Is Optimization the Ethical Form of Personalization?

Geese.png

I began organizing in the summer of 2008 during Obama’s first run for president. Gmail was widely in use, but not to the extent that it dominates the free email market today, Facebook had opened itself up to public use just 18 months earlier, and people still used Meetup.com regularly. DailyKos, the initial driver behind the now-massive Netroots Nation annual conference was already 6 years old but mainly read by the politicos who wrote there.

As an organizer, probably the coolest thing you could do on the internet was figure out how to hack Facebook’s messaging tool to ask huge swaths of newly addicted social networkers to turn up at your phone bank. But the most effective thing you could do with efficiency? Send a personalized email directly to someone’s inbox. PACs and campaigns alike were raising huge sums of money from small-dollar donors who’d only just begun subscribing to action-oriented email lists—long before McSweeny’s published the hilarious and devastating takedown “My Mom Interned at a Nonprofit and Now All Her Emails Are Different.” Personalization was still mostly [Insert_FNAME] tags and custom subject lines based on where you lived. And people LOVED it (Read: Personalization made people do stuff at high rates).

Somewhere over the last ten years or so, we’ve all become so used to personalization that obnoxious has transformed into a blasé indifference—a once-useless feeling that now actually converts sales better than actual delight or excitement.

Somewhere in there, though, big business caught on and Google got in on the game with Remarketing and changed everything. What was once a game owned by email only after you’d disclosed personal information to a business or organization was now a semi-anonymous game of follow the customer. The name Zappos became synonymous with the same ad showing the same pair of shoes you’d already decided not to buy every time you arrived on Facebook. Between 2000 and 2010, personalization went from cute to obnoxious. And somewhere over the last ten years or so, we’ve all become so used to personalization that obnoxious has transformed into a blasé indifference—a once-useless feeling that now actually converts sales better than actual delight or excitement.

Personalization has become a mirror that reflects the same things you already like and are back to you in ways that make you feel like you need more of what you already have and like.

Personalization has become a mirror that reflects the same things you already like and are back to you in ways that make you feel like you need more of what you already have and like. (Why? Because AI, which is the underpinning of most personalization software, leaves little room for evolution and agency: AI relies on datasets to learn you and then it has a hard time changing its “thoughts” about what you like or don’t like until you tell it otherwise—but that’s a thought for another time.)

Now that personalization is driven by AI eating up massive datasets—including your own choices—whatever game it was for organizers like me in 2008 to personalize an email to improve action rates, or for Zappos to get you to buy those shoes you already kind of have? AI won that game, and humans are no longer involved in the rule-making. It’s not that personalization isn’t about genuine connection anymore. It’s that—for the most part—the motivation of genuine connection has been supplanted by the motivation for genuine profits.

So how do nonprofits working to change the circumstances of our day to day inequities—emboldened by abhorrent data privacy regulations even in liberal democracies like the United States—keep up? Of course organizations must continue to use personalization as a tactic: When digital tastes change, it’s adapt or be left behind. My proposal is this: When it comes to data that you’ll use to develop genuine connections with individuals on a one-to-one level, where what data you collect will be limited to what’s necessary and access to that data will be limited to those who truly need it, go forth and personalize. But when you can’t control which 3rd parties will have insight into the decisions you drive from users and the data it generates, optimize instead.

But when you can’t control the which 3rd parties will have insight into the decisions you drive from users and the data it generates, optimize instead.

Segmenting groups rather than pinpointing individuals is one way to protect identities. It’s not a perfect solution, but one that nonprofits working to forward user rights and anonymity in digital should consider. Sure, you may lose a little calculated lift from your revenue efforts, or a couple people from event turnout, but message your organizational stance well and maybe—just maybe—noting that the digital tools you employ online won’t subverting the longterm impact your organization seeks to make IRL will soon be as compelling as Dear [Insert_FNAME] was 10 years ago.