How we forecast sales for a pharma brand with no historical data

The brand sat outside the category top-10, with no accumulated data and no stable benchmarks. Standard planning tools — market comparison and extrapolation of past campaigns — did not work here: there was simply no history.

In such conditions forecasts are usually assembled after the fact, around a media plan that already exists. We went the other way.

The idea: plan from sales, not from budget

We built a model that converts media contact into sales, and planned the campaign backwards — from target sales to the number of contacts and the structure of the media mix.

The task

• +33% sales in packs — twice the market rate of 15%.
• Forecast deviation under 5%.
• Achieved without historical data and without trade or promo support.

How it worked

• Reach by channel was converted into 'attentive reach' using perception coefficients, then into awareness growth.
• Awareness was recalculated into consumption and repeat consumption — allowing for purchase frequency, brand portfolio and distribution.
• The required media mix was derived from target sales, and impressions from reach.
• The campaign ran on hyper-segmentation with continuous testing of new segments; the whole funnel lived inside digital.

Results

• Actual: +32% sales against a plan of +33%. The category grew 15% over the same period.
• Market share grew 18% — from 4.4% to 5.2%.
• Distribution stayed stable, prices unchanged, no promotions ran — the media contribution is isolated.
• The model needed no corrections during the campaign: the forecast stayed within tolerance.

What it means

The approach transfers to other categories and brands wherever there is no accumulated media history and uncertainty is high. Media planning stops being budget allocation and becomes a tool for managing sales.