Estimating outcomes without treating historical averages as guarantees

Use historical results as a starting assumption, then test what changes when the audience, offer, price or measurement rule differs. An average describes the periods that produced it. It does not promise the next campaign's result or the return from extra spend.

Rebuild the historical measure

Choose the business outcome before calculating a rate. For enquiries, specify whether the count means form submissions, accepted enquiries or booked work. Check duplicates and allow the agreed time for a response to arrive.

Match spend and outcomes under a consistent campaign and follow-up window, cost boundary and GST basis. Keep raw spend and outcome counts beside any cost-per-outcome figure.

Do not average campaign rates without considering their denominators. In a hypothetical example, one flight costs A$1,000 for 10 accepted enquiries and another costs A$6,000 for 30. Their individual costs are A$100 and A$200 per accepted enquiry.

The simple average of those two rates is A$150, but the pooled cost is A$7,000 for 40 enquiries, or A$175 each. Both calculations are correct; they answer different questions. Neither predicts the next flight.

Identify what may no longer carry over

For each past period, record the market, audience, placement, creative, offer, media price and availability. Mark changes in outcome definition or collection.

A cheaper historical period may have reached an easier audience or used inventory unavailable at the same terms. Keep sparse periods and unusual promotions visible rather than hiding them inside a blended average.

Key Conditions to Verify Before Estimating Media Outcomes

Make a conditional estimate

Begin with the most relevant observed results, not the period that makes the target look easiest. State how many periods and outcomes inform the assumption. Set cautious, base and optimistic rates only after documenting expected changes in the proposed campaign.

Show the outcome counts implied by each rate at the proposed spend. If the history is too unlike the new buy, call the calculation a sensitivity exercise rather than a forecast.

Check also whether the tool's conversion definition matches the accepted enquiry definition used in your business records.

Media Outcome Estimation Best Practices

Review the prediction

Before launch, save the historical data, exclusions, assumed rate, forecast date and decision threshold. During delivery, check whether the market, placement, price and outcome collection still match. If they change, update the remaining estimate and explain why.

At close, compare the forecast with the result under the same counting rule, allowing for delayed or missing outcomes. Record which assumption held and which did not.

Reviewing a Media Outcome Prediction