Industry CommentaryTech & LaborOpinion

Do As I Say, Not As I Commit

May 23, 2026
reheight
0 comments

A developer’s perspective on the growing gap between what companies promise publicly and how they operate internally. From opaque worker-scoring systems to declining customer experience, this piece explores how “customer obsession” and corporate values often collapse under scale, incentives, and cost-cutting — and why the same broken systems eventually impact employees, contractors, and customers alike.

I build software for a living. I've also spent time inside the kinds of organizations that talk the loudest about quality — the ones with the polished engineering blogs, the laminated values on the conference room wall, the customer-obsession keynote. Those two vantage points have taught me something nobody says from a stage: the volume of a company's external promises is often inversely proportional to how seriously it honors them internally.

The pattern isn't unique to any one company or industry. It shows up in tech, logistics, finance, healthcare, retail — anywhere an organization is big enough that the people writing the principles are far removed from the people living under them. The company projects a standard outward. Inward, that standard quietly decays. And the gap between the two doesn't stay contained. It travels — through employees, associates, and contractors — and it lands, eventually, on the customer.

There's a Name for This

Researchers call it the "value-practice gap," or more bluntly, organizational hypocrisy. It has a fairly precise definition in the literature: a company is hypocritical in the prototypical sense when it makes a moral or quality claim, signals commitment to it, behaves inconsistently with it, and — in the strong form — obscures the inconsistency, all while reaping the benefits of the claim it isn't honoring.

That's not a rant. That's a framework. And the reason it matters is that the literature is also clear on the consequences: perceived corporate hypocrisy reliably reduces consumers' willingness to buy, increases employees' intention to quit, and erodes community trust. It doesn't damage one stakeholder group. It damages all of them, through the same mechanism.

The Standard Is Real — for Customers, on Paper

Here's the thing that gives the game away. When a company builds something for a paying customer, the standards are real and enforced. Data has to be accurate. Systems have to be explainable. Errors have to be correctable. If something breaks, there's an incident response, a post-mortem, a fix. That rigor is genuine.

Except increasingly, even that is slipping — and the slippage exposes the projection for what it is. U.S. customer experience quality has now declined for four consecutive years, hitting an all-time low, with analysts pointing to companies losing focus on customer-obsessed strategy and to falling employee engagement. The most telling number I've seen comes from a 2025 global study: 80% of business leaders believe they're meeting customer expectations, while only 24% of customers agree. In the same breath, 92% of leaders call customer experience a top strategic priority — but only 28% treat it as important enough to actually fund.

That is the projection-versus-reality gap rendered in survey data. The claim is loud. The investment is quiet. The customer can feel the difference even when the marketing insists there isn't one.

The People in the Middle Absorb It First

Before the gap reaches the customer, it passes through the workforce — and that's where the contradiction is sharpest, because the same company holding customer-facing systems to a high bar quietly drops that bar for the systems that govern its own people.

A productivity score outputs a number, and nobody will explain the formula. If the number is wrong, the burden of proof falls on the worker, who doesn't have access to the data needed to prove anything. That would never survive a code review on the customer-facing side. But it runs in production indefinitely on the worker-facing side, and nobody pages on-call.

The research backs up the cost of this. A 2024 Cornell study found that when organizations use AI to monitor and score employees, workers complain more, perform worse, and want to quit more — unless the system is framed as helping them develop rather than judging them. That caveat is the whole point: the same tool lands completely differently depending on whether the people under it are treated as professionals to support or output to police. A 2025 review in New Technology, Work and Employment reached a compatible conclusion, framing algorithmic management as a set of unresolved tensions — control versus autonomy, efficiency versus fairness, transparency versus opacity — and finding that transparency, when actually provided, improves outcomes. Opacity isn't a property of the technology. It's a choice about how to deploy it.

And Contractors Absorb It Worst

The further you get from the protected core of "employee," the wider the gap grows. Contractors and platform workers sit at the far end of it, frequently subject to the same control as employees while denied the corresponding protections. Misclassification — treating workers like employees in practice while labeling them independent contractors on paper — is the structural version of the projection problem: the company claims an arm's-length relationship it doesn't actually maintain, because the claim is cheaper.

The mechanics are familiar by now. Reclassifying contractors as employees raises labor costs by an estimated 20–30% once payroll taxes, workers' compensation, and benefits are included — a direct financial incentive to maintain the fiction. Platform workers are routinely managed by opaque algorithms that assign tasks, score performance, and can deactivate someone's entire income source without warning or any meaningful way to contest it. Human Rights Watch documented this across major U.S. platforms in 2025, noting that the same companies posting record revenue run "black box" systems their workers can neither see into nor appeal.

It All Lands in the Same Place

Here's why this isn't four separate complaints. It's one system with four exits.

The retail and service literature makes the connection explicit: the store associate, the delivery driver, the support rep — these are the company's actual interface with the customer. When those people believe the company is hypocritical in how it treats them, they have little reason to extend themselves on the company's behalf. The internal gap becomes the external experience. The customer waiting on hold, getting the runaround, dealing with someone who's checked out — they're standing at the far end of a decision made several layers up, where someone chose the appearance of a standard over the cost of maintaining one.

And customers do leave. More than half say they'll abandon a company after a handful of poor experiences, with trillions in global spending at risk from degraded customer experience. The projection that was supposed to protect the brand ends up undermining it, because you cannot indefinitely advertise a standard you've stopped paying for. People notice. They always eventually notice.

Why It Keeps Happening

I don't think it's usually malice. I think it's scale and incentive structure.

When you operate at massive scale, a systematic flaw in an internal system is a rounding error in the aggregate. At the individual level it's someone's standing, their pay, their job, their dignity at work. The cost of fixing it precisely is measured in engineering sprints or labor budget. The cost of not fixing it is borne by whoever is least able to push back — the contractor, the associate, and eventually the customer. That math produces a predictable result.

Harvard Business Review put the underlying dynamic plainly in an analysis of tech-company ethics: external pressure pushes organizations to challenge their own core practices, but the countervailing pressure to keep predictable processes running for the bottom line consistently wins — which ratchets down the capacity to object to things people internally already suspect are wrong. The people building and running these systems often know. Shipping them anyway is just how the incentives are arranged.

What the Fix Actually Looks Like

None of this is a hard engineering problem. That's what frustrates me most.

Explainable systems are not a technical challenge. If you can build a recommendation engine for 300 million customers, you can document the formula that scores 10,000 workers. If you choose not to, that's a policy decision wearing a technical disguise.

Correctable errors are not a technical challenge. Audit trails, contestation workflows, version-controlled logic — standard infrastructure on the customer side. Applying it to the worker side is a prioritization choice, not a research project.

And honoring the standard you advertise is not a technical challenge at all. It's just refusing to maintain two sets of books — one for the keynote and one for the people who actually live inside the company's decisions. The promise made to the customer and the system imposed on the worker should be able to survive being placed next to each other.

If a standard only applies when someone with leverage is watching, it was never a standard. It was marketing.

The Broader Point

I'm writing this as a developer, not an activist. I care because bad systems bother me — especially bad systems that hide behind the credibility of organizations that genuinely know better and have the resources to do better.

The companies doing this aren't incompetent and aren't unaware of the principles they're violating. They publish those principles. They hire the people who could implement them correctly. The gap between what they promise the outside world and what they practice on the inside isn't an accident or an oversight.

It's a choice. And it's one the whole chain — customers, associates, contractors, and the engineers building the systems — has every reason to be a lot less comfortable accepting.

#Programming#Big Tech#Labor#Software Engineering#Corporate Culture#Opinon

Discussion (0)

Join the conversation

Sign in to comment
💭

No comments yet. Be the first to share your thoughts!