The slide that is missing
Most campaign plans I have reviewed have a slide for reach, a slide for engagement and a slide for budget. Very few have the slide that connects those numbers to incremental sales. The plan says how many people will see the work. It does not say how many more units the business will sell because they did.
Marketers know this. The link is hard to make, the data is incomplete and the deadline is close, so the plan goes forward on media metrics and good intentions. When the finance team later asks what the campaign returned, there is no agreed answer because no expectation was ever written down.
AI has made the problem sharper. A language model can now produce a polished campaign plan in a few minutes. It will have objectives, audiences, channels and a timeline. It will read well, and it will contain numbers that nobody calculated.
Calculated or assumed
The most useful discipline in campaign planning is to label every number as either calculated or assumed.
A calculated number follows from other numbers by arithmetic. If the budget is fixed and the cost per thousand impressions is known, impressions are calculated. An assumed number is a judgement. The share of people who will notice the ad, the share who will consider the product and the share who will buy are all assumptions, however confident the planner sounds.
Most plans blend the two until they cannot be told apart. A reader sees a sales forecast to the nearest unit and takes it as fact. It is a chain of four or five assumptions multiplied together, and a small change in any one of them moves the result a long way.
Work back from the commercial goal
The fix is to plan in the opposite direction. Start with the commercial goal, stated as incremental sales or revenue over a period. Work out how many additional buyers that requires. Work out how many people must consider the product to produce those buyers, and how many must be reached to produce that consideration.
At each step, write down the conversion rate you are assuming and where it came from. Past campaigns are the best source. Category norms are second best. A guess is acceptable if it is labelled as one.
Consider a brand that wants 2,000 additional sales from an eight-week campaign. If it assumes that one in fifty people who consider the product will buy, it needs 100,000 people to consider it. If one in ten people reached will consider it, the campaign must reach a million people.
Now the reach target has a reason. The two assumptions behind it are on the table, where they can be challenged by someone who knows the category.
Run more than one scenario
A single forecast invites false confidence. Three scenarios do the job better: one where the assumptions hold, one where they are a third worse and one where they are a third better. If the campaign only pays back in the best case, that is worth knowing before the money is spent.
Scenarios also change the conversation with finance. A range with stated assumptions is easier to defend than a point estimate. When results come in, you can see which assumption was wrong, and the next plan gets better.
After the campaign, put the actual results beside the plan. Most teams skip this step because the next campaign has already started. It is the only way assumptions turn into evidence.
Why the assumptions stay hidden
There are understandable reasons the assumptions stay hidden. Agencies are paid to sound certain, and a range can look like weakness in a pitch. Marketing teams fear that a stated conversion rate will be used against them if results fall short. Finance teams, for their part, often ask for a single number.
I spent years building and running businesses across India and Southeast Asia. The plans that held up under scrutiny were the ones where someone had written the assumptions down early and invited challenge. The awkward conversation happened before the money was spent, when it was still cheap.
Regional campaigns raise the stakes. A conversion rate that holds in Malaysia may not hold in Indonesia or Vietnam, so each market needs its own stated assumptions.
A plan finance can read
A campaign brief that states the commercial goal, separates calculation from assumption and shows a range is a different document from the usual deck. Someone outside marketing can read it. It can also be wrong in a way that teaches you something.
MetaModel is the product we built for this job. A marketer describes a campaign and gets a free diagnostic. The paid campaign package, at US$49, adds the brief and the commercial simulations. It runs at metamodel.systems and needs no advisory engagement.
Use the tool or use a spreadsheet. Either way, write the missing slide before the campaign runs.