Customer ethics and the asymmetry most companies never name

Every commercial relationship runs on an asymmetry. You know what your product costs to make, what it does badly, what the renewal price will be, how hard it is to leave, and which of your customers are not using what they pay for. They know almost none of that.

The asymmetry is not itself wrong: it is what expertise is. Customer ethics is about what you do with it.

What to take away

The plain-words test

One test does most of the work here, and it is worth learning before any framework.

Take the practice and describe it to the customer, out loud, in the words you use internally. Not the marketing description. The internal one.

We make canceling hard so a percentage of people give up. We show a higher price to people whose browsing suggests they will pay it.

The free trial converts because most people forget the date, and we renew at a rate we hide until after the charge. Support is measured on retention, so they offer three things before letting you go.

If the sentence is not sayable, the practice is not defensible. The gap between the internal description and the external one is the exact size of the ethical problem.

This test has a useful property: it does not require agreeing on any moral theory. It only requires you to notice when you are relying on the customer not knowing something.

Where the pressure comes from

Nobody sits down to design a manipulative flow. What happens is that a conversion rate is a number, someone is accountable for it, and a series of individually small changes each move it up.

Removing a step. Pre-ticking a box. Making the decline option gray. Moving the cancel link one level deeper. Shortening the reminder. Each is defensible alone. The cumulative result is a system optimized against the customer's intent, built entirely out of reasonable-sounding decisions by people who would each have refused the whole thing.

The structural fix is to test changes against something other than the immediate conversion:

Practices worth naming

Consumer-facing selling is legally regulated in most markets. Several practices above are directly addressed by law in some jurisdictions and not others. In the United States, Section 5 of the Federal Trade Commission Act gives the FTC authority to address unfair or deceptive acts or practices in commerce. Nothing here is a statement of what is permitted where you operate. That is a question for counsel, and the answer differs by country and often by sector.

The FTC's telemarketing sales rule is one example of how specific such rules get about disclosure and authorization in the transactions it covers.

Data used for a purpose nobody agreed to

The characteristic modern version of this problem is not a breach. It is a legitimate collection reused for something the person never contemplated.

Someone gives you their address to deliver a parcel. Two years later it becomes a feature in a model that decides what they are offered.

Someone contacts support about a medical matter, and that becomes a segment. Someone's usage pattern, collected to keep the service running, is sold as an aggregate. It is not as anonymous as everyone assumes.

Practical questions before a reuse:

Before a secondary use, ask:

Data protection is heavily regulated, and rules differ across jurisdictions. In the United States, examples include the California Consumer Privacy Act (CCPA), as amended by the California Privacy Rights Act (CPRA), and the California Privacy Protection Agency (CPPA), a relevant state regulator. Coverage and obligations depend on the business, data, and use; differences include what counts as consent, what a legitimate secondary use is, and what rights people have.

In the United Kingdom, the Information Commissioner's Office guidance on what makes consent valid shows how narrow a permission obtained at collection can be.

Customers who cannot protect themselves

A large customer base includes people who cannot assess what they agree to. They may be in acute financial distress, recently bereaved, facing declining cognition, a language barrier, or emergency time pressure.

Standard practice treats these as edge cases. They are not edge cases; they are a predictable fraction of every book of business, and they are systematically the most profitable segment for a badly designed product, because they are the least likely to leave, complain, or read.

What can actually be done:

Financial services, credit, insurance, and health products carry specific legal duties toward vulnerable customers in many jurisdictions. In the United States, the applicable framework depends on the product and circumstances: the Consumer Financial Protection Bureau (CFPB) enforces federal consumer financial laws, including the Consumer Financial Protection Act's prohibition on unfair, deceptive, or abusive acts or practices for covered entities; state insurance departments oversee insurance; and the Health Insurance Portability and Accountability Act (HIPAA) applies to covered entities and business associates handling protected health information. These rules are sector-specific, so do not infer which duties apply to a product from a general article.

What you say about the product

Marketing claims are the easiest place to get this right and the easiest to let slip.

The test is not whether the claim is technically true. It is what a reasonable reader takes away. A claim can be composed entirely of accurate statements and still leave everyone with a false impression, and that is the version most likely to cause harm, because it survives internal review.

Pricing that you would explain

Differential pricing is not automatically wrong: volume discounts, student rates, and regional pricing are all differential and all defensible. What separates the defensible from the rest is whether the reason for the difference is one you would state.

"You pay less because you buy more" is fine. "You pay less because you are new and we want you" is fine, if the existing customer can see their rate.

"You pay more because our model thinks you are unlikely to shop around" is the one that fails, and it fails the plain-words test, not any technicality.

The related trap is the loyalty penalty: the price for staying quietly drifting above the price for arriving. It is easy to build without anyone deciding to, and it is hard to defend when a customer notices, because the answer is always some version of "we charged you more because you were not paying attention."

If you have a pricing structure like this, the useful question is not whether it is permitted. It is what happens when a customer puts the two prices side by side in public.

Letting people leave

How a company behaves at the end of a relationship reveals more than anything it does at the start, because there is no longer anything to gain.

The questions to answer honestly about your own product:

Lock-in built from genuine value is fine: people stay because leaving means giving up something good. Lock-in built from friction, data hostage-taking, or a deliberately painful migration is a different thing, and everyone inside the company knows which one they have built.

Complaints as information

Most organizations treat complaints as a cost to be minimised. The number that gets managed is complaint volume, which produces exactly the wrong behavior: the cheapest way to reduce complaint volume is to make complaining harder.

Better measurements to hold alongside it:

A functioning complaints process is one of the few genuine early-warning systems a company has. Treating it as a reputational containment exercise wastes it.

Where to start

If you are looking at a product or a funnel for the first time:

That list will produce more than any values statement, because each item ends in something specific that changes on Monday.

Related reading on this site: ethical frameworks covers reasoning when customer interest and another obligation genuinely conflict. codes of conduct turns the practices above into rules people can apply.

leadership ethics covers why the conversion target usually wins. employee ethics covers what it costs the support agent who notices.

Common questions

Is differential pricing wrong?

Not in itself. Volume discounts, student rates and regional pricing are all differential and all defensible. What fails is a difference you would not state, and the loyalty penalty is the common example.

How do we know whether a flow is manipulative?

Describe it to a customer in the words you use internally. If the sentence is not sayable, the practice relies on them not understanding it, and no amount of disclosure in the terms changes that.

What should we do when we overcharge people by mistake?

Identify everyone affected and refund in the form they paid, without waiting to be asked. A credit against future purchases keeps their money in your business, and customers read the difference correctly.