In retail, many consider value pricing – i.e. pricing a product based on its value to the customer – the gold standard. In most cases, the value of a product is determined for relevant customer segments – for example by looking at buyer personas. Consider the following ski jacket example. A retailer offers four models that are marketed to their previously identified customer segments. The jackets are differentiated by two value drivers: filling (down and synthetic) and water column (high and low, i.e. above or below 15,000 mm).
Applying value pricing and insights from identified buyer personas (see Figure 1), the retailer determined an average willingness to pay (WTP) of the base jacket (synthetic and low water column) of €250. Further, the average WTP for the down filling vs. the synthetic filling is an additional €50, and the WTP for the high water column (>15,000mm) vs. the low water column is also €50. Figure 2 shows the value price model and the resulting prices of the four jackets. The Blue model is targeted at premium customers, Red is aimed at skiers who want to stay warm but are less concerned about getting wet. The Black model targets sportive skiers who need to stay dry, and Gray is the budget model.
This pricing approach is straight forward and applied by thousands of retailers worldwide. However, it almost always leaves money on the table because in addition to ignoring costs, it more crucially ignores customers’ tradeoff decisions between offers, and how a portfolio can be optimized to consider these effects. For example, if the Blue jacket was €329, it might pull significant demand from the Red and Black jackets (and from current non-buyers), leading to a higher total profit.
Pricing with buynomics
To evaluate such scenarios, let us look a bit deeper into what is going on underneath the hood of value pricing. The simplification – manifested in buyer personas and customer segments – of treating all customers (in a segment) as their average helps marketers better conceptualize the market, but it comes at a cost. From empirical customer preference studies, we know that customers do not share a common WTP for a product. Rather, some would pay more and some less than the average. The result is a distribution like the one shown in Figure 3. This differs sharply from the (implicit) assumption of equal preferences made by value pricing. Taking advantage of variations in customers’ WTP requires a different set of tools. For example, if a marketer is only concerned with one product and the demand function is known (e.g., it is produced by the blue distribution in Figure 3), straight forward price elasticity evaluations provide a good tool for profit optimization.
In a multi-product case, optimization is more difficult. To illustrate, Figure 4 shows the ski jacket case reduced to just the two down jacket models. Each potential customer has a WTP for each of the two jackets. Someone with a strong preference for staying dry, will be willing to pay significantly more for the Blue model, than for the Red model. Overall, some potential customers will not buy either jacket at the current prices of €350 and €300, respectively (quadrant I). Some would only buy the Blue model (II) or the Red model (IV). Some customers would buy either one and must choose given the two jackets’ prices and their personal preferences for each of the two models (III). To maximize profit, marketers need to look at all relevant price variations and evaluate customers’ expected purchases. For more than two products, this becomes difficult to do with paper and pencil – or Excel. That’s why we built buynomics.
Based on customers’ preferences from a similar retail project, we created virtual customers for this ski jacket case study. Virtual customers share the demographics, preferences and WTP for products of their real-life counterparts and make – as a group – the same purchase decisions when faced with the same offers. This makes them an ideal tool for analyzing the effects of price changes on the demand of different products. Figure 5 shows the comparison of the value pricing solution versus the profit optimum identified by buynomics. The buynomics method leads to a profit improvement of 8% against value prices*. In this case study, the price spread between the top (Blue) and the bottom (Gray) models was reduced to better facilitate upselling, which was the main profit lever. Note for example, that in the optimum the Gray jacket has very low sales (13 units), because the price is very close to that of the Red and Black jackets. If the price of the Gray jacket is reduced by €20, sales increase to 1038 units and the profit loss is limited. The pricing manager should analyze the effects of price changes on sales, revenue, and profit and then also evaluate the price structure. For example, what is the effect of a low entry price on profits? How does a close competitor affect sales? Both questions can be studies with buynomics.
Please note that the proposed reduction in the price spread of the portfolio is specific to this example and will not always be the result of price optimization. However, we frequently find that standard value pricing methods lead to a too wide price spread in a portfolio and profits significantly below the optimum.
Please take a look into the buynomics app to see for yourself, how sales react to price changes. In this example, the virtual customers were constructed from historical sales data, validated by the client’s marketing and sales teams, and confirmed by subsequent sales. buynomics can train virtual customers to perform a wide range of behavioral traits such as susceptibility to price thresholds or a preference for a middle option – where this is relevant. The methods can be used both to develop and price new products and to improve existing prices. Further, because the method produces very reliable predictions of future sales volumes depending on their prices, buynomics is also a very capable aid for sales forecasting.
Want to know more?
* In the actual retail case, the profit increase was also 8%, and sales volumes, prices, and costs were proportional to those shown in this case study.