What Is Media Planning? The Process, the Maths and the Proof
The full planning process, the reach and frequency maths behind it, and the measurement stage most guides leave out.
Definition
Media planning is the process of deciding which media channels a brand should buy, how much to spend in each, and when to run, in order to reach a defined audience against a business objective. It is distinct from media buying, which executes the plan, and it is only as good as the measurement evidence behind the allocation.
Media planning is the process of deciding which advertising channels a brand will buy, how much budget each one receives, when the ads run and how often the audience should see them. It turns a commercial objective into a costed schedule of channels, flight dates, reach and frequency goals, and the measurement method that will later prove whether the plan worked.
One clarification before anything else, because search results for this topic are crowded with the wrong subject. Media planning in advertising is the allocation of paid media budget across channels such as television, connected TV, online video, paid social, search, retail media, radio, print and out of home. It is not the business of scheduling organic posts on a social account. The two disciplines share a word and almost nothing else.
What is media planning in marketing, and who does it
Media planning sits between strategy and execution. Strategy decides what the brand is trying to achieve and what it will say. Media planning decides where that message is placed, in what volume, at what weight and over what period, then sets the standard by which the placement will be judged.
The people doing it fall into three groups:
- Agency media planners, who build plans for client brands and hand them to a buying team.
- In-house media or performance leads at brands, who own the budget and defend it internally.
- Marketing effectiveness and analytics teams, who supply the evidence the plan rests on and who audit it afterwards.
The third group matters more than most guides to media planning admit. A plan is a set of predictions about how channels will behave. If nobody checks those predictions against outcomes, the plan is a preference dressed as a forecast.
What a media plan actually contains
A media plan is a document, and a good one is specific enough that a buyer could execute it and an analyst could grade it. Most published examples stop at channel and budget. A plan that can be defended to a finance director carries a measurement method on every line.
Here is what that looks like for a hypothetical $1.2m quarterly plan for a mid-market consumer brand.
| Line item | Role in the plan | Budget | Share | Flight | Reach and frequency goal | How it will be measured |
|---|---|---|---|---|---|---|
| Connected TV | Reach growth in the core audience | $420,000 | 35% | Weeks 1 to 10, continuous | 45% reach, average frequency 4 | MMM contribution plus geo holdout in 12 markets |
| Online video | Extend CTV reach into light TV viewers | $180,000 | 15% | Weeks 1 to 10, continuous | Incremental reach of 12 points | Reach and frequency overlap study, MMM contribution |
| Paid social | Mid-funnel consideration and creative testing | $240,000 | 20% | Weeks 2 to 12, pulsed | 8 exposures per reached user over the flight | Platform conversion lift test, MMM contribution |
| Paid search, non-brand | Capture demand created upstream | $180,000 | 15% | Always on | Impression share 65% on priority terms | Incrementality test on a matched market pair |
| Retail media | Convert at the point of purchase | $120,000 | 10% | Weeks 4 to 12 | Category share of voice 20% | Retailer sales lift plus MMM contribution |
| Out of home | Regional presence in two launch cities | $60,000 | 5% | Weeks 3 to 8, flighted | 60% cover of the city population | Geo lift against matched control cities |
Two things about that table are unusual, and both are deliberate. Every line names a measurement method rather than promising to monitor performance. And the methods are not all the same, because no single method reads every channel well.
Media planning vs media buying
These are separate jobs, often done by separate people, and confusing them is the most common error in briefs. Media planning decides what to buy and why. Media buying decides what to pay and negotiates the terms.
| Media planning | Media buying | |
|---|---|---|
| Core question | Which channels, at what weight, when | At what price, on what terms, from whom |
| Primary output | Channel mix, budget split, flight plan, reach and frequency goals | Insertion orders, programmatic line items, negotiated rates and make-goods |
| Core skills | Audience research, forecasting, econometrics, budget modelling | Negotiation, platform operations, inventory and yield knowledge |
| Time horizon | Quarterly or annual, revisited monthly | Daily and weekly |
| Success looks like | Budget sitting in the channels that generate incremental return | Buying the planned impressions at or below the planned price |
| Characteristic failure | Correct price, wrong channel | Correct channel, overpaid |
The distinction carries real financial weight on the buying side. The Association of National Advertisers found that only about one third of every programmatic dollar reaches the end user, in a study covering 21 brands and roughly $123 million of spend, as reported by Adweek. A plan that is right about channel and wrong about supply path leaks most of its value before an impression is served.
Types of media planning
There are two useful ways to cut this, and planners use both.
By ownership
- Paid media planning. Bought placements: television, connected TV, online video, paid social, search, display, audio, retail media, print and out of home. This is what people normally mean by media planning.
- Owned media planning. Sequencing of channels the brand controls, such as email, the website, the app and the existing customer base.
- Earned media planning. Public relations, influencer relationships and organic amplification, planned for timing rather than bought outright.
By channel character
- Traditional media planning. Linear TV, radio, press and out of home. Bought on audience estimates and modelled reach, not user-level tracking.
- Digital media planning. Search, social, programmatic display and video, bought on auction mechanics with granular in-platform reporting.
- Integrated cross-channel planning. The mix treated as one system, where each channel is credited for its effect on the whole rather than on its own reported conversions.
- Performance-weighted planning. Budget allocated against a cost-per-acquisition or return-on-ad-spend target, usually over a short horizon.
- Brand-building planning. Budget allocated against reach and mental availability over a longer horizon, where the return arrives late.
The integrated view is the one most organisations claim and least often practise. Nielsen's 2025 Annual Marketing Report found that only 32% of marketers globally say they measure their media spending holistically across both digital and traditional channels, falling to 23% in Europe. The report is based on 1,400 global marketing professionals surveyed in early 2025. If two thirds of the market cannot see the whole mix, two thirds of media plans are being built on a partial picture.
The media planning process, step by step
The steps below are the common sequence, with one addition that most published versions of the media planning process leave out.
- Define the commercial objective, not the media objective. "Grow category penetration by three points" is an objective. "Achieve 200 GRPs" is a tactic, and treating it as the goal is how plans drift.
- Define and size the audience. Who buys, who could buy, how many of them exist and what they consume. Sizing matters as much as profiling, because it sets the ceiling on achievable reach.
- Audit the evidence you already hold. Prior model results, past experiments, last year's plan and what it actually delivered. Skipping this is how organisations rediscover the same finding every eighteen months.
- Select channels against a stated criterion. Write down why each channel is in the plan. If the reason is "we always run it", that is a finding, not a rationale.
- Set reach and frequency goals per channel. These are the levers that make a plan checkable rather than aspirational.
- Allocate budget across the mix. Covered in detail below, because this step determines most of the outcome.
- Build the schedule. Continuity, flighting or pulsing, mapped against seasonality, competitor activity and production timings.
- Measure, then feed the result back into the next plan. This is where the process either becomes a loop or stays a straight line ending in a slide deck.
Related product
Aryma Nebula
An agentic spend optimizer that turns MMM outputs into weekly and monthly budget allocation decisions.
The maths of media planning: reach, frequency, GRPs and CPM
Almost every commercial guide names these metrics and works through none of them. They are simple, and knowing them is the difference between reading a plan and interrogating one.
The standard definitions come from the advertising literature, principally Belch and Belch, Advertising and Promotion. Reach is the number of individuals or households exposed to the message at least once over a given period. Frequency is the average number of times those reached are exposed. Average frequency is derived by dividing gross rating points by reach. Cost per mille answers what it costs to reach a thousand people.
Take a single campaign against a target population of 4,000,000 adults, with a media budget of $72,000.
| Metric | Formula | Worked example |
|---|---|---|
| Reach | Unique people exposed divided by target population | 1,200,000 of 4,000,000 = 30% |
| Impressions | Total exposures delivered | 4,800,000 |
| Average frequency | Impressions divided by unique people reached | 4,800,000 divided by 1,200,000 = 4.0 |
| GRPs | Reach percentage multiplied by average frequency | 30 multiplied by 4.0 = 120 GRPs |
| CPM | Cost divided by impressions, multiplied by 1,000 | ($72,000 divided by 4,800,000) times 1,000 = $15.00 |
| Cost per point | Cost divided by GRPs | $72,000 divided by 120 = $600 per GRP |
Two practical consequences follow. First, GRPs are a volume measure, not a quality measure. A plan delivering 120 GRPs as 30% reach at four exposures is a different campaign from 12% reach at ten exposures, even though the headline number is identical.
Second, effective frequency, the number of exposures below which the message does not register, decides which of those two shapes you want. If effective frequency in the category is three, the first plan works and the second wastes most of its weight on people already convinced.
The reach question is not academic. Nielsen's ROI report found that only 63% of ads across desktop and mobile in the United States reach their intended age and gender targets. Planned reach and delivered reach are different numbers, and only one of them appears on the plan.
How to allocate budget across the media mix
This is the step where media planning either creates value or quietly destroys it, and it is the step most published guides hand-wave. Six methods are in common use, and they are not equally defensible.
- Percentage of sales. Set the budget as a fixed share of revenue. Simple, stable and backwards: it makes spend a consequence of past sales rather than a cause of future ones.
- Competitive parity or share of voice. Match spend to a target share of category advertising. Useful as a sanity check, weak as a primary method, because it outsources your budget logic to competitors who may be wrong.
- Objective and task. Cost out the reach and frequency needed to hit a stated objective, then sum it. Intellectually honest and the standard textbook answer, but only as good as the assumed response rates.
- Historical plus or minus. Last year's plan with an adjustment. Fast, and the dominant method in practice. It preserves every error in the previous allocation.
- Response curve allocation. Model each channel's diminishing return and move budget until the marginal return is equal across channels. This is the economically correct answer.
- Model-derived optimal allocation. The same principle, but with the response curves estimated from data through marketing mix modeling rather than assumed.
The cost of getting this wrong is measurable. Nielsen's ROI report describes a "50-50-50 gap": 50% of media plans are underinvested by a median of 50%, and brands can improve ROI by 50% with optimal budgeting. The same report found social media delivering 1.7 times the ROI of television while receiving less than one third of television's budget. Those are not rounding errors. They are the difference between a plan that pays for itself and one that does not.
Two constraints keep the theoretical optimum from being the answer you can actually execute. Response curves saturate, so past a certain point additional spend in a channel buys progressively less. And real plans carry commitments: upfront television deals, minimum spends, contracted retail media, agency fee structures and creative that only exists in certain formats. An allocation that ignores those produces a recommendation nobody can act on. Turning modelled curves into a weekly or monthly allocation that respects the constraints is exactly the job Aryma Nebula exists to do.
Media scheduling: continuity, flighting and pulsing
Every definition of media planning includes deciding when to advertise, and then almost every guide omits the three canonical patterns for doing it.
- Continuity. Even weight across the whole period. Suits categories with steady, year-round demand and protects brand memory, but spreads budget thin.
- Flighting. Concentrated bursts with dark periods in between. Suits seasonal categories and launches, and buys real weight in short windows, at the cost of decay between flights.
- Pulsing. A continuous base layer with bursts on top. The compromise most large advertisers use, and usually the safest default when demand is year-round but uneven.
Dark periods are not free. Nielsen has found that a brand loses an average of 2% of future revenue for every quarter it stops advertising, and that a one-point gain in brand metrics can drive a 1% increase in sales. That figure is the strongest available argument for pulsing over hard flighting when a budget cut is on the table.
Why attribution quietly distorts media plans
Most media plans are built on numbers that systematically favour the channels closest to the click. Three forces cause it.
Signal loss and consent restrictions have shrunk the deterministic user-level data that last-click attribution depends on, and the gaps get filled with modelled estimates that differ by platform. Walled gardens grade their own homework, reporting conversions against their own attribution windows, so the same conversion can appear in three dashboards at once. And last-click models credit the final touchpoint, which structurally underprices everything upstream: television, connected TV, online video, out of home and audio.
The planning consequence is predictable. Upper-funnel channels look expensive, budget migrates down the funnel, short-term efficiency metrics improve, and total demand slowly shrinks because nothing is refilling the top. The plan looks better every quarter while the business gets harder.
This is why the measurement method belongs on the plan itself, line by line, before money is committed. It is also why the 32% holistic-measurement figure from Nielsen matters so much. A channel that no method in your stack can read fairly will lose every budget argument regardless of its actual contribution.
Closing the loop: how to prove a media plan worked
The honest test of a media plan is not whether the campaign ran to schedule. It is whether the business would have been worse off without it. Three methods answer that, and they work best together.
Marketing mix modeling uses aggregate time-series data to estimate each channel's contribution to sales, including channels with no click and no user-level data. It reads the whole mix on one basis, which is precisely what last-click cannot do, and it produces the saturation curves budget allocation needs. Marketer demand reflects that: an eMarketer survey conducted with TransUnion found that 46.9% of US brand and agency marketers plan to invest in MMM over the next year, and 27.6% named it the most reliable measurement methodology available.
Geo experiments and holdout tests withhold or increase spend in matched markets and measure the difference. This produces a causal read on one channel at a time, which is narrow but very hard to argue with.
Calibration joins the two. Experiment results are used to constrain the model, so the MMM is anchored to at least one channel where the true incremental effect is known. Google's open-source MMM, Meridian, is explicit about this direction. Its documentation describes modelling reach and frequency to make video measurement more actionable, providing budget optimisation, and a forthcoming GeoX component designed to combine MMM with incrementality by calibrating models with experiments across channels.
For a planner, the practical output of all this is not a report. It is three numbers per channel that go straight into the next plan: incremental return, the point at which returns start diminishing, and the confidence interval around both. Getting from a finished model to those three numbers, in a form a planning meeting can use, is the work MMM Singularity and Aryma Nebula automate.
Media planning for B2B and considered purchases
Nearly every worked example in the published literature is consumer packaged goods or retail. B2B and high-consideration categories break several assumptions at once.
- The addressable audience is small, sometimes a few thousand accounts, so reach ceilings arrive quickly and frequency does most of the work.
- Sales cycles run for months or quarters, so the lag between exposure and revenue is longer than most attribution windows and longer than most reporting cadences.
- Buying is committee-based, so the person reached is often not the person who converts, and single-person attribution is misleading by construction.
- Volumes are low enough that conversion data is noisy, which weakens experiments and makes prior evidence more valuable.
The practical adjustments are to plan reach against accounts rather than individuals, to set frequency goals deliberately high within a tightly defined audience, to measure against pipeline creation rather than closed revenue, and to lengthen the modelling window so lagged effects become visible at all.
Where AI actually changes media planning
The AI claims in this category are mostly unspecified. It is worth being concrete about what changes and what does not.
What changes is speed and cadence. Scenario simulation that once took an analyst a week can be run across dozens of budget splits in minutes. Budget optimisers can respect real constraints such as minimum spends and contracted commitments rather than producing an unusable theoretical optimum. Model results can be refreshed and reallocated weekly or monthly instead of being read once a quarter. And the interpretation layer, turning coefficients into a sentence a planning meeting can act on, stops being a bottleneck.
What does not change is the statistical core. Variable selection, functional form, adstock and saturation specification, and causal validity checks stay human-led, because that is where the credibility of the whole exercise lives. We call this split Peripheral Agentic MMM: agents around the model, rigour inside it.
The market is moving this way. The IAB's 2026 Outlook Study, based on insights from more than 200 brand and agency buyers, found advertiser focus on cross-platform measurement rose to 72% from 64% year over year, with two thirds now focused on agentic AI for ad buying and campaign execution. The same study forecasts 9.5% growth in US ad spend for 2026, with social media up 14.6%, connected TV up 13.8% and linear television down 1.7%. A mix shifting that fast makes annual planning cycles look slow.
Frequently asked questions
What is the difference between media planning and media buying?
Media planning decides which channels to use, how much budget each receives, when campaigns run and what reach and frequency to target. Media buying then negotiates and executes those placements, securing inventory at the best available price and terms. Planning answers where and why; buying answers how much and from whom. In smaller organisations one person does both, but they remain distinct skills with distinct failure modes.
What are the 5 M's of media planning?
The 5 M's are Mission, Money, Message, Media and Measurement. Mission is the objective the advertising must achieve. Money is the budget available and how it is set. Message is what the creative says. Media is the mix of channels and vehicles chosen. Measurement is how success will be judged. It works well as a plan review checklist, because a plan missing any one of the five is usually missing the argument for itself.
How do you calculate GRPs and CPM?
Gross rating points equal reach percentage multiplied by average frequency. A campaign reaching 30% of the target audience an average of four times delivers 120 GRPs. Average frequency is GRPs divided by reach. Cost per mille is media cost divided by impressions delivered, multiplied by 1,000. A $72,000 buy delivering 4,800,000 impressions has a CPM of $15.00 and, at 120 GRPs, a cost per point of $600.
How do you measure the success of a media plan?
Start by rejecting delivery metrics as proof. Impressions served and budget spent show the plan ran, not that it worked. Credible measurement combines marketing mix modeling, which reads every channel on one basis including offline media, with geo holdout or incrementality experiments that give a causal read on individual channels. Calibrating the model with experiment results is the strongest option, because it anchors modelled estimates to a known truth.
What is marketing mix modeling and how does it relate to media planning?
Marketing mix modeling is a statistical method that estimates how much each media channel, along with price, promotion and external factors, contributed to sales. It relates to media planning in two directions. Backwards, it grades the plan that just ran. Forwards, it supplies the response and saturation curves that determine how the next budget should be split, which is the input the allocation step needs and rarely has.
What tools do media planners use?
Planners typically use audience research platforms for sizing and profiling, planning and buying platforms or demand-side platforms for inventory and forecasting, competitive intelligence tools for share of voice, and analytics or measurement platforms for outcomes. The gap in most stacks sits between the measurement output and the plan itself, where model results have to be translated by hand into a budget split. That translation step is increasingly automated.
How Aryma approaches it
Media planning is a forecasting exercise, and forecasts are only as good as the evidence behind them. Most of the difficulty in this discipline is not choosing channels. It is knowing what each channel actually returned last time, where it stopped returning, and how to move money accordingly without breaking commitments already in place.
That loop is what we build for. Aryma Nebula takes marketing mix modeling outputs and turns them into weekly and monthly budget allocation decisions grounded in measured ROI, saturation points and real-world constraints, from $99 per month. The statistical core stays rigorous and human-led. The allocation work around it does not need to stay manual.
If you want the reasoning rather than the tooling, our product range covers how the models are built, validated and interpreted, and MMMGPT answers methodology questions directly from a decade of MMM knowledge.
Aryma Labs is a marketing mix modeling consultancy founded in 2019. Aryma AI is its Gen AI division, applying agents to the periphery of MMM while keeping the statistical core human-led.
Explore the Aryma AI suite
Gen AI products for marketing mix modeling, built on a human-led statistical core. Explore the suite, or talk to the team.
More from Aryma
Keep reading
Peripheral Agentic MMM
A deep dive into how AI is reshaping Marketing Mix Modeling (MMM), yet not replacing the foundation of MMM
Why LLM Observability Matters: How We Measure Every Query Inside MMMGPT
How Aryma Labs uses Langfuse to trace every MMMGPT query end to end, turning AI debugging from guesswork into evidence.
The Art of Subtraction: Training AI Agents in Marketing Mix Modeling with 'Via Negativa'
How exclusionary prompts and restraints are reshaping the future of smarter, more nuanced MMM AI - lessons from Aryma Labs