Essay
Most marketing organisations are extremely good at a job that stopped creating advantage some time ago. They plan, produce and measure campaigns with real craft, hit every number they were given, and close the year with the company’s underlying position exactly where it started.
The problem sits in the structure. A campaign is a monologue with a start date and an end date. It rents attention, converts a slice of it, and then stops existing. Nothing it built stays on the balance sheet the following morning.
The test
One question separates the two modes of work: if we stopped spending tomorrow, what would keep working?
A product people prefer keeps working. A community that talks to itself keeps working. A distribution advantage keeps working. A reputation for solving one specific problem better than anyone else keeps working. A finished campaign leaves nothing running, however well it performed while it was live.
Campaigns borrow attention. Businesses own demand.
That is the reasoning behind the doctrine on this site: build businesses, not just campaigns. Campaigns are not worthless, but they are the smallest unit of value in a discipline that keeps getting asked to deliver much bigger ones.
From message to reason
The shift is from designing what you will say to designing why someone would choose you voluntarily. Those are different jobs requiring different skills. The first needs creative and media capability. The second needs an understanding of behaviour, a view on the business model, and the authority to change the product, the price or the experience when the evidence calls for it.
Which is exactly why this transition tends to stall in the organisational chart rather than in a strategy document. Marketing gets asked to create demand while holding no control over anything that actually creates it. The fix is not a better brief; it is giving the demand-creation function a seat where product, price and distribution actually get decided.
- Are we measured on activity delivered or on demand created?
- Can we change the product when the behaviour tells us to?
- What did we build this year that will still be earning in three years?
Building inside a company that already exists
Corporate builders inherit two advantages that startups pay dearly for: distribution and reputation. They also inherit a constraint startups never face — everything they launch reads as a statement about the parent brand. Ignore that and a promising internal venture gets quietly shut down after a strong first year, for reasons nobody quite writes down.
Starting something new inside an established business is usually filed as a portfolio problem. In practice it is a trust problem. The new venture has no track record, fights the proven core for resources, and inconveniences people who are measured on something else entirely.
So the first milestone is never commercial. It is organisational: earning enough internal credibility to be allowed to keep going. Teams that get this design their first year around producing evidence and allies rather than a launch. Teams that miss it produce a great launch, and then discover the organisation was never actually behind them.
It also changes the calendar. Campaign organisations plan in bursts because their unit of work has a start and an end. Business builders plan in cohorts and in versions: what did we learn from the people who arrived last quarter, what does that change in the product, what does that do to retention and to price. The rhythm looks slower from the outside and runs considerably faster underneath, because each cycle keeps the value the previous one already earned.
The objection: but campaigns are measurable and businesses aren’t
The strongest argument against this whole doctrine is not that it is wrong — it is that it is inconvenient to report on. A campaign produces a number by Friday: reach, clicks, conversion, cost per acquisition. Building something people choose produces something far harder to fit in a weekly deck, because the return shows up as retention, referral and pricing power, months after the work that caused it. Organisations built around quarterly reporting will keep drifting back toward the thing that is easy to measure, even after agreeing, in principle, that it is the smaller prize.
The honest answer is not to pretend the measurement problem does not exist. It is to measure the right things badly rather than the wrong things well. A cohort’s month-six retention, a repeat-purchase rate, the share of new customers arriving through referral instead of paid media — these are slower, noisier signals than a click-through rate, but they tell you whether you built something or just rented attention for a while. A leadership team that will not tolerate a slower, noisier metric has already chosen campaigns, whatever the strategy document claims.
What compounds and what decays
It helps to be precise about the mechanism, because ‘building versus campaigning’ can sound like a taste preference rather than an argument. A campaign’s effect decays from the day it stops running — the curve is well understood, and it is the entire reason media plans exist. A product improvement, a piece of earned reputation or a distribution relationship does the opposite: it keeps producing value with no further spend, and often compounds, because satisfied customers bring others at zero marginal cost.
Put both curves on the same chart over three years and the campaign looks better in month one and worse in every month after. Most marketing budgets are still allocated as though month one were the only month that mattered, because the people approving the budget are evaluated on a cycle that ends before the compounding curve overtakes the decaying one.
The failure mode when this is understood too late
The costliest version of this mistake isn’t the company that never tries to build demand instead of renting it. It’s the company that tries, builds something genuinely better, and then reverts to campaign thinking the moment growth slows, because campaigns are the muscle everyone in the building already has. The product stops improving in response to what customers are actually doing, the team that understood the behavioural signal gets reassigned to a launch, and the business slides back into monologue while insisting internally that it still believes in the dialogue.
The tell is almost always the same: budget gets cut from product and research first, because those costs are visible and their absence is not felt immediately, while campaign spend survives because a quarter without it produces an uncomfortable, obvious gap in the numbers. By the time the underlying preference erodes, the campaign spend required to compensate for it has quietly doubled, and nobody can point to the week the trade got made.
- Which of last year’s cuts fell on the things that compound rather than the things that decay?
- If growth slowed tomorrow, would the instinct in this building be to fix the product or launch a campaign?
- Who in this organization is rewarded for retention and referral rather than for reach?
The talent this actually requires
None of this works with a marketing team hired and trained to produce campaigns. Identifying why customers actually choose you, arguing for a product change on the strength of that evidence, and staying accountable to a retention number rather than an impressions number is a different discipline — closer to product management with communications fluency than to traditional brand marketing. Organisations that decide to build rather than campaign and then staff the effort with the same team, the same incentives and the same quarterly rhythm as before have not made the shift. They have just renamed the campaign.
That has a direct implication for how these people are managed. They need a mandate that survives a bad quarter, because the compounding curve described above is, by its nature, slower to show results than the decaying one it replaces. A team measured with campaign patience on a building timeline gets disbanded exactly when the work was about to start paying off.
The commercial argument for protecting that mandate is straightforward, even if the emotional pull toward cutting it is strong. A campaign team disbanded mid-flight has simply stopped; nothing was owed to the previous quarter’s spend. A building team disbanded mid-flight destroys the very asset that was supposed to justify the patience in the first place — half-built preference, half-shipped product improvements and half-earned reputation do not compound, they just stop, and the next attempt starts from closer to zero than anyone in the room remembers agreeing to.
What people actually choose
People choose things that make their life easier, make them feel competent, or connect them to others. That is unglamorous and it has held for decades. What changed is that they now have infinite alternatives and near-zero switching cost, so tolerance for anything merely adequate has collapsed.
In that market, communication cannot make up for a product nobody prefers. It never really could. It used to just take longer for that to become obvious.
So the ambition has to move up: not a better campaign about the thing, but a thing worth choosing — and only then the campaign. In that order, the marketing work gets considerably harder and considerably more valuable, and it finally compounds into something the company actually owns.