Forecast complexity

Life is complex. This makes predictions – e.g., the results of soccer matches, the weather, or financial markets – difficult. Luckily, there are often simple rules that provide strong guidance. If you bet on a 2:1 victory of the stronger team in a soccer match, you have a better chance of being right than the experts on TV. In meteorology, the forecast “tomorrow will be just like today” is right in about 2 out of 3 days and was, until very recently, better than most professional predictions. In finance, simply buying the market portfolio will beat most other investment strategies.

In pricing, we are interested in understanding what product offer and prices optimize revenue or profit. Here, cost-plus pricing is a simple rule that produces decent and reliable results. Despite being ridiculed by many pricing professionals, cost-plus pricing is still widely and successfully applied by decision makers around the world, as was recently discussed in the Harvard Business Review. To beat it, pricing professionals have thrown in many different ideas that typically put more emphasis on the customer and competition, rather than on the product and its costs. Two prominent examples are value-based pricing and price elasticity-based pricing. Yet, both often fail to perform better than cost-plus pricing, because they build on oversimplified core assumptions.

In value pricing, a group of customers is typically replaced by ‘the customer’ who is the average of the customers in a segment to make the market situations easy to understand. The joke about the statistician who missed the rabbit first to the right, then to the left, and then claimed that the rabbit is dead – on average – is well known. However, marketers are making the same mistake if they tailor their products and pricing to some average customer who does not exist in the real world.

Another flaw in value pricing comes from focusing solely on value and ignoring costs. Particularly, in the case of slim margins this can be incredibly wrong. The Harvard Business Review explains value pricing with the following example. Let there be two competitors, A and B, in the big-screen TV market segment, each offering one TV model. A determined that because of customer preferences, her offer is worth $150 more than B’s offer. By pricing relative to the next best offer, which is B’s TV at $799, A determines a value-based price of $949. If B reduces the price to $399, then A must follow and price at $549. This can easily turn into a losing business if costs are not considered. Value pricing very often misses a consideration of how profit or revenue are affected by the price difference between products (here, A’s and B’s TV) at different price levels. Certainly, value pricing can be done better – and it often is – e.g. if it is based on a conjoint study and market simulation. However, conjoint studies are very costly, take a long time, and are thus only applicable in limited situations – and they are inflexible and cannot easily be adjusted to answer new questions.

Using the price elasticity to compute volume reactions to price changes is another example of dangerous conceptual oversimplification. Frequently, such calculations implicitly assume that elasticity is constant, or that the price demand function is linear across a larger price range. In a market with just one product, this might be acceptable. In a market with multiple products and competitors, such assumptions are grotesquely wrong. If a price is increased above that of a close competitor, the elasticity can easily jump from -2 to -10. Price elasticity is good for educational or recreational pricing exercises and back-of-the-envelope estimates, but not for actual price decisions if real money is on the table. The differential calculus, elasticity calculation is based on, was used to calculate the motion of the moon with pen and paper 1,500 years ago. Astrophysics has long moved forward and now relies heavily on computer simulations. It is now time for pricing to also make the next step and leave pen and paper behind.

buynomics' way

We believe that for a true leap forward, pricing needs to be built on a precise model of the underlying purchase and market dynamics. To see why this matters, consider Nate Silver’s comparison of progress in weather and earthquake forecasting in his 2012 best-seller “The Signal and the Noise”. Despite what you may think, weather forecasts have become increasingly better over the past decades. On the other hand, earthquakes remain notoriously unpredictable. Why? Because meteorologists have developed an increasingly better computer model of weather dynamics – e.g. they simulate the movement of clouds or the build-up of storms. The more precise the model and the faster their computers, the better their predictions. On the other hand, the tectonic dynamics that cause earthquakes are still largely unknown. Maybe, this is because they happen deep under the surface and are hidden from our sight. The history of science teaches that a clear understanding and modeling of the underlying dynamics is the starting point of improvement.

To get pricing right, therefore, we need to get customers’ behaviors right. Just like the weather dynamics in meteorology, we need a model of customer behavior. However, much like in geology, today’s pricing concepts are mostly blind to the underlying dynamics responsible for customers’ purchases, and thus are not improving.

Unhappy with this situation, we have developed buynomics. buynomics is a technology that makes full use of recent advances in large-scale computer simulations, big data analyses, and artificial intelligence (AI) to model the buying behavior of customers. buynomics evaluates market data such as sales figures, price information, expert knowledge, and market studies – and combines these with the knowledge from our customer behavior database. Using AI powered algorithms, these insights are used to invigorate virtual customers that behave just like their real-life counterparts. This gives pricing professionals full transparency on how to steer and optimize sales, revenue, and profit.

buynomics can beat conjoint based pricing in established markets with available market data. buynomics predicts more precisely how price and product choices affect sales, and it pinpoints the best pricing strategy to optimize revenues or profits. As a true software solution, buynomics can economically price large portfolios and not just selected high-value products.

We are very excited about the potential of buynomics and how it improves pricing. Subscribe to our blog, if you want to stay in touch learn more. In the coming weeks, we will share many pricing insights and show you how buynomics will make your pricing easier, better, and more profitable.

Want to know more?

Get Your Demo