TL;DR: Most policy communication doesn’t fail because it’s badly written. It fails because plain language is only one link in a four-part chain (audience, messenger, channel, and pre-test) and the other three are usually under-resourced. Fifty years of evidence from law, public health, and behavioural science keep pointing to the same conclusion: reach is decided by who delivers a message, to whom, and in what format, far more than by the words alone.

Your Content Goes Here

Government communicators have been told for decades that the reason their work doesn’t land is that it’s written badly. Cut the jargon, shorten the sentences, run the readability check, and citizens will follow. The instinct isn’t wrong, but it’s incomplete. Plain language matters, and the evidence on its benefits is real. The trouble is that plain language is doing only one job in a chain of jobs, and the other jobs (who delivers the message, what format it arrives in, whether the audience can read it at all, and whether the messenger is trusted by the people on the receiving end) are where most of the reach gets lost.

If you build a policy communication plan on the assumption that the writing is the bottleneck, you will under-resource the parts that actually move behaviour. At Exaltus, we plan reach as a chain rather than a final-step polish, because the evidence on what reaches people is unusually settled. The pattern across half a century of government, public health, and behavioural research keeps coming back to four things, and “make it shorter” is only one of them.

Has the plain-language theory of reach actually been tested?

The most ambitious test of the writing-only theory was published by Martinez, Mollica, and Gibson in PNAS in 2024. Working from a corpus of roughly 225 million words covering every law passed by Congress from 1951 to 2022, they compared the structural complexity of federal law against four baseline genres of English. The conclusion is uncomfortable for plain-language advocates: laws remain dense with the features associated with processing difficulty (center-embedding, passive voice, low-frequency jargon, capitalization runs, and long sentences) and most of these features have not meaningfully declined since the plain-language movement began. As the authors put it, top-down efforts to simplify legal texts have so far been largely ineffectual, despite how tractable the changes look on paper.

The point isn’t that plain language doesn’t work. It does. A widely cited Cooley Law School review of more than 60 empirical studies found documented benefits of plain language for both agencies and the public, including measurable drops in confused calls, errors, and re-work. The point is that the underlying texts citizens have to interact with (laws, regulations, benefits letters, public-health notices) keep regenerating their complexity even after writers know better. If reach depended only on the writing, fifty years of effort would have moved the needle further than it has.

What happens when the audience can’t read it?

The second job in the chain is the audience’s actual literacy. The Canadian Public Health Association’s “A Vision for a Health Literate Canada” summarizes findings from the International Adult Literacy and Skills Survey and reports that roughly 60% of adults in Canada, and 88% of adults over 65, lack the capacity to obtain, understand, and act upon health information well enough to make appropriate decisions on their own. The British Columbia Medical Journal’s overview of the same data ranked Canada 11th of 18 OECD countries on adult literacy.

These numbers don’t refute plain language. They make it more important. They also explain why a brochure rewritten at a Grade 8 level still misses most older adults and many newcomers: not because the text is wrong, but because reaching the audience requires more than the right text. It requires meeting people in the format and channel where they already process information.

Who should deliver the message, and through which channel?

The third job in the chain, and the one most policy plans under-budget, is the messenger. The UK Behavioural Insights Team’s tax-compliance experiments, summarized across the OECD’s 2021 Behavioural Insights for Better Tax Administration, showed that location-specific normative messages added to reminder letters could raise compliance by roughly 15% over standard letters.

They also showed something more striking. As contact moved from impersonal to personal, the same message reached very different numbers of people:

  • Letter: about 8 in 100 recipients paid.
  • Email: roughly double that.
  • Phone call: about 30 in 100.
  • Personal visit: close to 9 in 10.

Same agency, same legal authority, same underlying message. The channel did the work. The same pattern shows up in public-health communication. The evidence review of COVID-19 communications by Ghio and colleagues, published in BMJ Open in 2023, found that what made pandemic messaging move behaviour was clarity without conflicting messages, delivery by trusted sources, and a tone that was authoritative without being controlling.

A complementary review of public trust in COVID-19 information sources, in Parliamentary Affairs in 2024, found that scientific experts and local officials carried disproportionate weight, especially for audiences who already felt distance from national government. Reach is not a property of the text. It’s a property of who hands it to whom.

Why does pre-testing decide whether policy communication gets through?

The fourth job, pre-testing with the actual end audience, is the one most often cut for time. The GAO’s review of Veterans Benefits Administration claimant letters (GAO-02-395) found that the agency rewrote its letter templates without testing them with veterans, because it lacked contracts to run focus groups.

The federal plain-language guidance the GAO cited is explicit on this point: testing documents with the audience is supposed to be integral to the writing process, not a final-step nicety. When that step is skipped, problems ride into print.

This is the step that turns the other three from intent into reach. A clearly written notice, delivered by the right messenger, in the channel where the audience already pays attention, still has to be checked against the audience’s actual understanding before it ships. The cost of that step is small; the cost of skipping it is invisible until uptake numbers come in.

What does policy communication that reaches people actually do?

A communication plan that takes reach seriously names four things and budgets for each:

  • The audience. Not “the public” or “Canadians.” The specific person, in their specific situation, deciding the specific thing the policy is designed to change.
  • The messenger. A trusted local actor, a community organization, a primary-care clinician, a cultural broker, paired with the formal authority of the agency. A message from someone the audience already trusts will outperform the same message from someone they don’t.
  • The format and channel. A letter is not a phone call is not a short video at point-of-care is not a poster at the bus stop. The choice of channel is not about preference; it’s about how much reach it produces per dollar.
  • The pre-test. A small focus group, a quick comprehension test, a one-week pilot with the actual end users, built into the timeline before the notice goes wide.

That’s the chain. Most policy communication fails to reach people not because the writers are careless, but because three of those four jobs were under-resourced or never named. Naming them is what turns a notice into something the audience hears.

Is plain language enough?

Plain language is necessary. It is not sufficient. The reach problem in government and public-health communication is mostly a problem of who, where, and when, wrapped in the appearance of a writing problem.

Fixing the writing without fixing the rest produces a clearer text that still doesn’t get through. Fixing the rest is what turns policy communication into something the public can use.

That’s the part the evidence has been pointing at for fifty years. It’s also the part most communication plans still treat as someone else’s job.

Frequently asked questions about policy communication

It usually fails on delivery, not wording. Plain language is necessary, but reach also depends on who the message comes from, the channel it arrives through, and whether it was tested with the audience first. When three of those four jobs are under-resourced, even a clearly written notice doesn’t land.

Yes. Reviews of more than 60 studies link plain language to fewer confused calls, fewer errors, and less re-work. The point is not that plain language fails, but that it can’t carry reach on its own.

It depends on the stakes, but personal contact consistently outperforms paper. In the UK tax experiments, letters moved about 8 in 100 recipients, phone calls about 30 in 100, and personal visits close to 9 in 10, with the same underlying message.

Heavily. Roughly 60% of Canadian adults, and 88% of those over 65, fall below the literacy level needed to act on health information without help. A Grade 8 rewrite still misses many of them, which is why channel and messenger matter as much as the text.

It means checking the message against the actual audience before it ships: a small focus group, a comprehension test, or a one-week pilot. The GAO found agencies skipping this step for lack of contracts to run it, and the problems rode into print.

Name and budget for four things: the specific audience, a trusted messenger, the right format and channel, and a pre-test. Plain language is one job inside that chain, necessary but not the whole plan.

If you’re planning a policy or public-health communication that has to reach a wide, diverse audience, we can help you build reach into the brief rather than bolt it on at the end.

You can also see how visual storytelling fits inside a government communication plan, or browse examples of our work.