The Paradox of Choice

The Jam Catastrophe

In the year 2000, a psychologist named Sheena Iyengar walked into a Draeger's supermarket in Menlo Park, California, and set up a jam tasting booth. Some days she offered 24 varieties. Other days, just 6. What happened next became one of the most cited experiments in behavioral science — and one of the most controversial.

The big display drew crowds. Sixty percent of shoppers stopped to sample when faced with the magnificent wall of 24 jams, versus only 40 percent for the modest display of 6. So far, so obvious: more stuff attracts more eyeballs. This is why supermarkets have 47 kinds of toothpaste.

But here's the twist. Of those who stopped at the 24-jam table, a mere 3 percent actually bought a jar. At the 6-jam table? Thirty percent.1 The big display was ten times worse at converting browsers into buyers. Something about all those options — the Boysenberry Delight sitting next to the Black Cherry Preserves sitting next to the Blackberry Compote — didn't inspire appetite. It inspired paralysis.

Iyengar's result was elegant because it confirmed something we all feel but rarely articulate: sometimes, when you face a wall of options, you don't feel empowered. You feel exhausted. And when you're exhausted, you do the easiest thing available, which is nothing at all.

Barry Schwartz, a psychologist at Swarthmore, built an entire book around this idea. The Paradox of Choice (2004) argued that modern abundance was making us miserable.2 More options in jeans, retirement plans, phone carriers, romantic partners — all of it was creating not liberation but anxiety. The subtitle said it all: "Why More Is Less."

Schwartz's argument was persuasive and popular (his TED talk has over 17 million views). But here's what makes the story interesting for us: the paradox of choice isn't just a vibe. It has mathematical bones. And those bones tell a more nuanced story than "fewer is better."

The Mathematics of Too Much

Let's put some rigor behind the intuition. Suppose you're choosing among N options and you genuinely want the best one. What does the math say about your prospects as N grows?

Problem 1: Finding the Best Gets Harder

If you're evaluating options sequentially and you can't go back — like apartment-hunting in Manhattan — you face a version of the famous secretary problem. The optimal strategy is to reject the first roughly N/e candidates (about 37%), then accept the next one that beats everything you've seen.3 Even with this optimal strategy, your probability of picking the absolute best is only 1/e ≈ 36.8%, regardless of how many options there are.

That's actually the good news. The bad news is that most of us don't use the optimal strategy. And even if we do, we still fail to pick the best two-thirds of the time.

Problem 2: Comparison Costs Explode

If you want to compare every option against every other option (which is what a maximizer feels compelled to do), the number of pairwise comparisons grows as:

This is O(N²). Your cognitive budget doesn't grow quadratically with the number of options on the shelf. Something has to give — either you stop comparing (satisficing) or you collapse in the jam aisle.

Problem 3: Regret Scales with N

Here's perhaps the cruelest part. When you finally do choose, the opportunity cost of your decision — the nagging sense of what you might have missed — grows with the number of unchosen alternatives. More precisely, if option values are drawn from a fixed distribution, the expected value of the best unchosen option grows as O(log N).4

This is Hick's Law meeting regret theory. William Edmund Hick showed in 1952 that decision time increases logarithmically with the number of choices. Your brain needs log₂(N) bits of information to pick one item from N, which is why 100 options don't feel 50 times harder than 2 — they feel about 7 times harder. But "7 times harder" is still a lot harder when you're standing in the toothpaste aisle at 11 PM.

Try It Yourself

Theory is nice, but let's make this visceral. Below is a three-round experiment. Each round, I'll ask you to pick your favorite color. I'll measure how long you take and ask how satisfied you feel. Then we'll see if the paradox shows up in your own behavior.

The Optimal Number of Options

So if too few options leave you with a bad match and too many drown you in comparison costs and regret, there must be a sweet spot. And there is. The math gives us an inverted U-curve: satisfaction rises with more options (better chance of a good match), peaks, and then falls (as cognitive costs and regret overwhelm the marginal gain).

We can model this. Suppose each additional option has a small probability of being better than your current best, but every comparison costs you some cognitive effort, and your regret over missed options grows logarithmically. The net satisfaction curve looks something like this:

Play with the calculator below to see how the sweet spot shifts depending on how costly your comparisons are, how much you value finding the absolute best, and how your options are distributed.

The Secretary Problem, or: How to Stop Looking

If the paradox of choice is the disease, the secretary problem is the cure — or at least the best treatment mathematics has to offer.

The setup: you're interviewing N candidates for a position. You see them one at a time and must decide on the spot whether to hire each one. No callbacks. The optimal strategy, as we noted, is to reject the first 37% and then hire the next candidate who beats all previous ones.

This strategy is elegant because it converts an impossible problem (evaluate everything simultaneously) into a manageable one (set a benchmark, then act decisively). It's the mathematical formalization of "date around, then commit" — which, not coincidentally, is exactly the advice that satisficers follow by instinct.

Netflix Knows This

Netflix has roughly 18,000 titles in its global catalog. By the paradox of choice, this should be a disaster — users staring at their screens in catatonic indecision, eventually giving up and reading a book. And in the early days of streaming, that's more or less what happened. Netflix engineers found that if users spent more than 60–90 seconds browsing without selecting something, the probability of them watching anything dropped to nearly zero.6

Their solution was the recommendation algorithm: a massive machine learning system whose primary job is not to show you more but to show you less. The Netflix home screen presents roughly 40–50 titles from those 18,000. It's an artificial constraint on choice, engineered to put you in the satisficer's sweet spot.

Every recommendation engine — Spotify's Discover Weekly, Amazon's "customers also bought," YouTube's autoplay — is fundamentally a choice-reduction machine. They exist because the companies learned, empirically, what Iyengar demonstrated in the jam aisle: unconstrained choice is the enemy of action.

Wait — Is Any of This Actually True?

Here's where the story gets properly interesting for anyone who cares about mathematical honesty. In 2010, Benjamin Scheibehenne and colleagues published a meta-analysis of 50 studies on choice overload. Their finding was striking: the average effect size was essentially zero.7

Some studies replicated Iyengar's jam result. Others found the opposite — more choice led to more purchases, not fewer. The overall picture was not "choice overload is real" or "choice overload is fake" but rather "it depends."

Depends on what? The meta-analysis identified several moderating factors:

This is the nuanced truth that the math predicts and the data confirms: the paradox of choice is not a universal law. It's a conditional phenomenon. It emerges when comparison costs are high relative to the marginal quality gain from more options — which is exactly what the inverted U-curve above models. Slide the cost slider up, and the optimal number of options drops. Slide it down (expertise makes comparisons cheap), and you can handle many more.

The replication crisis didn't kill the paradox of choice. It refined it. We went from "more is less" (a bumper sticker) to "more is less when the cost of comparing exceeds the expected gain from a better match" (a theorem). The math was right all along — we just needed the right equation.

Living in the Sweet Spot

So what does the mathematics of choice actually tell us about how to live?

First: be a satisficer on purpose. Set a threshold before you start looking. "I want jeans that fit, cost under $80, and come in dark blue." Then buy the first pair that clears the bar. This isn't settling — it's the optimal stopping strategy with a known utility function.

Second: reduce options deliberately. The Netflix engineers aren't hiding content from you maliciously. They're doing you a favor. When you curate your own choices — picking three restaurants to consider instead of scrolling the entire Yelp catalog — you're performing the same service for yourself.

Third: know when more is actually better. If you're an expert, if your preferences are clear, if the decision is reversible — go ahead and browse the full catalog. The inverted U-curve shifts to the right for you. More options genuinely help when you have the cognitive infrastructure to evaluate them cheaply.

And finally, the deepest lesson: the goal is not the best choice. The goal is a good choice, made at reasonable cost, with minimal regret. Herbert Simon won a Nobel Prize for figuring this out. The rest of us are still standing in the jam aisle, paralyzed by Boysenberry Delight, hoping someone will just hand us the strawberry.8