Designing Behavior for Others — Ethically

Every product you build is a behavior-design machine, whether you planned it that way or not. A button placement, a notification timing, a progress bar — each one nudges people toward an action. The question is not whether you are shaping behavior. You always are. The question is how consciously you do it, and in whose interest.

This section is about wielding behavioral science as a builder — honestly. We will cover how engagement mechanics actually work (the psychology under the hood), where the ethics line sits, and a set of concrete tests you can run on your own product to stay on the right side of it.

How Retention and Engagement Actually Work

Before you can behave ethically, you need to understand the tools. Engagement is not magic — it is applied behavioral science. Here are the main levers.

The Hook Model (Nir Eyal)

In his 2014 book Hooked, researcher and author Nir Eyal described a four-step loop that habit-forming products run over and over:

  TRIGGER
  (external: notification, email, ad)
  (internal: boredom, anxiety, loneliness)
       |
       v
  ACTION
  (the simplest behavior in anticipation of a reward)
       |
       v
  VARIABLE REWARD
  (the payoff — but unpredictable in size or timing)
       |
       v
  INVESTMENT
  (the user puts something in: data, content, time, social connections)
       |
       v
  TRIGGER  (now mostly internal — loop tightens)

Each pass through the loop builds a stronger habit. The trigger moves from external (a push notification) to internal (you open the app automatically when you feel bored). That shift — from external nudge to internal craving — is what separates a habit from a one-off action.

Variable Rewards — Why Unpredictability Is Addictive

The most powerful element in the Hook Model is the variable reward — a payoff that varies in size or timing so you can never quite predict what you will get. This is not new: psychologist B.F. Skinner showed in the 1950s that animals on a variable-ratio schedule (reward comes after an unpredictable number of responses) press a lever far more than animals rewarded every single time.

Social media feeds are a textbook variable reward machine. Every scroll might reveal something exciting or nothing at all. That unpredictability triggers the brain's dopaminergic system — the same circuit that fires for gambling and drugs.

Progress Bars and the Endowed Progress Effect

A progress bar is not decoration. It is a behavioral trigger. Two well-documented psychological effects explain why:

• Zeigarnik Effect — named after Soviet psychologist Bluma Zeigarnik, who noticed in the 1920s that people remember unfinished tasks far better than completed ones. An open progress bar is an unfinished task your brain keeps wanting to close.
• Endowed Progress Effect — people work harder toward a goal when they are given a head start. If you give a user a checklist with the first item already checked ("Account created ✓"), they complete more steps than if the checklist starts at zero. The artificial head start makes the finish line feel closer.

This is why good onboarding flows show "3 of 5 steps complete" immediately after sign-up. Research shows 25% of apps are used only once — a well-designed onboarding that uses these effects dramatically narrows that gap.

Streaks and Loss Aversion

A streak is a count of consecutive days (or sessions) a user has performed an action. Streaks work through loss aversion — the well-established finding (from Kahneman and Tversky's Prospect Theory, 1979) that losing something hurts roughly twice as much as gaining the equivalent thing feels good. Apps using streak mechanics see up to 35% higher daily active user rates compared to those that rely on positive rewards alone.

Notifications — The External Trigger That Can Turn Toxic

Notifications are the most direct behavior-design lever you have. They literally interrupt a person's life. The ethical rule is simple: a notification should earn its interruption. It should deliver real value to the user at that moment — not just pull them back to your engagement metrics.

The Ethics Line: Persuasion vs. Manipulation

The word "persuasion" means helping someone make a decision that serves them. The word "manipulation" means pushing someone toward a decision that serves you, often by bypassing their rational judgment. The gap between them is where ethics lives.

Dark Patterns — Named and Defined

The term dark patterns was coined by UX designer Harry Brignull in 2010 to name UI tricks that deliberately deceive users. Here are the most common ones builders encounter:

• Roach Motel — easy to get in, very hard to get out. The name comes from the old pest-control ad: "You can check in but you can't check out." Example: signing up for a gym membership online in two minutes but being required to visit a physical branch in person to cancel. Amazon settled a USD 2.5 billion lawsuit in part for using this pattern with Prime subscriptions.
• Confirmshaming — the decline button is written to shame the user for saying no. Example: an email signup popup where the "No thanks" option reads "No, I'd rather pay more" or "No, I don't want to save money." It is designed to make you feel foolish for opting out.
• Fake Scarcity — inventing urgency that does not exist. "Only 2 left!" counters that reset daily. "Sale ends tonight!" banners that run permanently. These exploit the psychological fear of missing out (FOMO) to force a fast decision before rational thought kicks in.
• Hidden Costs — prices are low in the cart but fees, taxes, or "service charges" inflate the total at the final checkout step, after the user has already committed mentally.
• Misdirection — the design draws attention to one thing while hiding another. Pre-ticked checkboxes for marketing emails or add-on purchases, buried in a wall of text, are a classic example.
• Trick Questions — confusingly worded opt-in/opt-out language where the "safe" choice is actually the one that opts you into something you don't want.

Eyal's Manipulation Matrix

Nir Eyal offered a practical decision tool in Hooked: a 2×2 grid that plots your product against two questions. It does not give you a final answer, but it gives you a clear starting point.

Question 1: Does this product materially improve the user's life?
Question 2: Would the maker use this product themselves?

                  MAKER USES IT
                  Yes            No
                +-------------+-------------+
  IMPROVES  Yes | FACILITATOR | PEDDLER     |
  USER'S        |             |             |
  LIFE          | (build it)  | (question   |
                |             |  your       |
                |             |  motives)   |
                +-------------+-------------+
              No | ENTERTAINER | DEALER      |
                |             |             |
                | (ok if      | (don't      |
                |  harmless)  |  build it)  |
                +-------------+-------------+

• Facilitator — you use the product yourself and believe it genuinely helps users. This is the ethical green zone. Build it.
• Peddler — you believe the product helps users but you would not use it yourself. You are guessing at someone else's needs from a distance. Proceed with caution and more user research.
• Entertainer — you use it yourself but it only entertains, does not meaningfully improve lives. Games and media often sit here. Not unethical, but not transformative either.
• Dealer — you would not use it yourself and you know it does not really improve users' lives. You are extracting value from users, not creating it for them. Do not build this.

The Regret Test

The Regret Test is the simplest practical ethics check for any behavior-design decision. Ask:

"If the user completes this action and then thinks about it the next day, will they regret it?"

If the answer is yes — or even "probably" — do not use the technique. The regret test works because it shifts your perspective from the moment of the nudge (when you control the frame) to the moment of honest reflection (when you do not).

Healthy Engagement vs. Addiction

Nir Eyal draws a sharp line between habits and addictions: addiction is always harmful by definition. A habit can be good or bad. The same behavioral mechanics that build a healthy morning exercise routine can also build a harmful gambling dependency. The mechanics are identical; the consequences are not.

How do you tell the difference when designing a product?

The tell-tale sign of addictive design: the user's behavior diverges from their own stated preferences. They say they want to spend less time on the app. The app is designed to ensure they spend more. That gap — between what users want and what the product drives them to do — is where manipulation lives.

Concrete Dos and Don'ts for an Honest Product Builder

Do

• Use real progress bars — show users their actual progress toward a real goal (profile completion, order status, onboarding steps). They reduce anxiety and drive genuine engagement.
• Send notifications that earn their interruption — "Your order shipped" is valuable. "We miss you!" on day 3 is not. Set a high bar: would a friend send this message?
• Make cancellation as easy as signup — one-click unsubscribe, no "call us to cancel," no guilt-trip retention flow that buries the cancel option. Users remember how you treated them on the way out.
• Use real scarcity honestly — if you only have 3 units left, say so. If the sale ends at midnight, let it end. Real scarcity is a legitimate reason to act quickly. Fake scarcity is a lie.
• Default to what the user wants — 90% of users want the sensible default. Pre-tick the option that helps them; never pre-tick the option that benefits only you.
• Explain disabled states — if a button is greyed out, tell the user why and what they need to do. A disabled button with no explanation is a tiny dark pattern.
• Run the Regret Test on every nudge — build it into your design review checklist.

Don't

• Don't use confirmshaming — write the decline option with respect: "No thanks" is always acceptable. Never shame users for opting out.
• Don't build a roach motel — anything easy to start must be easy to stop. Subscriptions, accounts, marketing emails.
• Don't manufacture urgency — no fake countdown timers, no "X people are viewing this" when it's fabricated, no stock counters that reset every day.
• Don't hide costs — show the final price as early as possible. Surprises at checkout destroy trust and spike cart abandonment.
• Don't bury opt-outs in walls of text — if a user is giving you permission, make sure the checkbox is clearly visible and clearly labeled.
• Don't design infinite scroll where it serves only you — if your product has no natural stopping point, consider whether that serves users or traps them.

Tying This to Non-Technical Store Owners

If you are building tools for non-technical people — store owners, small business founders, people who did not grow up reading tech blogs — the ethical obligation runs even higher. These users are often less able to detect when a UI is working against them. They trust that the software they pay for is on their side.

When a store owner sets up their print shop, they are not reading UX research papers. They navigate by instinct and trust. If your product hides the cancel button, uses fake scarcity in its own upgrade prompts, or sends guilt-driven "Your trial ends tonight!" emails at 11 PM — you are exploiting that trust. You are not being clever. You are being dishonest with someone who cannot easily fight back.

The standard is simple: build the product you would want your own parents to use. Clear labels. Honest pricing. Easy cancellation. Notifications that help, not harass. Defaults that serve them, not your metrics.