Nearly every company has a code of conduct, and nearly every employee has signed off on it. That is the good news in LRN’s 2026 Code of Conduct Report, “The Culture Behind the Code.” The harder news is that the signature may be doing most of the work. The report combines an analysis of more than 1,000 codes with a survey of roughly 2,000 employees across multiple countries, and its central message is that the compliance challenge has moved. Writing a code is no longer difficult. Making it something people actually reach for is.

Adoption Is Near Universal, Use Is Slipping

Ninety-five percent of surveyed employees say their company requires them to certify or acknowledge the code, and for 68% that happens every year. By the usual program metrics, this looks like success. But the report separates reach from use, and the use numbers are moving in the wrong direction. Sixty-three percent of employees say they have used the code as a resource, down from 70% the year before. Only about one in three says they use it often or very often outside of formal training.

The format question may be part of the story. Most employees still receive the code as a PDF (65%) or a printed copy (48%), while website access, at 38%, grew six points. A static document is a poor match for how employees now look for answers, which is to search, expect personalized results, and expect an answer right away. The report is careful to say the data do not prove why use has declined, and I would be equally careful. But the pattern is a fair prompt for any compliance officer who counts annual certification as evidence of an effective code.

Findable Is Not the Same as Usable

When employees do open the code, many hit friction. Only 42% report no challenges. The most common complaint, from 24%, is that the code is too long or hard to navigate. Twenty percent say it lacks practical guidance for real situations, 19% find it static, and 18% have trouble finding specific information.

The code analysis helps explain this. Most codes have the basics of good design: 83% have a good layout, 80% link to related policies, and 66% have a table of contents. The tools that help someone decide what to do in a hard moment are far rarer. Only 39% include an ethical decision-making model, and only 29% include scenario-based questions and answers. In other words, many codes tell employees where to look but not how to think. The report points to LRN’s earlier culture research, in which 23% of employees said it is acceptable to break the rules to get the job done, and 14% reported personally violating their company’s code or standards in the prior year. Those are exactly the moments when a rule on a page may not be enough, and a short scenario or a simple set of decision questions might be.

The Code Is Behind the Risk Curve

On content, the foundations are solid. Conflicts of interest and confidential information each appear in 98% of codes, bribery and corruption in 96%, and gifts and entertainment and company assets in more than nine in ten. Emerging and operational risks are a different story. Labor laws appear in 51% of codes and international trade in 53%. Security and crisis management and competitive intelligence appear in roughly a third. And only 9% of codes explicitly address AI or technology ethics.

Readers of this blog will recognize the problem. Trade, forced labor, sanctions, and AI are among the areas where enforcement attention and corporate exposure are growing fastest, and they are the topics a typical code addresses least. LRN’s recommendation is sensible. Do not try to add every new risk to the code. Tie the code to the company’s latest risk assessment, connect new technology guidance to principles employees already understand, such as accountability, fairness, transparency, and protecting confidential information, and link out to supporting policies and training so that guidance can change without the document ballooning.

Training Is Not Enough, and Speak-Up Confidence Is Slipping

Training works when employees get it: 97% of those trained say it helps them understand how to use the code. But training reached 78% of employees in 2026, a decline from the prior year, and only 61% of employees say their manager talks about the code, down from 67%. The manager gap is also uneven. Seventy-seven percent of executives say their manager discusses the code, compared with 60% of supervisors and managers, and fewer than half of office and administrative staff and trades and technical employees. The people furthest from senior leadership hear the least.

The speak-up data may be the most important in the report. Reporting channels appear in 97% of the codes analyzed, yet only 42% explain what happens after a report, such as the investigation process. On the employee side, 69% say their code explains the company’s position on retaliation, down from 75%, and only 66% are confident they can report misconduct without retaliation, down from 71%. The report does not say why confidence declined. But a company can have a hotline, a policy, and a certification rate near 100% and still have a third of its workforce unsure that speaking up is safe. DOJ’s evaluation of compliance programs asks whether employees are comfortable reporting and whether the program works in practice, and these are the numbers a prosecutor or regulator would want to see.

Depth Versus Usability

The most interesting finding concerns length. Codes LRN rated as more effective scored 91% on its effectiveness framework, compared with 53% for less effective codes. They also run longer, averaging 36 pages and about 10,400 words, versus 11 pages and about 3,600 words. LRN is clear that this relationship is associative and not causal: more effective codes cover more of the elements it evaluates, and covering more takes space. The takeaway is not to lengthen the code. It is to make depth manageable through structure, search, and links.

The usability cost of getting this wrong is not shared equally. Thirty-four percent of Gen Z employees say the code is too long or hard to navigate, compared with 27% of Millennials, 16% of Gen X, and about one in ten Boomers. Eighty percent of employees whose first language is not English say the code is available in their language, which also means a meaningful share cannot count on that. A code tested only on the compliance team will not tell you how it lands with a new hire in a distribution center.

What Compliance Officers Should Do

LRN’s recommendations translate into a practical agenda, which I would organize as follows.

First, measure use and not just completion. Track code searches, page visits, and hotline questions that trace back to code topics, and ask employees in pulse surveys whether they used the code and whether it helped. Certification shows reach, and that is all it shows.

Second, make the code a front door. Publish it in a web-based, searchable format and link each section to the relevant policy, training module, and reporting channel. Keep essential guidance in the code and push detail to the linked material.

Third, add decision tools. Put a short decision-making model and scenario-based Q&As in the sections where employees face real pressure, such as gifts, third parties, customer demands, and information sharing.

Fourth, tie the code to your risk assessment. If trade controls, sanctions, forced labor, data protection, or AI are among your top risks, the code should say something about them, even if the detail lives in a policy.

Fifth, explain what happens after someone speaks up. Describe the investigation process, who is involved, how confidentiality works, and what protections against retaliation look like. Then test whether employees believe it.

Sixth, equip managers. Give supervisors short, role-specific talking points so the code comes up in team meetings and not only in annual training. Prioritize frontline and technical populations, where the reinforcement gap is widest.

Seventh, test usability across the workforce. Review results by generation, language, location, and work arrangement, and fix the places where the code is hardest to use.

The Bottom Line

A signed acknowledgment is evidence that a code was delivered. It is not evidence that the code was read, understood, or applied, and it says nothing about whether an employee facing a hard decision will turn to it. LRN’s report is a useful reminder that effectiveness is measured at the moment of decision: can the employee find the answer, trust it, and trust that raising a concern will be safe. Companies that can show that, with data, will have a far stronger story than those that can only show a 100% certification rate.

Source: LRN, 2026 Code of Conduct Report: The Culture Behind the Code. Figures cited are from LRN’s code analysis (more than 1,000 codes) and employee survey (approximately 2,000 respondents).