Promotions can generate up to 40% of revenue, but they also bring significant uncertainty and often excessive discounts, costing companies millions in lost margin each year. Today this can be measured and recovered.
Billions flow through promotions. Why is margin leakage so high?
Ernest Wojciulewicz: Because this is an area of high uncertainty. Discounts, promotion types, media, product selection, market context, and time pressure all happen in parallel. When outcome forecasts are uncertain, organizations pay a “certainty premium” and choose the safer option - usually a discount that is too deep. With PLN 1 billion in promotional turnover, just 1 percentage point of excess discount means roughly PLN 10 million of margin given away.
But this cost is not visible in the P&L...
E.W.: Exactly. No one reports a “safety buffer,” yet we see that often 30-60% of promotions do not deliver the expected growth and profitability. These are not people errors; this is the cost of uncertainty: hundreds of decisions, high factor complexity, and limited data.
What does Effiana change?
E.W.: We turn intuition into a measurable process and remove uncertainty from it. Effiana is an AI model built individually for each client, integrating marketing data, commercial data, and market context. It forecasts promotion outcomes and ROI. It helps teams choose optimal parameters: discount, mechanics, media, and assortment. It reduces excessive investment and enables recovery of 4-7% of promotional margin - in mid-sized and large companies this is usually from several to several dozen million PLN recovered annually.
Why now? And are data not a barrier?
E.W.: Now, because this is an AI use case with fast, measurable return. Promotional decisions are repeatable and financially significant, and the technology is mature enough to optimize them. Data can be a barrier, but in this area organizations usually have more of it than they think.
Are leadership teams afraid of implementation risk?
E.W.: They mainly need credible ROI. That is why we start with a PoC. In 10-12 weeks, on the client’s own data, we show model effectiveness: how much uncertainty currently costs and how this cost is recovered. Only then is a full implementation decision made.
If margin is leaking, does every month of waiting have a cost?
E.W.: Yes. If annual margin loss is PLN 12-24 million, each month means PLN 1-2 million. Additionally, competitors may already be moving to this standard - just as they previously moved to price and revenue management. It is not worth waiting.
One sentence to close?
E.W.: Margin leakage in promotions is neither a team mistake nor a necessity. It is the cost of uncertainty. It can be measured and recovered quickly.
