What is Media plan?

A plan needs 500 sales. At a 2.5% conversion rate that is 500 ÷ 0.025 = 20,000 clicks. At a 1.2% click-through rate that is 20,000 ÷ 0.012 = 1,666,667 impressions. At a A$9 CPM the budget is 1,666,667 × 9 ÷ 1,000 = A$15,000, or A$30 a sale. Every one of those three rates is an assumption, and the plan's honesty is in writing them down where somebody can argue with them.

A media plan states what advertising is expected to produce, what it will cost, and on what assumptions.

The core of it is a short arithmetic chain run in reverse. Start from the outcome the business needs and divide back through the funnel until it reaches money.

Conversions needed, divided by the conversion rate, gives the clicks required. Clicks divided by the click-through rate gives the impressions. Impressions priced at the expected cost per thousand gives the budget.

Budget = (conversions ÷ conversion rate ÷ click-through rate) × CPM ÷ 1,000

Each step is a division by a rate, and each of those rates is an assumption.

The forecast will be wrong. That is not a failure of planning; it is what forecasts are.

What the plan provides is a written record of what was assumed, which is what makes the eventual difference informative. A plan that assumed a 2.5% conversion rate and delivered 1.8% has produced a finding. A plan with no stated assumptions has produced a disappointment.

So the discipline is to write the rates down explicitly, sourced where possible to the account's own history, segmented to match what the plan describes.

Three cautions on where the rates come from.

The account's blended average is usually wrong for a specific plan. A blended conversion rate mixes branded and non-branded, retargeting and prospecting, and a plan about prospecting in a new market should not inherit it.

Industry benchmarks describe businesses whose conditions are invisible. A published figure is not evidence about your plan.

Where no history exists, say so. A guess labelled as a guess, with a small test planned to replace it, is far more useful than a guess presented as a projection.

The same chain run in the other direction answers a different and frequently more useful question: what can this budget be expected to produce.

Businesses more often have a budget than a target, and the forward calculation turns a number somebody committed into an expectation somebody can hold. It also surfaces the mismatch early, when a target and a budget that cannot reach it are sitting in the same document.

The audience cannot supply the impressions. The arithmetic can call for more impressions than the targeting can deliver in the flight window. That constraint is checkable before the money is committed and is routinely not checked, and the symptom later is a campaign that will not spend its budget.

The target was a wish. A conversion count derived from a revenue goal rather than from what the account has achieved produces a plan that was never going to work, and the failure is attributed to execution.

The implied cost per sale exceeds the ceiling. The plan can be internally consistent and still describe activity the business cannot afford. Dividing the budget by the conversions and comparing against contribution catches it in a minute.

Frequency was ignored. Impressions divided by a small audience is repetition, not reach, and a plan that needs more impressions than its audience can absorb without heavy frequency has a fatigue problem written into it.

The flight is too short. Compressing a budget into a narrow window raises the daily spend rate, which buys further down the inventory curve and raises the price per result above what the plan assumed.

The outcome required, in the business's own terms — sales, customers, revenue, leads that close.

The chain, with each rate stated and sourced.

The budget and the flight, with the implied daily rate computed rather than left as a total.

The audience size, checked against the impressions required.

The ceiling, so that the implied cost per outcome can be compared against what the business can pay.

The split across channels or campaigns, with a reason for each share rather than a habit.

What would falsify it. The rate that, if it comes in materially below assumption, means the plan should be revised rather than persevered with.

That last item is the one most often missing and the one that turns a plan into an instrument. A plan nobody can be wrong about is a budget request with arithmetic attached.

Plans meet reality in the first fortnight, and the useful response is to compare the observed rates against the assumed ones rather than the observed outcome against the forecast.

If the click-through rate is close and the conversion rate is half what was assumed, the problem is on the site rather than in the media, and no amount of budget reallocation fixes it.

If the impression price is double the assumption, the market is more competitive than the plan believed and the budget buys less.

If the audience is delivering fewer impressions than required, the plan needs a wider audience rather than more money.

Each of those is a different action, and only the stated assumptions make it possible to tell which one applies.

Where a plan spans channels, the chain is run once per channel rather than once in total, because the rates differ enormously between them.

A search click and a feed click do not convert at the same rate, do not cost the same, and do not come from audiences at the same stage. Running one blended chain across both produces a plan that describes neither and a budget split derived from an average nobody will experience.

The practical form is a row per channel — target, conversion rate, click-through rate, price, implied budget — with the totals at the bottom. It is longer and it is the version that can be checked.

It also makes the interaction visible. Upper-funnel activity in one channel raises the branded search volume in another, so a plan that cuts the first and holds the second constant has assumed something it should say out loud.

A plan's second use is the one most accounts skip: comparing it against what happened, rate by rate, when the flight ends.

That comparison is where the assumptions get better. An account that has run four plans and recorded how each rate came in against its assumption has a set of priors grounded in its own history, which is worth more than any benchmark and takes a spreadsheet rather than a project.

Without it, every plan starts from the same guesses as the last one, and the discipline never compounds.