Jan 9, 2024

Introduction

In a recent Hacker News discussion, commentators criticized Google following its latest layoffs. The prevailing sentiment is one of disappointment; Google, once a beacon of innovation and accessible technology, is now perceived as prioritizing profits over ethics. This is evident in their increased reliance on ads, the introduction of dark patterns, sunsetting of products, and their overall corporate behavior, now seen as one of the least ethical in tech, a far cry from their early mantra of “don’t be evil”.

Personally, I have lost a lot of faith in Google after its closure of Inbox, Stadia and Domains, and have started to migrate away from most of its services. I am also angered by the way it stripped away the livelihoods of so many of its employees with multiple rounds of layoffs.

My feelings towards Google are mirrored towards other technology companies initially driven by visionary founders and groundbreaking ideas, that eventually succumb to the temptations of scale and profits. Such companies tend to mineralize into bureaucratic behemoths, engage in anti-competitive actions, degrade their products, and foster toxic work culture.

There are many apologists for the behavior of tech companies in Silicon Valley. The misuse of user data or the closure of services is hand waved away as “just business”. Layoffs are justified as the “trimming of fat”. Without employees and users though, these companies are nothing.

Nutrition labels

I am hopeful for greater accountability in tech leadership, driven by consumer choices. Just as there has been a growing awareness about the quality and origin of our food (e.g. microplastics, organic farming, GMOs etc.), there’s potential for a similar consciousness to develop around the quality and ethics of our technologies and the companies behind them.

To that end, I propose nutrition labels for technology companies and the products they make. My guiding principle is that these labels are primarily factual and easy to verify.

I am hesitant to introduce subjective judgements or calculated scores to these labels. In other words, I do not want to give companies like Google a grade (whether to give them an A or an F would be too hotly contested), rather I want users to evaluate factual data for themselves.

Sometimes a company is synonymous with its sole product e.g. Reddit. Other times, a company builds many products e.g. Microsoft. I think it’s worth having two labels: one at the company level, and one for each product.

Existing work

Apple has privacy nutrition labels in its App Store. Here is an example for Facebook Messenger:

These are a step in the right direction, but lacking a few characteristics. They do not:

• Link out to external references
• Each label is categorical instead of ordinal, which makes it difficult to assess its extent

There are proposed nutrition labels for AI models from Twilio, which seem more useful and provide inspiration:

However, we can see from the mockup that these labels are a bit vague and almost read like a marketing tactic.

Company labels

My proposed company label would include:

• Venture capital
  • Amount raised to date
  • Series
• Founder profile
  • Previous employment
  • Main accomplishments and failures
• Revenue streams
  • Major products and their % revenue
  • Most recently retired products
• Employees
  • Estimated employee count
  • Number of offices
  • Hiring rate
  • Turnover rate
  • Remote, hybrid or in-person work
  • Glassdoor satisfaction ratings
  • Most recent layoff and count
• Environmental impact
  • Energy usage
  • Carbon footprint

Admittedly, some of these labels are hard to quantify or find information about in a way that is comparable between companies. However, the label does not need to be perfect, at the very least it can link out to salient news coverage or other sources of information that can help a consumer get the big picture.

Product label

My proposed product label would include:

• Data privacy
  • The major pieces of collected data
• Pricing
  • Subscription or one-time costs
  • Additional in-app purchases
• User base
  • Estimated user count
  • Average rating across different platforms’ app stores or review sites
• Vendor lock-in
  • Supported export/import formats
  • Contributions to open source

Conclusion

When nutrition labels are primarily factual, they become easier to crowdsource. I hope to either work on this project myself, or see someone else champion it in the future. I believe that providing clear, digestible information about a company’s products and operations can help consumers vote with their wallets for what they want technology companies to become. I hope that we arrive at a future where stablity is prized over scale, innovation over profit, and ethics over expediency.