Virtual Market Surrogates

From Predicting Market Response to Testing Commercial Decisions

Why pricing and promotion intelligence may need to move beyond simply predicting what happens next.

Prediction only gets us so farMost pricing technology is trying to answer a perfectly reasonable question: What is likely to happen?If we change the price, how will demand respond? If we run a promotion, what lift might we get? What is the likely revenue or margin impact? All useful questions. But when you are actually making the decision, there is usually another question sitting behind them: What would happen if we made a different decision? What if we changed the price by 5% instead of 2%? What if we did not run the promotion? What if we changed the offer completely? That is a different problem. Prediction tells us what is likely to happen. Decision intelligence needs to help us think through the alternatives.

History has its limits

Of course, much of what we know about a market comes from history. We look at what we did, what the market did in response, and try to learn from it. But there is an obvious limitation. Historical data tells us what happened after the decisions we actually made. It cannot directly tell us what would have happened if we had made a different decision. There is another complication. We never see the market perfectly. We can observe prices, sales, promotions, competitors and customer behavior. But things like willingness to pay, buyer confidence, substitution pressure or changing market sentiment are much harder to see directly. So the challenge is not simply having more data. It is understanding enough about the market to make a better decision. Building a useful representation of the market

That is what led us to the idea of a Virtual Market Surrogate.

The term may sound technical, but the thinking is fairly straightforward. If we cannot observe the entire market, can we build a useful representation of the parts of the market that matter to the decision we are trying to make? That representation can bring together what we can observe with what we can derive, model and infer. The objective is not to build a perfect digital copy of the market. That is neither realistic nor necessary. The objective is to understand the market well enough to support the decision at hand. And as new decisions are made, the market responds. Those responses give us new information. Change a price and something happens. Run a promotion and something happens. Make an offer and the customer accepts, rejects or behaves differently. Each response tells us a little more.

From predicting outcomes to testing decisions

This is where the idea becomes interesting for us. If we can build a richer understanding of the market and keep learning from how the market responds, we can start moving beyond simply forecasting what happens next. We can begin to evaluate alternative decisions against that understanding of the market. In simple terms:

Understand the market. Test the alternatives. Make the decision. Learn from the response.

That is what we mean by Virtual Market Surrogate — and an important part of how we are approaching the future of pricing and promotion decision intelligence at Gestalt and Price-It-On®. __________________________________________________________________

Gestalt has filed a provisional patent application covering aspects of the Virtual Market Surrogate technology.