If you want to know how to plan a programmatic campaign for the first time, start with the smallest scope that can answer one specific question. In other words, your first campaign should give you an answer good enough to guide your next decision. This guide walks through the programmatic media plan, the programmatic campaign budget and the programmatic campaign setup in ten steps, and ends with a worked example.

Key takeaways

• A first campaign should answer one clear question, not prove that programmatic works.
• Build the budget from the audience and exposure you need, and narrow the scope if it does not fit.
• Prepare tracking, baselines and reports before the flight, because later changes cannot recreate missing data.

Where should you start when you plan a programmatic campaign?

Start with the MINT Square planning sequence: objective, audience, customer journey, channel roles, budget, creative, measurement and optimisation. However, that sequence assumes an organisation that already plans campaigns. A first-time buyer therefore needs three additions: a statement of what programmatic should add, a decision about technology and operating model, and a documented learning step at the end.

To plan a programmatic campaign on that basis, follow ten steps. Each one ends with an output you can check:

• Business objective: a commercial outcome the budget owner accepts.
• Audience: the people or households you need to influence.
• Advertising job: demand creation, demand capture or a clearly separated combination.
• What programmatic adds: a capability hypothesis and acceptable evidence.
• Environments: a role and a reason for each one.
• Access and operations: a workable setup, owners and dates.
• Budget and duration: enough delivery and observation time for the question.
• Creative: suitable assets and a production schedule.
• Measurement: tested tracking, baselines and agreed reports.
• Launch and review: an evidence-based recommendation for what happens next.

Steps 1 to 3: objective, audience and advertising job

These first steps of programmatic campaign planning decide what success means and who you need to reach. If you get them wrong, a campaign can hit its media target and still miss the business outcome.

Step 1 is the business objective. State what needs to improve commercially, for example more qualified enquiries or additional profitable sales. A rate such as a lower CPA (cost per acquisition) can come from a small group of easy conversions, so check it against volume and value. Then agree the objective with the person who approves the budget, and share it before launch, even when nobody has asked.

Step 2 is the audience. Describe people by their needs and situation first, for example “households planning a renovation who may need to replace their heating system in the next year”. A platform segment is only one way to approximate that. Also check what your data represents, because a converter profile may mainly describe people already ready to buy.

Step 3 is the advertising job. Decide whether the activity introduces the brand to future buyers (demand creation), helps interested people act (demand capture) or does both. Our article on brand and performance marketing explains the difference, and the Institute of Practitioners in Advertising (IPA) adds background in its work on finding the right mix of advertising activities. If you combine the two jobs, separate their budget and reporting. Otherwise, a blended cost per acquisition becomes the number everyone quotes.

Steps 4 to 6: from hypothesis to programmatic campaign setup

These steps turn your intent into a programmatic media plan you can run. Each one ends with a check before you move on.

Step 4: state what programmatic adds

Bring the capability hypothesis from deciding whether programmatic is worth it into your brief, tied to a specific gap. For example, you might expect audio placements to add listening opportunities beyond your existing screen-based activity. Then state what the evidence can and cannot show. For instance, customer suppression as a proxy for prospecting has matching limits and does not prove unique reach.

Step 5: select the environments for your programmatic media plan

Choose each environment because it performs a needed role, and one or two may be enough. Audio plus display does not guarantee different people; instead, it may give you different moments with overlapping households. Do not pick a format just because a competitor used it or a DSP (demand-side platform) offers it. If an environment lacks a clear role, budget or suitable creative, take it out of the programmatic media plan.

Step 6: confirm the programmatic campaign setup

Choose the access arrangement and decide who runs the work day to day. Next, put contracting, tracking, data, quality checks, assets and approvals for the programmatic campaign setup into one schedule with named owners. If the launch date is fixed, narrow the scope to what can be ready and say what learning you lose.

How big should your programmatic campaign budget be?

There is no universal minimum budget that makes every test useful. Instead, build your programmatic campaign budget from the audience and exposure you need, then check a forecast with your partner:

• Estimate the audience in a consistent unit, such as households or recognised devices.
• Choose a planning frequency that fits the message, purchase cycle and campaign period.
• Multiply planned reach by average frequency to estimate impressions.
• Apply the quoted CPM (cost per thousand impressions) for your inventory and deal mix.
• Add platform, data, verification, service, creative and research costs.

In short, media cost = planned impressions ÷ 1,000 × quoted CPM. This is only an initial estimate, because actual reach depends on available opportunities, recognition and delivery quality. With several environments, never add their reach figures and call the result unique people.

Protect the exposure the question requires

A fixed programmatic campaign budget forces a trade-off between audience breadth, repeated exposure and the number of environments. For a demand-creation question that needs repeated exposure, our starting recommendation is to narrow the audience or number of environments before reducing every part of the campaign to a thin layer of delivery. Reaching fewer relevant people with enough exposure is generally more useful than a broad campaign that cannot test its own premise. However, this is a planning recommendation, not a frequency rule.

How long should a first programmatic campaign run?

MINT Square often plans roughly ten to twelve weeks for a first campaign across two environments. That range comes from our delivery experience, not a standard for every objective. In addition, purchase cycles and seasonal events can change it, so work back from the date the result must inform the next decision.

In practice: where programmatic campaign budget gets stranded

In our own delivery work, four patterns recur. This is MINT Square’s experience, not a published finding. One is targeting too narrow to deliver at the assumed daily rate. Another is a conservative frequency cap on an already limited audience. Then there are guaranteed deals that pace to their own schedule while the flexible part of the buy absorbs the difference. Finally, an environment with a longer creative approval starts late and never catches up.

At the end, unspent budget gets pushed into whatever is easiest to buy, which is usually the least informative part of the plan. As a result, the campaign closes on budget and answers a question nobody asked. So name who reviews pacing and how often, and decide in advance what happens to underspend, including not spending it.

Steps 8 to 10: creative, measurement and review

The last three steps protect your ability to learn. Several of the choices they involve are difficult to repair after launch, because later changes cannot recreate missing data or exposure.

Step 8 is creative. As a working approach, MINT Square plans at least three creative executions per environment where budget and format support a useful comparison, plus a refresh for flights beyond six weeks. These are our practices, not industry rules. If you cannot produce suitable assets, reduce the scope of the programmatic media plan, because an unreadable DOOH (digital out-of-home) layout tests the creative as much as the environment.

Step 9 is measurement. Before launch, test website or app events as part of the programmatic campaign setup, save the baseline and separate reporting for each advertising job. Also arrange any survey, control group or holdout before the start. Our guide to choosing the right KPIs explains which evidence fits, while this step makes sure you can obtain it.

Step 10 is launch and review. Check the programmatic campaign setup and delivery before you make performance decisions. Then investigate technical, safety or major pacing issues promptly, but allow enough delivery before you read early differences. Record every material change and its reason. At the review, document the hypothesis, what ran, what changed, what the evidence shows and what remains uncertain. A learning result can justify a sharper second test, but not automatically more investment.

What is the smallest programmatic campaign that answers something?

In programmatic campaign planning, the minimum is a coherent set of decisions, not a budget figure. That means one objective seen by the budget owner, an audience described independently of platform settings and a clear advertising job. Add a capability hypothesis with accepted evidence and a feasible operating model with dates and owners. In contrast to two underfunded environments, a properly funded single-environment test can be the better design.

Worked example: planning a first campaign for a heating brand

This illustrative example is based on planning patterns MINT Square encounters, not on a client case.

A mid-sized advertiser sells heating systems through an installer network in Germany. Its small team runs Paid Social and works with an agency on Search. Its cost of qualified enquiries has risen for several quarters, despite better bids, creative and landing pages. The existing data does not settle why. After checking frequency, coverage, enquiry quality and conversion handling, the review supports investigating an additional media role, without proving that programmatic will improve acquisition cost. Specifically, the test asks whether suitable added placements can reach the target household profile in useful contexts.

The scope narrows from display, audio, CTV (connected TV) and retail media to contextual display and digital audio. CTV is deferred because there is no suitable video or funding. Retail media drops out because the installer-led model does not fit. The resulting programmatic media plan focuses on demand creation over a twelve-week flight. Because time is the team’s main constraint, a managed service handles the programmatic campaign setup and daily operations.

During the flight, the team handles under-delivery in week two, a request for a CPA comparison with Search in week four and a weak creative in week seven, each by the agreed rules. Finally, the review separates observed facts, estimates and interpretation, and no commercial outcome is claimed.

FAQ

Where to go from here

When you plan a programmatic campaign this way, it earns its value by narrowing the next decision. The evidence, rather than the wish to validate your first choice, should decide whether you continue, change the question or stop.

Our playbook Getting Started with Programmatic Advertising in 2027 covers every step and the full worked example in more detail.