Ask a room of marketers what an integrated campaign is and most will describe the artwork: the same image, the same colours, the same slogan, appearing everywhere at once. That is not integration. It is repetition with a shared palette. An IMC campaign earns the name when something stricter is true — when every channel the audience meets is carrying the same claim, stated in the form that channel handles best, so that a person who encounters only one piece of the campaign still receives the whole argument, and a person who encounters all of them never receives two different ones.

The claim is the unit of integration. The press release states it as fact, with dates and evidence attached. The advertisement dramatises it, compressing it into a scene the audience can feel. The social conversation tests it, in public, where weak claims are found out quickly. The owned pages document it, holding the detail the other channels point back to. Integration is therefore an editorial discipline before it is a media plan: the plan decides where things appear, but the editorial work decides whether they agree. This article sets out how that agreement is built, and the ways it quietly falls apart.

What Integration Actually Means

Integration operates at the level of the claim, not the execution. Two executions can look nothing alike and be perfectly integrated, because both assert the same thing and both can prove it. Two executions can be visually identical and contradict each other completely — the advertisement promising a transformation the press release, written by a more careful hand, described only as an improvement in one measured respect. The public does not experience a campaign as departments. It experiences a company saying things, and it keeps a rough, unforgiving account of whether the things match.

This is why the useful test of an IMC campaign is subtraction. Remove every channel but one, and ask what a stranger would conclude the company is claiming. If the advertisement alone, the release alone and the social feed alone each produce the same sentence, the campaign is integrated. If each produces a different sentence — the advertiser’s ambition, the press officer’s caution, the social team’s joke — the campaign is three campaigns sharing a logo, and the audience will believe whichever version it met last, or whichever a critic chooses to set against the others.

None of this requires the channels to use the same words. Register differs by channel and should: a release that read like an advertisement would fail as a record, and an advertisement written like a release would fail as an advertisement. What must not differ is the substance — the facts relied on, the size of the promise, the conditions attached. Integration is consistency of substance under variety of form, and holding that line, across teams and weeks and temptations, is the actual work.

The Building Blocks

Integration rests on three pieces of work done before any channel is booked: a message architecture that fixes the claim, channel roles assigned by honesty about what each channel does well, and a sequence that puts the evidence in public before the amplification begins. Skip one and the campaign may still look busy; it will not hold together.

One Message Architecture

The architecture begins with a single sentence: the claim every channel must be able to carry. The sentence has to be specific enough to be checkable — “the dependable choice” is not a claim, it is a mood — and modest enough to survive contact with the evidence. Underneath it sit the proof points, ranked: the two or three facts, demonstrations or records that make the claim believable, in the order of their strength. Every channel works from the same page. The advertisement may lead with the second proof point because it dramatises best; the release leads with the first because it is the most verifiable. The ranking can flex. The set cannot grow in one channel and shrink in another.

The architecture also records what the campaign is not claiming. This negative space does more integration work than any brand guideline: it is the list a copywriter checks before writing the line that goes one step further than the evidence, and the list a community manager checks before answering a challenge with a promise the company has not made. Most public contradictions are not lies; they are local enthusiasms nobody checked against the centre. A written architecture, with its exclusions stated, is how the centre holds.

Channel Roles, Assigned Honestly

Channels earn their place by being assigned the work they are genuinely good at, not the work the org chart happens to own. The standard map of the territories — paid, earned, shared and owned media, set out in the shared media article in this series — is the starting point. Paid media buys controlled repetition: the claim compressed and placed, exactly as written, in front of chosen audiences. Earned media supplies independent judgement: coverage is the claim tested by someone with no reason to pass it. Shared media is the audience’s verdict, distribution performed by people for their own reasons. Owned media is the archive: the pages where the full evidence lives and stays.

Assigning roles honestly means accepting each territory’s limits as part of the assignment. Paid media cannot supply credibility, however often it repeats the claim. Earned media cannot be scheduled, however carefully the story is prepared. Shared media cannot be commanded, only earned sideways by material people judge worth passing on. Owned media cannot create an audience on its own. A plan that asks a channel to do what its territory cannot do — the advertisement expected to make the claim believed, the release expected to make it loved — has misassigned the work.

Sequencing: What Must Be True First

Integration has an order, and the order is not cosmetic. The governing rule is that amplification comes after verification: the claim should be true, public and checkable before the paid layer spends money repeating it. In practice that means the record is established first — the announcement made, the evidence published on owned pages where anyone can inspect it — and only then do the dramatisation and the conversation scale up. An advertisement that runs ahead of the announcement it refers to sends its audience to a record that does not yet exist; a social push that precedes the facts invites the audience to fill the gap with speculation the campaign will spend its second week correcting.

Sequencing also protects the campaign against its own news. Announcements attract questions the advertisement never anticipated; if the evidence and the record are already public, those questions have somewhere accurate to land. Where the sequence must compress — a launch date fixed by events outside the campaign’s control — the discipline survives in miniature: record and amplification may land on the same morning, but the record lands first, and nothing is amplified that the record does not support.

Where the Press Release Fits

Inside an integrated campaign, the press release plays one role and plays it soberly: it is the campaign’s verifiable record. It is the dated, attributed, checkable statement of what the company is announcing — the document the advertisement’s dramatisation must remain consistent with, the link the social conversation can point back to, the text a journalist works from. Other channels may be more vivid; the release is the one that can be quoted against the company later, which is precisely why it disciplines all the others. A claim that cannot survive being written into the release, with its facts and conditions attached, does not belong in the advertisement either.

The craft of the launch announcement itself — the release that opens a campaign, its structure, timing and targeting — belongs to the campaign announcement discipline covered elsewhere in this series, and this article leaves that ground to it. What belongs here is the release’s position in the integration: written from the same message architecture as everything else, issued before the amplification it supports, and maintained afterwards. Campaigns run for weeks; facts occasionally change mid-flight. When one does, the record is corrected openly and the other channels follow the correction. An advertisement still running with the superseded claim is not a scheduling oversight; it is the campaign contradicting itself in public, at paid rates.

The Campaign Archetypes

Most integrated campaigns are variations on a small set of recurring shapes. The archetypes below are described generically — how each works, what integration demands of it, and where each typically breaks. The details are illustrative, not reportage.

The Launch

The launch introduces something new: a product, a service, a company. Its claim is existence plus significance — this exists now, and here is why it matters — and its integration challenge is sequencing under time pressure, because launch dates are usually fixed by operations rather than by communications. At its best it is a staged reveal: the record established with the announcement, the evidence published for inspection, the dramatisation following once the facts can bear the attention. Its characteristic failure is the reverse order — a teaser campaign that builds curiosity about claims nobody has yet had to state precisely, so that by launch day the advertisement has promised, by implication, more than the announcement can deliver. The gap between the teased version and the real one becomes the story.

The Repositioning

The repositioning campaign argues that what the audience believes about the company is out of date. Its claim is a correction, which makes it the most evidence-hungry of the archetypes: the audience is being asked to replace a belief it already holds, and assertion alone will not do it. Integration here means the proof points lead everywhere — the owned archive rebuilt so the new claim is documented, earned coverage sought on the strength of the evidence, advertising restrained to what the record supports. Its characteristic failure is cosmetic: a new look and a new slogan laid over an unchanged record, so every channel says the new sentence while every checkable fact still supports the old one. A repositioning can only announce a change that has actually happened.

The Seasonal Campaign

The seasonal campaign borrows a moment the audience is already living — a holiday, a season, a returning event — and attaches the company’s claim to it. Its advantage is relevance supplied by the calendar; its danger is that the moment does the creative work and the claim becomes decoration. Integration demands the same discipline as any other archetype, applied quickly: the seasonal execution must still carry the central sentence and stay within its proof points, and the record — prices, dates, availability, terms — must be exact, because seasonal campaigns make dated promises that expire in public. Its characteristic failure is disconnection: a charming seasonal film that could belong to any company in the category, integrated with nothing, remembered — if at all — without the brand attached.

The Cause-Linked Campaign

The cause-linked campaign connects the company to a social or environmental cause. It is the archetype where integration is most severely tested, because the claim is not about a product but about the company’s conduct — and conduct is checked against the record, not the advertisement. The governing rule is the one the CSR communication discipline states: the programme must exist before it is announced, described at the level the evidence supports, with amounts stated as spent rather than pledged and activities described as done rather than intended. An IMC campaign built on a cause the company supports only inside the campaign is not integration; it is a costume, and the audience’s shared channels — where the discrepancy will be documented — are part of the campaign whether the plan includes them or not. Run honestly, with the programme real and the claim sized to it, the archetype integrates powerfully precisely because every channel is telling a story that checks out.

The Always-On Programme That Is Not a Campaign

Finally, a boundary worth drawing. Much good integrated work is not a campaign at all: the standing programme — the maintained record, the regular coverage, the ongoing conversation — that keeps a company’s claim visible between campaigns. It deserves the same message architecture and the same honesty, but it should not borrow the campaign’s vocabulary or its measurement habits. A campaign has a defined claim, a beginning and an end, and a change it intends; a programme has continuity. Calling the programme a campaign flatters it in reports and corrupts both: the programme gets judged on timescales it was never built for, and real campaigns lose the discipline of an ending. Know which of the two you are running.

Planning Honestly

An honest campaign brief names two things most briefs avoid: the audience’s current belief, stated unflatteringly and in the audience’s own terms, and the change the campaign intends — a different belief, a behaviour, a consideration that did not exist before. The change is written as an intention, not a promised outcome, because no plan can promise what an audience will conclude; the moment a brief writes the intended change as a deliverable, every downstream judgement is corrupted, and measurement becomes an exercise in finding the promised number. The brief also fixes the message architecture, the channel roles and the sequence, and names one owner for the whole — the person whose job is the agreement between the parts, with authority to stop a channel that has drifted from the claim.

The planning process has one meeting whose product matters more than any document: the integration review, where every channel’s near-final material is laid side by side and read as the audience will read it — as one company’s voice. Its real output is a list: the contradictions caught before the public catches them. The advertisement’s implication the release does not support. The social copy answering a question with a claim from a different campaign. The landing page whose figures are last quarter’s. Campaigns that survive contact with the public have had this meeting, formally or otherwise; campaigns that fall apart in public are discovered, on inspection, to have skipped it.

Measurement Without Fiction

Integrated campaigns attract a characteristic measurement fraud: the summing of channel metrics into a single impressive total. Reach from the paid report, impressions from the social platform, circulation from the coverage, visits from the owned pages — added together into one figure that counts the same people repeatedly, mixes estimates with counts, and presents the result as the campaign’s audience. It is fiction with arithmetic on it. The honest practice reports each channel in its own terms, each figure labelled as what it actually counts, and resists the total. A reader asking whether the campaign worked is not served by a bigger number; they are served by knowing which parts can be verified and which are estimates wearing a uniform.

The campaign-level question is different, and better: did the claim travel accurately? Was the sentence the audience repeated — in coverage, in shares, in the questions the sales team started hearing — the sentence the campaign set out to establish, with its conditions still attached? That judgement belongs to the evaluation discipline covered elsewhere in this series: objectives stated in advance, indicators agreed before the work begins, results reported with their limits visible. Attribution, finally, is claimed modestly or not at all. An integrated campaign is precisely the setting where single-channel credit is least knowable — the channels were designed to work together, so their effects arrive together — and a report that assigns the outcome neatly to one channel is describing the reporting tool, not the campaign.

Integration Failures

The failures recur often enough to name. The commonest is structural: the channels were briefed by different teams from different documents, so the campaign was never integrated at the source — the advertising agency working from a creative brief, the press office from a news plan, the social team from a content calendar, each internally coherent and mutually inconsistent. A close relative is the improvised second message: mid-campaign, a channel team under pressure invents a sharper claim than the architecture allows, and for a week the company runs two campaigns, the official one and the popular one, until a journalist or a critic places them side by side. A third is the integrated launch date with unintegrated facts: every channel live at the same hour, each carrying a different figure, date or name, because the choreography was rehearsed and the proofreading was not.

And there is the failure that looks most like success: the campaign whose integration was a colour palette. Every execution matches visually, the toolkit was followed, the toolkit was beautiful — and the claim was never fixed, so each channel is saying something different in the same shade of blue. It passes internal review because internal review checks what can be screenshotted. It fails in the only place that matters, which is the audience’s reconstruction of what the company said. The remedy is the subtraction test applied before launch, by someone with the authority to delay it: one channel at a time, one sentence each, and no launch until the sentences agree.

The Discipline Underneath

An IMC campaign is, in the end, a promise-keeping exercise conducted in several media at once. The claim is the promise. The proof points are its evidence. The sequence, the architecture and the integration review exist so that the promise is made nowhere more strongly than the evidence allows, and kept identically everywhere it is made. Done that way, integration stops being a production technique and becomes what it actually is in the durable campaigns: a form of respect for the audience’s intelligence, which assumes people will meet the fragments, compare them, and draw conclusions from the comparison. They will. The only choice a campaign really makes is what the comparison will show.