, and if so, voila.” The marketing manager clicked to the next slide.

The software engineer frowned. On the marketing manager’s screen was a slide that showed almost the same screen as before, but now in the lower right corner was a small advertisement for a company named Big Star, which was an advertisement for buying a DVD not from Blockbuster, but instead from this new online video company.

The software engineer made a few notes, then looked at the list of use cases in the back of the MRD. “What happens if the user doesn’t want to see the ad?”

“There’s an x-close box in the upper right corner. Isn’t that a Windows interface convention? Clicking that will close the ad. Or, if they go to another web page, the ad goes away.”

“And if a user clicks on the ad?”

“You launch a new instance of the browser, and whatever URL is encoded in the ad, that’s the web page the user sees. In this case the new page would be at the Big Star site and would show a coupon for ten dollars off the next DVD rental. Oh, and by the way,” the marketing manager paused for effect, in a slow, deadpan voice said, “Users seeing ads and clicking on them is how we will make money. Money.”

The software engineer smiled and nodded. He liked the marketing manager, even if he was a bit different. Gold rushes attracted all types, whether it was for mining for real gold in the foothills of the Sierra Nevada 150 years ago or digital wares in Silicon Valley at the turn of the millennium.

In the past, marketing managers had been former engineers, tired of coding, with a sense of business, maybe had good people skills. Or they might have been engineering students in college, but after graduation had decided not to become developers. In either case, the marketing managers were familiar with the software development process, knew how to write good requirements documents, and were reasonable in negotiating the trade-offs regarding features going into a product weighed against engineering time needed to code and test.

Now most were MBAs, and came from large companies which made consumer products: dog food, baby shampoo, canned pork n’ beans, or razor blades. They were unfamiliar with the software development process. Still, they did learn quickly, and were obsessed with a positive user experience.

“How do you think users will react to these ads?” said the software engineer. “I’m worried because we have a lot of users who like our utility. But I’m afraid they will uninstall it because of these ads.”

The marketing manager bit his lip. What was this guy’s problem? The landscape for many technology companies had changed ever since Netscape started giving away its browser. Engineering was stuck in a 1989 revenue model: selling shrink wrapped software at a retail outlet. These days, unless the company made very specialized software, like Adobe’s Illustrator or FrameMaker, or unless the company provided expensive consulting services like Oracle, then there were few options available to web site properties or mainstream software publishers. The company neither sold software nor provided consulting services. It only had a very cool utility that might soon be copied by other companies, unless legal finally filed all the patents.

None of this was bad news to the marketing manager. Once he understood the consumer information available through their technology, he quickly realized the potential. Before moving to Silicon Valley eight months ago, he had worked as a marketing director for two clothing companies: J. Peterman and J. Crew. These companies had spent hundreds of thousands of dollars on various marketing campaigns in different mediums: radio, magazines, television, and direct mail catalogs. Yet they were always shooting blind: at best the company might have some demographic information about who watched a particular television show, or the median income of a particular zip code. The company would buy mailing lists from retailers targeting a similar demographic, maybe from Pottery Barn or Whole Foods. For all this, marketing departments rarely knew if someone bought a product because of one of their advertising campaigns.

Not any more—the quality and quantity of information available online was nothing short of mind boggling.

The marketing manager exhaled slowly then said, “Well, how do you react to ads on television? How do you react to ads in newspapers and magazines?”

“I don’t know. Do I react to them? I don’t think of it that way. They’re just there.”

“Exactly. You’re just not used to seeing ads this way. But look, newspapers and magazines make their money by advertising, not by subscription fees. Same thing with television—television programs have advertising.”

“Yeah, but—.”

“And look, you have to admit that at least this ad is relevant. That doesn’t happen with newspapers or television, I mean not really. At best in a newspaper’s sport sections there’s an ad for tires, not women’s clothes. This is much better. We’re going to be very smart about what ad we show and when. We’re not showing an ad for propane products to someone who is shopping for lingerie at the Victoria’s Secret web site. Maybe the user will click on the ad, maybe not, but there’s nothing wrong with offering a contextual ad.”

“Of course, we could also—.”

“And these ads are much better. Think about those stupid flashy, jiggy ads you see embedded in web sites. My wife says she thinks she’s going to have a seizure when she sees them. They say something like you’re the millionth customer and just won a new blow up sheep doll or something stupid. We’re much better than that. Advertising can be done well. I won’t say it’s art, but a good ad can be useful to a customer and maybe even pleasing. Half the time people watch the Super Bowl, it’s not to watch the football, but to see the ads. And they’re sometimes good, right?”

This time the software engineer didn’t try to speak, only sat and waited.

“Also, think about those catalogs you get in the mail. All that wasted paper. You probably just throw most of the out, right? I mean, what a waste of trees. Catalogs and direct mail are just wasted paper. With our ability to precisely understand what really interests someone, we can target and present ads, useful ads, that frankly, I think will surprise and delight users.”

The marketing manager seemed done. Put that way the software engineer couldn’t disagree, although he wasn’t sure about delight, but end users certainly might be surprised to see an ad appear.

Yet it remained distasteful, the manner of presentation. In this scenario their software utility was doing something that maybe the user wanted, but maybe didn’t. By experience and preference, he felt software should only function in a manner consistent with the user’s explicit intention: a mouse click or menu pick or series of keystrokes. An advertisement, like the Big Star banner, appearing in a separate window independent of the user’s action, seemed wrong. It was like a word processing program suddenly starting to print a document even if the user didn’t initiate a print job.

More than that, it felt demeaning to be coding in the service of advertising. It felt like a step down, a waste of his ability and his talent. His first job had been as a staff software engineer for NASA’s mission control center in Houston. Then he had moved to Mountain View and worked for Sun, then SoftWare Publishing, then a small start up that went out of business after three years. As his career progressed so had his engineering ability, such that he could cleanly program complex features that sat behind a simple, easily understood interface, useful software that people were happy to buy.

Almost eleven months ago he and a few other engineers had started developing the company’s form fill utility, before there was any talk about advertising. The first time he showed the utility to his father, a retired civil engineer from Bechtel, a reluctant computer user, he had immediately grasped the usefulness. “Oh yeah, I get that. That’s good. Especially the password fill in thing. I need that. Well done.”

He wondered what would his father say about the advertising.

A friend, a developer at another company, wasn’t having any of the soul searching. “What’s wrong with making a bunch of money from advertising technology? Someone’s got to do it. It’s simply where the VC money is right now. If it was 1990 you’d be making shrink wrap software. If it was two, three years ago you’d be building a web site to sell shit, or tools to build web sites to sell shit. Now it’s adtech. Who knows what the next hot thing will be in five years, but you can be sure a lot of smart developers will be attracted to it. Take the money you make and use that to save the world.”

“Yeah, but -”

“Besides, from what you’ve told me the advertising could actually be decent, right? It also subsidizes your cool form fill tool, which people seems to really like. And, a lot of the backend technology sounds pretty cool.”

The technology had been part of the attraction, not just all those stock options. The software engineer had been intrigued by the challenges: the client code, server code, how to make it all seamless and efficient, then scaling out server farms at co-location facilities around the country as their customer base grew. There was also a completely new domain, at least for him, one about which the marketing managers were obsessed: end user data, specifically data to help them show the right ad at the right time. The hardware and software requirements for all that data processing and analysis and reporting—it was something he couldn’t wait to work on.

He had done some reading about advertising after the board of directors announced the company’s revenue plan. In a book his wife had from one of her college marketing classes, he learned about concepts of primary and secondary demand, advertising as a defensive mechanism, as a means to keep existing customers rather than attract new ones. Later he asked her, “But doesn’t advertising divert consumers from buying a better product at a cheaper price? I mean, if the persuasion is strong enough, then consumers are not making rational decisions. And wouldn’t that affect price competition?”

His wife, an economics and math major, whom he always thought should have been an engineer, had sighed, then said, “ Okay, Thorstein, I see you’re at the chapter about the classical theory of the market. Yeah, if there’s prefect competition, which pre-supposes similar products, many customers, many vendors, and a buyer knows everything about a product, including all the prices, then advertising might be a distraction. But that prefect environment doesn’t exist except in books, lectures halls, Ayn Rand horror novels, and the editorial pages of The Wall Street Journal.”

Looking again at the marketing manager’s computer, he now asked, “How do we think the web site owners will react? You know, like in this case Blockbuster. What if they find out we’re running an ad on their site?” said the software engineer.

“Well as I understand it we are not running an ad on their site. Right? But regardless, they probably won’t like it. But we’re not exactly sure, since no one had done this before. Still, as you know better than I do, we’re not doing anything to their web site. It’s our application that is putting up the ad, the user has installed our application on his computer, so unless the web site owner think they have a say on what software a user runs on his computer, they don’t have a case. And besides, if the web site has a better proposition for the user, then the user will ignore our coupon. It’s competition, pure and simple.”

The software engineer looked out the window. The tide was out and there was only a ribbon of water surrounded by mud in the Belmont slough. He’d have to walk his dog after the meeting.

Was it pure and simple? The software engineer was one of the first ten employees, and he had hundreds of thousands of stock options, more than all previous jobs, combined. An initial public offering or acquisition could be a life changing event. A chance like this would not come his way again. He had worked at another start up for three years, at a reduced salary even, and nothing had come of that. This time there was a solid product and, however distasteful he found it, a revenue plan that was believable. There had been some favorable press about their form fill utility, and the company had even made one of the many lists of companies to watch next year. The CEO was in the process of locking up another round of funding, and soon a new set of venture capitalists would be replacing two of the three board members.

The software engineer looked at the marketing manager’s computer: a laptop from Dell, nothing special, yet it had a hundred times the capability of his first computer. In his twelve years in the work force the computer had changed considerably: decreasing component costs and increasing processing power, along with faster and faster high speed internet access had transformed the computer from a tool reserved for scientists and engineers to something close to a household appliance. Everyone had one now.

He looked at the screen: in addition to running Power Point the marketing manager was running a browser, a recent version of Internet Explorer. Until recently Netscape’s Navigator had been dominant, but Microsoft had come from behind to improve Internet Explorer, and Navigator was losing ground in both features and market share. And yet a browser was nothing special, just another application that opened and ran files. Yet put altogether, the computer and the browser and the internet, it seemed like nothing more than a glorified television set with an infinite number of channels: shopping, porn, videos of babies dancing. Now advertising would further pollute the waters.

Or would it? He had recently finished Berners-Lee’s memoir about creating the web, a vision that included academics, surfing, commerce. Wasn’t advertising a key part of commerce? The software engineer made a few notes in his notebook, then said, “Okay, let me see what I can do.”

 

Notes

The title of this essay is a variation on the book title, The Soul of a New Machine, by Tracy Kidder.

Thorstein Veblen was an economist, best know for his book, The Theory of the Leisure Class.

Ayn Rand is the author of the libertarian dystopian fantasies The Fountainhead and Atlas Shrugged.

Tim Berners-Lee is a British computer scientist, credited with the creation of the world wide web.

Leave a Reply


Your email address will not be published. Required fields are marked

{"email":"Email address invalid","url":"Website address invalid","required":"Required field missing"}