What an internal communications plan is, and what it isn't
An internal communications plan is a short document that answers three questions: who do we need to reach, with what, and who does it every week. Everything else is padding.
It is not a brand guidelines document. It is not a list of channels. And it is not a software project: the tool gets chosen at the end, once you already know who you have to reach, how often, and what needs a record kept of it.
A sensible length for a company of a few hundred people: six to ten pages. Past twenty, not even the department that wrote it reads it.
Section 1: the map of who never finds out
The diagnosis is not an engagement survey. It is a headcount, and you do it with the staff list in front of you.
- How many people are on each site, in each department and on each shift.
- How many of those people have a company email address. And of those, how many actually open it.
- How each group gets information today: email, the noticeboard in the changing room, the supervisor, a WhatsApp group, a sheet of paper stapled to the payslip.
- What happened the last time there was an announcement that mattered. Who found out, when, and who never found out at all.
An example of what a finished diagnosis looks like: a plant of 480 people, 60 of them in offices with a company email address they use daily, and 420 on the floor across three shifts, each with a named email account nobody opens because there is no computer at their workstation. Email covers twelve per cent of the workforce. That is a diagnosis, and it fits in two sentences.
Give it a week, and talk to the site supervisors before you talk to anyone else. They know exactly what did not land last time, and why.
Section 2: objectives you can defend
'Improve internal communication' is not an objective. An objective is a sentence with a number in it, a deadline, and somebody's name against it.
- Next month's rota seen by 80 % of the shop floor before the 25th of the month before.
- The new forklift procedure recorded as received by all 96 people who drive one, before 30 June.
- No site below 60 % reach on official announcements for two quarters running.
- Operational alerts stop going out from a supervisor's personal mobile and move to a channel where they leave a trace.
Three objectives, not seven. And at least one of the three has to be uncomfortable: the one that makes you look at your worst site instead of the company average.
Section 3: audiences, taken seriously
Segmenting is not splitting people into 'staff' and 'management'. It is recognising that the same piece of news arrives by different routes, and hours apart, depending on where somebody is standing.
| Audience | How they hear things today | What they care about |
|---|---|---|
| Offices and head office functions | Email and meetings | Results, projects, reorganisations |
| Shop floor, day shift | Handover and the changing-room noticeboard | Rota, safety, line problems |
| Shop floor, night shift | Whatever the outgoing shift tells them | The same things, always late |
| Shops and branches | A phone call or a message to the manager | Promotions, prices, floor procedures |
| New starters | Their buddy, if they have time that day | Where to clock in, who to call, which kit is theirs |
| Employee representatives | Meetings and the union noticeboard | Everything, and before everybody else |
The last row is the one most plans skip, and it is the one that decides whether the rest goes anywhere. A tool that reaches the personal phone of every person on the payroll goes to the employee representatives, or to the recognised union, before you sign it. Not after.
Section 4: what gets said where
One rule saves a lot of arguments: every type of message has one main channel, and only one. When something important can travel by four routes, it ends up travelling by whichever is easiest for the person writing it, which is almost never the one that reaches the person who has to read it.
01
The obligatory
Health and safety, payslips, policies, changes to the collective agreement. This goes through the channel that keeps a record of who received it and when. Never through a messaging group.
02
The operational
Line stoppage, rota change, a problem in one shop. This goes through the fastest channel that reaches the affected shift, and it goes segmented: if it is about the Derby plant, Manchester does not get it.
03
The institutional
Results, restructures, appointments. This goes to the employee representatives and to line managers first, then to everybody else, and only then outside the company. The order matters more than the wording.
04
The stuff that builds the habit
Long-service awards, retirements, the photo from the shutdown, who has just joined. This is what makes people open the channel on their own. Without it, the channel only ever carries bad news and gets abandoned inside a month.
A ratio that works: for every obligatory announcement, three or four of the others. It is not window dressing. It is what keeps alive the place where you later have to publish the things that matter.
Section 5: one quarter of editorial calendar
This is the section that turns the plan into work. One table, one quarter ahead, reviewed on the first Monday of every month.
| When | What | Who writes it | Who it goes to |
|---|---|---|---|
| 20th of every month | Next month's rota | Shift manager | Shop floor, by shift |
| First Tuesday | Monthly production and safety summary | Plant management | The whole site |
| As and when | A change to a safety procedure | Health and safety adviser | Only the affected roles, with a read receipt |
| Every fortnight | One piece of company news: new contract, new client, an anniversary | Comms or HR | The whole company |
| Quarterly | Short survey, five questions, always the same five | HR | By site |
Leave two slots empty on purpose every month. They fill themselves, and if you do not leave them, whatever comes up will push out whatever you had planned.
Section 6: how you will know whether it works
A plan with no measurement gets defended with anecdotes, and anecdotes last until the first budget cut.
- The percentage of the workforce that has seen each official announcement, broken down by department.
- That same figure sorted lowest first: what matters is the worst site, not the company average.
- How many people have opened the channel at least once in the last month.
- How many read-receipt documents are still outstanding, and in which roles.
- One question to the workforce every quarter, always the same one, so you can compare quarter to quarter.
What is not worth measuring: the number of posts per month, impressions, and likes. They measure how busy the department is, not whether anybody found out.
Who does this if you have no comms department
Most companies of two hundred to a thousand people do not have an internal communications department. They have somebody in HR who does this as well as their job.
That works, on two conditions. That the time is written down and protected, two hours a week rather than 'when I can', and that the person is allowed to publish without collecting three signatures first.
- One named owner.
- A deputy, for the summer holidays and for sick leave.
- One correspondent per site, usually the supervisor: they are the ones with the photos and the news.
- A short approval route, written down: who reviews what, and within how many hours.
What usually sinks a plan is not the budget. It is Tuesday's announcement waiting for sign-off from a director who is away until Thursday.
The five mistakes that keep repeating
01
Starting with the tool
Somebody buys a platform and then goes looking for things to put in it. The right order is the other way round, and it makes the purchase cheaper too: once you know what you actually need, half the modules on the quote turn out to be spare.
02
Launching with a welcome to the new tool
The first post has to be something people want to read: the rota, who is working over Christmas, the photo from the shutdown. Nobody opens the introduction to the new intranet.
03
Opening every site at once
If the rollout goes wrong on one site, you fix it in a week. If it goes wrong on twelve, the conclusion upstairs will be that this does not work.
04
Presenting the average
An average reach of 68 % can hide one site at 95 % and another at 20 %. The average reassures the board and fixes nothing.
05
Going quiet over the summer
The habit dies in three weeks and takes a quarter to rebuild. Leave two posts written and ready before the holidays start.