The Show So Far

dratner

by Leave a Comment

Famous Firsts

This summer I decided I wanted to write some code again, but I also wanted to write a casual multi-player party game that I could play with friends over the Internet. What I came up with is Famous Firsts.

Famous Firsts is a literary quiz game. In each round, your group of 3-5 players will be presented with the summary of a real book and you’ll be challenged to write the first or last line. Then all the players are presented with the list of submissions and challenged to pick the real one. You get a point for guessing correctly and another point for each person you fool. Trash talk is conveniently enabled via a side-bar chat function.

The game is nothing fancy, but it was a fun little project and I’ve open sourced it. I hand coded every line of JavaScript, CSS, and Go code, using only one nonstandard library (xid) and no frameworks, external data storage, or cruft of any kind. It’s 100% hand-made, locovore and organic code. Hope you enjoy!

by Leave a Comment

How to Reform Section 230

I’ve spent more than two decades as an engineer and executive building digital platforms for everything from helping people find childcare (Sittercity.com) to supporting President Obama’s 2012 election campaign. Many of the products I’ve worked on were made possible by the protections built into Section 230 of the CDA. While I’ve been a strong supporter of the law in the past, the time has come to reform it. The big platforms (including Facebook and Twitter) have shown that self regulation isn’t a possible solution unless it is spurred on by real legislative change. While different sides have different views of the problem, the fact is that these companies have developed not just business monopolies, but effective monopolies on the flow of information in society. That creates a public interest that an unmodified Section 230 no longer serves.

A priori, since the objections to repeal or revision of Section 230 tend to be rooted in free speech rights, it’s worth pointing out that these rights, while broad and sacred, are not absolute. Perjury, fraud, libel, assault, certain kinds of pornography, and hate speech are all common exceptions to free speech. At the moment, due in significant part to Section 230, all of these regulated types of speech thrive online.

The platforms have a grain of truth in their claims that they should not become the arbiters of truth, or, implicitly, of these laws. But rather than effectively vacate them, a better option would be to shift them back to where they belong: in public and in the courts. 

How could this be done?

First, require the platforms to better verify the people who use their platforms even if they don’t make that information generally available to the public. If liability is to be shifted from the platform to the person responsible for the content, it must be straightforward to identify to the level of legal certainty who the person is. While some will point out that anonymity is important for users in authoritarian regimes, it may well be that the platforms need different operating plans in different countries (just as they currently have in China) and that anonymity may well be causing more harm that it is avoiding. Advertisers and other sites that support social sign-in should support these changes as well, since they would also reduce bots and other fake traffic. This would be a trimmed down version of KYC in financial services.

Second, the platforms should be required to create a machine-readable near realtime public archive of all posts that exceed a certain threshold of views. If people truly desire privacy, they can be given the option to limit the reach of their posts to below that threshold (likely based on Dunbar’s number – the quantity of real social connections a person can maintain – which is around 150.) If a person doesn’t accept that limit, it is because they desire their post to reach not just their friends but the largest possible audience – effectively, they want to be publishers. Given this, it is reasonable that they should agree that their post will be part of the archive and that they will take public responsibility for it. 

This would provide transparency and allow researchers to prove or put to rest claims of algorithmic bias (a key Republican concern.) It would also have the incidental benefit of chipping away at the monopoly status of these companies since these archives would provide troves of AI training data that potential competitors need to create comparable products.

Third, since this will no doubt lead to an explosion of claims, the platforms should be subject to a tax to offset the costs to the judicial system. This would be a straightforward tax on earnings so those who benefit most would pay the most. These companies should embrace such a tax since the alternative would be to vastly expand their content moderation teams which have already shown they are unequal to the job.

This proposal is relatively modest compared to many others under consideration and attempts to balance the interests of individual people, the platforms, and society. It may not be enough, but it would be an important first step. 

by Leave a Comment

You Can Have Targeted Ads AND Data Privacy

The Internet is having an autoimmune reaction. The basic business plan that supports much of it (targeted ad revenue) has become anathema to users and has led to a progressive arms race of ad blockers, privacy laws, and user outrage while the industry itself has doubled down on the model, finding ways to accumulate yet more user data for yet more targeted advertising.

What may be most surprising about this is that it is happening in an industry that is known for questioning assumptions, innovation, and generally giving users what they want. Yet the basic orthodoxy of the current model has remained largely unquestioned since the 1990s even though it is only now finding its apotheosis in the form of Facebook. It’s worth unpacking some of it to see how we got here.

The Internet Must Be Free

In his great article, The Internet’s Original Sin, Ethan Zuckerman traced the history of the Internet business model and previewed a lot of the problems we now face. In short, to defeat and deter competition, build network effects (and, more recently, outmaneuver regulators), internet companies must grow exponentially. And in order to grow exponentially, they must remove all friction, including that most fundamental and pernicious friction: asking users to pay for their services. But they must still generate revenue and so they turn to sponsorship and advertising. Since very little blanket digital advertising actually works, they must promise more and more targeted advertising and this, in turn, is what drives the pressure to collect more and more user data.

In some ways this model has succeeded brilliantly. We have free services for so many things that used to cost money. News, phone calls, music, and so much more. At one level, you’d think we must be better off. But clearly that is not true since for the most part both advertisers and end users are profoundly dissatisfied. Only the big internet companies themselves have truly benefited.

The Ultimately Targeted Ad

Zuckerman refers to the myth of the ultimately targeted ad and it is a seemingly compelling one: if only I could properly define the right universe for my message I can be successful. The parameters may be demographic, psychographic, or behavioral and can be defined as narrowly as I’d like. Surely this should allow me to overcome John Wanamaker’s perennial challenge: “Half the money I spend on advertising is wasted; the trouble is I don’t know which half.”

But is that really what’s happening? Are the benefits of the each progressive layer of targeting really that clear or are Facebook et als really just successful because of the breadth of their audience? People expect everything in their internet experience to benefit them in some way and, especially as more and more of us experience the internet on the small screen of a mobile device rather than the big screen of a desktop, each and every pixel needs to be justified.

Give The Customer What They Want

As a simple contrast to the behavioral targeting model, consider Google’s now venerable search business. Instead of targeting “24-38 year old women interested in fitness and Christianity living near St Louis” (or whatever), you could simply offer to sell your product to someone who is searching for something very much like it. This contextual advertising doesn’t necessarily require much if any user data but can be much more effective that even the most targeted ad since it is in fact giving users what they want.

The problem with paid search is that while it works for users, it isn’t nearly as good for advertisers. This is just because of supply and demand: search volume isn’t growing as quickly as social media volume so there’s only so much inventory to buy while demand from advertisers has continued to grow. Since Google works as an auction, that means the price to buy an ad on your preferred terms goes up to the point of efficiency – that is to say when the gross margin for the product being sold net of advertising approaches zero. In this case the only one making money is Google.

Get The Best of Both

But why isn’t is possible to give users what they want (a largely ad free internet that only includes calls to action that are relevant to what a user actually wants to buy or do) while still creating volume for advertisers? Since what a user wants to accomplish can almost always be inferred by immediate context (e.g. the search term I entered, the content I am viewing), this can be done without needing to store users’ personal data.

The key to doing this is in a closer integration not between user data and ads but between content/context and ads. A personalized experience can be derived from the content, but it is not going to be as simple as searching content for keywords. A semantic understanding of context is necessary with calls to action that match. That requires intelligence, but machine learning has gotten to the point where such as thing is possible.

A Case In Point

We live these ideas at Public Good. Our technology uses machine learning to derive context from a news story or other piece of digital content. It then uses that context to create a personalized call to action. For example, if you are reading a news story about homelessness in San Francisco, you would be served a unit allowing you to volunteer with or donate to a Bay Area homeless shelter. But we also consider a deeper understanding of the context: if the article you were reading related to a homeless person with a severe substance problem you might see calls to action about substance abuse or if they had been attacked it might be about how you can help reduce violence.

We can create enough scale to get meaningful change by analyzing all the content produced by many of the nation’s biggest news publishers in real time as they are produced, a task that requires substantial automation through AI.

But ours is just a narrow use case of the broader possibility. The same thing could (as is starting to be) done for a variety of other contexts and products in places like Pinterest and Instagram and, for the most part, users really like it. As more platforms adopt this “deep contextualization”, the volume can start to grow sufficiently to make the pricing work for advertisers.

The Cure

The internet doesn’t need to be at war with itself. Users don’t need to sacrifice their personal data to enjoy ad supported services. And advertisers can get unstuck from the myth of the perfectly targeted ad in favor of the perfectly placed one.

by Leave a Comment

Devil’s Dictionary of Tech

Every so often I see an homage to Ambrose Bierce’s original Devil’s Dictionary updated for the digital age. Since a dictionary can have many contributors, I’d like to offer some definitions of my own…

Privacy Policy: A legal agreement obfuscating how a company can take your data and keep it private from you.

Pinwheel: Trying to make colossal failure looking deliberate by making it witty or pretty. See also beachball, fail whale, oh snap.

App: software that gets between you and what you want to accomplish.

To Google Someone: See “peeping Tom” (arc).

Facebook Live: Indecent exposure.

User Generated Content: The wretched refuse of a mass of men leading lives of quiet desperation. Also ridden with plagiarism and misappropriation.

Anonymous: Unaccountable.

Crypto: The intersection of money laundering and mass speculation.

Scalable: A crime too big to be punished.

Disruptive: Breaking laws so quickly, brazenly, and massively that you get away with it for awhile.

Direct Message: A message meant only for you, the recipient, the platform, the NSA, and any interested advertisers.

Cookie: The bribe offered by would be abductors trying to get you into their van.

Internet Poll: Popularity contest.

Algorithmic Matching: popularity contest.

Internet Famous: Incredibly well known by 15 people for 15 seconds.

Infinite Scroll: what you get when you dip below the bottom of the barrel and just keep going down.

Ad Blocker: Free rider.

Peer-to-Peer: A technologically facilitating interaction between two people that somehow requires thousands of employees and billions of dollars in California.

Kambucha: A fermented beverage lacking alcohol but still useful for hazing or indoctrination. It smells like a gas leak, tastes like a swamp, and flows down the gullet with the consistency of sewage. A serving costs more that US minimum wage and it is de rigeur to offer free in all tech company cafeterias.

Uptime: The casino’s hours.

Two Screening: When you need a distraction from your distraction.

Texting: When you don’t care enough to think.

User: A short circuit between the screen and the chair.

Hacker: The resale shop for your sensitive data after vendors and governments have finished with it.

<noun> Hacking: something that used to have its own gerund but is rebranded so it can be claimed by Silicon Valley. See dieting/body hacking, subleasing/rent hacking, etc.

Silicon Valley: To the twenty first century what Detroit was to the twentieth. 

Experienced: An employee who might recognize we have no idea what we’re doing and who will look weird in a company hoodie.

Senior: Expensive.

Machine Learning: when you can’t find a use for all the data you stole so you need to ask a computer to help.

Cloud: Nebulous and squishy agglomerations of vapor, full of darkness, turbulence, and occasional bursts of fatal lightning. Also where your data is stored.

Smartphone: A device to which the wearer’s intelligence has been outsourced.

Microsoft: See Apple (10 years ago.)

Apple: See Microsoft (10 years ago.)

Evil Empire: Whatever big tech you are talking to at the moment.

Don’t Be Evil: Do as I say, not as I do.

Ecosystem: Walled garden.

Sharing Economy: Neither shared nor an economy. Simply a class of companies that rents your possessions and labor in return for a pittance.

Content: The inconvenient text or video in the middle of all the beautiful ads. Like a lure for a fish, you don’t even get to keep it.

by Leave a Comment

Social Impact: A Tale of Two Stocks

In 2017 Kraft Heinz tried to acquire rival global food company Unilever for $143 billion. The deal did not complete and at least part of the reason that then-Unilever CEO Paul Polman revealed later was that it would require too many compromises in Unilever’s commitments to positive social impact.

I’m often challenged by people who believe that social impact is a nice-to-have and that big companies really can’t afford to worry about it for fear of being punished by shareholders. But while it’s anecdotal rather that a full-on trend, the story of these two stocks (UL and KHC for those of you listening at home) tells a very different story. I find it interesting since these are two huge, established, highly diversified and very sophisticated companies (each can claim to reach a broader user base that even Facebook), not a couple of hot new startups hoping to disrupt a market.

Kraft has been following the course that would be suggested by pure, traditional capitalism. Its investors introduced an approach called zero-based budgeting meant to squeeze every unnecessary penny in costs out of the system. At its core, the system calls for assuming no costs and having to justify each and every expense in each and every budgeting cycle. Theoretically, this should maximize earnings in a low-growth, somewhat commoditized industry. (It’s hard to increase the overall demand for products like ketchup, so you either need to move market share or lower costs in order to grow earnings.)

Unilever, on the other hand, has embraced social impact as a key part of its corporate strategy, both at a corporate level through its commitment to sustainable packaging and through individual brand projects like Vaseline’s Healing Project. This approach is supported by research done by leading marketing firms like Edelman and Porter Novelli that in 2018 US consumers said they valued buying a brand that agrees with their social values behind only product quality and ahead of price. This is a 20-year rising trend and is particularly noticeable in highly competitive product categories with the differentiation between two items on the shelf can be relatively small.

While this has not been the only difference in the two companies’ approaches, it has been a very significant one. And over the period since the merger fell apart, the results are stark. Between early 2017 and today, Unilever rose from around 41 to 57 while Kraft Heinz went from around 89 to 32.

It seems social impact has a strong place in capitalism after all.

by Leave a Comment

Lie, Damn Lies, and Economic Statistics

The title of this article is a minor adjustment of a favorite saw of Mark Twain’s, but I can’t imagine he’d grudge me the use because the intent is the same as the original. Every times I read business or economic news I become more frustrated by ubiquitous (and to some degree iniquitous) use of three economic indicators. Here’s what they are, why they are wrong, and simple ways to fix them:

GDP

GDP is meant to measure the overall size of the economy and, for some reason, has now become a proxy for the overall health of the economy. If GDP is growing, things are good, right? Well, no. GDP can go up while the business sector is shrinking so long as the gap is made up for in government spending (even deficit spending.) After all, it’s just defined by this simple formula:

GDP = C + I + G + (Ex – Im)

Where C is consumer spending, I is business investment, G is government spending, and (Ex – Im) is net exports. But could it really go up while business is in decline? Sure. GDP growth in the US over the years 2012-2015 has averaged around 3.57%. Over roughly the same period, government deficit as a percentage of GDP has been 4.5% and the trade balance has remained negative. Which means that if you subtracted out the “unsustainable” or deficit part of government spending, you’d end up with a net negative GDP.

But there’s an even bigger problem. GDP can go up even while wealth and incomes consolidate to a smaller and smaller number of people. So if it’s being used to inform policy about when stimulus is needed or how taxes should raised, lowered, or structured it gives a complete false impression.

Still, GDP is a single, easily accessible number so you can see why it would be useful for sound bites and news articles, right? Maybe, but not so much since there are better numbers that do something pretty similar. Even just using median income (not average income) would be much better if you want to know how average families are actually feeling.

Unemployment

This is another frustrating one. The unemployment rate is supposed to show how easy it is for people to get jobs. Since there’s always some natural and seasonal churn in the economy it constantly needs to be adjusted and a number of around 4-5% is considered “full employment”. But this one is broken, too. It measures jobless claims as a percentage of the labor force. So it doesn’t take into account the long-term unemployed (who no longer file for benefits) or those who have given up looking since it’s just too hard.

A much better statistic to use is the labor participation rate. This is the ratio of people who are working to the overall population that’s of working age. Admittedly, it still doesn’t help identify patterns among undocumented workers, cash employment, and various other categories, but it’s a darn sight better than the unemployment rate if you only get one data point for measuring the employment situation. It would take some time to adjust since you need to know what “healthy” is, but so did unemployment rate (it isn’t obvious that 4% rather than 0% is “full” employment.)

The DOW Jones Industrial Average

The DOW is the most commonly reported stock market index and it’s also the worst. No serious analyst thinks the DOW is a good measure of the market for a few reasons.

First, it’s small. It only includes a limited number of stocks and those stocks are of the biggest companies out there. If you believe that small business can have ups and downs independent of big business (and the data would support you if you do), the DOW provides a very incomplete view.

Second, it’s price weighted. Each component of the index is weighted by price rather than by market capitalization. This is a problem since price per share is market cap divided by the number of shares and the number of shares is completely arbitrary because a company can just pick how many shares to have outstanding through splits and reverse-splits. Net net, this means a company with a high share price can have a disproportionate impact on the DOW just by having a high share price even if it’s a relatively small part of the overall economy.

For all that, using the DOW to report on the market is probably a less of a big deal than GDP or unemployment since the DOW and broader indexes like the S&P 500 move together roughly 98.5% of the time and never, to my knowledge, move in completely different directions over any significant time frame. Having said that, it’s also the easiest to fix since indexes like the S&P 500 are already widely reported and are much more accurate.