In government operations, a campaign only earns its keep if it helps people act: understand a change, trust a message, complete a task, or respond quickly. I treat an information campaign as an operational tool, not a publicity exercise, so the real measure is whether it reduces confusion and supports delivery. This guide breaks down the parts that matter most: the objective, audience insight, message design, channel mix, accessibility, evaluation, and the mistakes that quietly weaken results.
The best campaigns change behaviour because they solve a real audience problem and prove it with evidence.
- In UK public-sector practice, campaign planning usually follows OASIS: objective, audience insight, strategy, implementation, and scoring.
- Strong campaigns focus on one primary action and a few supporting messages, not a long list of things to say.
- Accessibility is part of performance: clear language, captions, alt text, contrast, and accessible formats matter from day one.
- Testing works best when it isolates one variable and answers a real policy or delivery question.
- Outputs are not outcomes; if the campaign did not change understanding or behaviour, the numbers are incomplete.
What this type of campaign is really for
In the public sector, the job is rarely just to “raise awareness”. More often, the campaign has to move people through a practical step: renew a licence, use a digital service, follow a safety instruction, understand a policy change, or trust a new process enough to act on it. That is why I think of campaign work as a bridge between policy and behaviour. If the bridge is weak, the policy may still be sound, but the public experience will feel confusing or fragmented.
That is also why I separate campaigns from routine notices. A routine email or service alert may inform; a campaign should change the odds of a specific action happening. In UK government operations, that distinction matters because time, budget, and attention are always limited. If the campaign is not tied to an outcome, it becomes a content exercise instead of a delivery tool. Once that is clear, the next step is choosing an objective that can actually be measured.
Set the outcome before you set the creative
I start with the outcome and work backward. A strong objective is specific, measurable, and tied to a real operational result. “Increase awareness” is too soft on its own; “increase online form completion by 15%” or “reduce avoidable calls to the contact centre by 20%” gives the team something concrete to plan around.
The simplest test I use is this: if someone outside the team cannot tell whether the campaign worked, the objective is still too vague. Good objectives answer four questions:
- Who are we trying to move?
- What action do we want them to take?
- By when should that happen?
- How will we know it worked?
That discipline matters because campaign teams often collect plenty of outputs but very little decision-quality evidence. A page view, a post, or a video impression tells me the content was delivered. It does not tell me whether people understood it, trusted it, or acted on it. Once the outcome is explicit, audience insight becomes much easier to use rather than something the team adds later to justify a creative idea.
Read the audience before you pick the channel
This is the part that most teams think they understand and usually do not. An audience is not one flat group. In government work, I often have to think in layers: the end user, the influencer, the gatekeeper, the local partner, and sometimes the staff member who has to make the process work behind the scenes. Each one may need a different message, but they still need to be pulled toward the same operational outcome.
Segmentation should be practical, not academic. I usually look at:
- Where the audience lives or works.
- How they currently interact with the service.
- What they already believe or mistrust.
- What gets in the way: time, money, language, literacy, disability, internet access, or competing priorities.
That last point matters more than teams sometimes admit. In the UK, about 1 in 4 people are disabled, so accessibility is not a niche issue; it shapes whether people can use the campaign at all. I also pay attention to emotional barriers. People may not act because they are anxious, embarrassed, overloaded, or sceptical. If the campaign does not account for that, better copy alone will not fix the problem. The audience picture then tells you which channels deserve attention and which ones are only doing superficial work.
Build the message and channel mix around the job to be done
I usually keep GOV.UK as the source of truth and use every other channel to drive people there or reinforce the action. That keeps the operational message stable while allowing different channels to do what they do best. The message itself should be short, plain, and action-led. Government guidance consistently pushes clear language, active voice, and simple sentence structures, and I think that is the right instinct: if a message needs decoding, it is too expensive for the public.
My rule of thumb is one primary ask, one supporting reason, and no clutter. If a campaign tries to ask people to do three different things, the audience usually remembers none of them. The same is true for channel choice. Different channels solve different problems, and the mix should reflect that.
| Channel | Best use | What can go wrong |
|---|---|---|
| GOV.UK pages | Authoritative instructions, step-by-step action, and service signposting | If the page is confusing, every other channel inherits the confusion |
| Email and SMS | Deadlines, reminders, and transactional updates | List quality, timing, and consent controls have to be right |
| Social media | Reach, repetition, and short prompts | Good for attention, weak for complex instructions on its own |
| Local partners and third parties | Trust, community reach, and hard-to-reach groups | Requires shared briefing and tighter coordination |
| Paid media | Scale, precise targeting, and fast awareness | Budget can disappear quickly if the landing page or service path is weak |
I also use a behavioural check here. If the desired action is hard, I ask whether the message makes it easy, attractive, social, and timely. That is a simple test, but it cuts through a lot of vague creative discussion. Once the channel mix is set, the workflow needs structure, or the campaign turns into a pile of tasks without a clear sequence.
Use the OASIS cycle to keep the work disciplined
In UK central government practice, OASIS remains the clearest way to keep campaign planning honest. It forces the team to move through five questions in order instead of jumping straight to assets and headlines. I find that useful because weak campaigns usually do not fail for one dramatic reason; they fail because several small decisions were never tested against a shared plan.
- Objective - What measurable result are we trying to achieve?
- Audience insight - What do people need, and what is stopping them?
- Strategy - What choices are we making about approach, priorities, risks, and resources?
- Implementation - How will the work be delivered, by whom, and on what timeline?
- Scoring - How will we monitor and evaluate outputs, outtakes, and outcomes?
I like OASIS because it keeps the campaign connected to delivery reality. If operations, policy, legal, and communications are not aligned, the campaign can be beautifully written and still fail. This is where public-sector leadership matters: the person running the work has to keep the team focused on one outcome, not a bundle of internal preferences. The process still falls apart if people cannot access the message, which is why accessibility needs to be built into the first draft.
Make accessibility part of the design, not a final check
Accessibility is not a finishing touch. It is how you make sure the campaign reaches the full audience, including people who would otherwise be left to work around the system. I keep the wording simple, use active voice, avoid idioms, and aim for short sentences. That is not style theatre; it lowers the cognitive load on the reader and helps the message survive on mobile screens, in a busy workplace, or under stress.
For digital and video assets, I check the basics early rather than patching them later:
- Write human-edited alt text, not auto-generated filler.
- Use captions and transcripts for video.
- Check colour contrast and never rely on colour alone to carry meaning.
- Provide accessible formats such as Easy Read, Large Print, Audio, or Braille when the audience needs them.
- Use British Sign Language where it is appropriate for the audience and the content.
- Set language correctly for multilingual content and avoid unnecessary jargon or abbreviations.
The strongest rule here is also the simplest: if the content is hard to read, hard to hear, or hard to navigate, it is not operationally ready. I see accessibility as a performance issue, not a compliance box. Once that is in place, the campaign can be tested properly instead of merely launched.
Measure what changed and learn while the campaign is live
I separate measurement into three layers because each one answers a different question. If you mix them together, reporting becomes noisy and management decisions get worse. The easiest layer to track is delivery; the hardest is behaviour. The most valuable campaigns track both.
| Layer | What it tells me | Examples |
|---|---|---|
| Outputs | What was delivered | Impressions, reach, clicks, page views, emails sent |
| Outtakes | What people understood or felt | Recall, comprehension, confidence, intention, trust |
| Outcomes | What changed in the real world | Applications completed, calls reduced, compliance improved, appointments booked |
Test-and-learn works best when the question is narrow. Split testing can compare two message variants, in-channel optimisation can improve performance within a channel, and regional or audience segmentation can show whether one group responds differently from another. In government guidance, the planning window for this kind of work is often practical and time-bound: identifying what to test can take 1 to 3 weeks, building the hypothesis 1 to 2 weeks, and developing the test plan 2 to 4 weeks. The same guidance also points to a minimum test spend of £10k and a common ceiling of 10% of total budget, depending on channel and duration. I use those figures as a sanity check, not a target; if the question is too small, a test may be wasteful, and if the question is too broad, the result will not be useful.
One thing I do not ignore is risk. Experiments can produce unexpected effects, and sometimes those effects are negative for the wider campaign. That is why learning has to be monitored, not just reported after the fact. When measurement is tight, the team learns what to keep, what to cut, and what to reuse across future campaigns. The opposite is also true, and it is where many public-sector campaigns quietly lose value.
The mistakes that quietly sink public-sector campaigns
Most weak campaigns do not collapse because the creative is ugly. They fail because the operating model is sloppy. The same mistakes keep showing up:
- Trying to speak to everyone, which usually means the message lands with no one.
- Mixing awareness, recruitment, reassurance, and compliance into one vague brief.
- Reporting reach or clicks as if they were proof of impact.
- Launching before service teams, policy teams, or local partners are ready.
- Leaving accessibility and translation work until the final review.
- Using paid media as a substitute for weak content or a broken service journey.
- Forgetting to document what was learned, so the next team starts from zero.
The deeper problem is often coordination rather than messaging. A campaign can only do so much if the underlying process is unclear or the service is not ready to absorb demand. That is why I prefer campaigns that are designed as part of the operational system, not bolted onto it. If the team gets that right, the next launch becomes easier to run and easier to trust.
The habits I would keep for the next launch
If I were briefing a new public-sector team, I would insist on five habits before anything goes live: one owner for the objective, one clear action for the audience, one source of truth for the content, one measurement model that tracks real change, and one accessibility check that is done early enough to matter. Those habits sound basic because they are basic, and that is exactly why they work.
The campaigns that last are rarely the loudest. They are the ones that respect the audience, align the organisation, and prove their value with evidence. That is the standard I would keep, because it turns communication from a one-off burst into a repeatable part of government delivery.
