Last week, we explored how excessive discounting might be training your customers to wait for your next promotion. Certainly, an occasional discount can help an otherwise undecided customer pull the trigger. But how do brands know whether discounts actually work? And how can you know whether you can safely raise prices without making sales drop precipitously? That’s where price elasticity comes into consideration. Pricing – one of the four Ps of the marketing mix – is just as strategically important as promotion. Your brand’s prices – just like its promotions – must be positioned carefully. Understanding price elasticity helps brands price more strategically.

How Sensitive Are Your Customers to Price?

Price elasticity measures how demand changes after a change in price. For example, a DTC chocolatier brand might notice a 20% increase in subscriptions when the monthly subscription price is reduced by 5%. In this case, demand for the premium chocolate subscription is elastic: a small drop in price resulted in a much larger increase in demand.

Conversely, a change in price might have little effect on demand. For instance, a DTC supplier of diabetes accessories like needles, syringes, and strips, might increase the price of any or all of these by 10%. Demand might go down only slightly – perhaps 1-2 percent, if at all. This is because diabetes patients need those accessories to manage their condition. Demand in this case is inelastic: a spike in price didn’t change demand much. Of course, in the long run, these customers might shop around for better prices on their accessories, so brands shouldn’t rest comfortably in inelasticity.

Why Many Brands Misread Promotional Performance

Naïve analysis of elasticity can mislead. Many brands misread or misinterpret their promotional performance because they focus more on moving products than on building incremental demand. And this fatal error can lead to undesirable changes in customer behavior.

We gave the example last week of a DTC coffee brand that discounted its premium flavor, and saw the unintended outcomes:

• Sales spiked and then collapsed
• Customers moved their planned purchases earlier, cannibalizing future orders
• Margins eroded as customers anchored the value of the premium flavor to the discounted price
• Customers became dependent on promotions, waiting for the next sale.

The coffee company could also have cannibalized sales of its lower-priced flavors, as customers wait to purchase the premium brand during the next sale, eroding margins even further.

Without understanding how promotions affect customer behavior and future sales, brands risk pricing peril.

Fortunately, you can model your customers’ price sensitivity; one popular approach is log-log regression.

How Analysts Estimate Price Elasticity

High sales volume during promotions does not automatically mean customers are price elastic, for the reasons given above. And a tiny drop in sales after a modest price hike doesn’t automatically mean customers are price inelastic. Promotions can create dependency, while perceptions of premium value or necessity – brand equity – can supersede perceptions of “expensive.” Elasticity modeling can help you determine whether discounts create incremental demand or price hikes increase revenue.

Log-log regression models independent variables such as promotion and pricing against a dependent variable like quantity sold. The natural logs of the dependent and independent variable values are computed before constructing the model.

Suppose a luxury skincare brand tests several prices for a new facial cream. It examines the units of the cream it sells at each price, shown in the table below.

Graphing these raw prices and sales, we get the following chart:

The relationship between price and units sold is curved rather than perfectly linear, so a simple linear regression may not accurately capture customer price sensitivity.

But if we compute the log of the prices and quantities and then perform log-log regression, we can estimate the percentage response to price rather than raw unit differences. That’s much more realistic: businesses often think in terms of percentages rather than absolute unit changes, making log-log models especially useful for pricing analysis.

Like regression, log-log regression will generate a formula, but the output and the coefficients will represent log values:

The relationship between price and units sold is curved rather than perfectly linear, so a simple linear regression may not accurately capture customer price sensitivity.

But if we compute the log of the prices and quantities and then perform log-log regression, we can estimate the percentage response to price rather than raw unit differences. That’s much more realistic: businesses often think in terms of percentages rather than absolute unit changes, making log-log models especially useful for pricing analysis.

Like simple regression, log-log regression will generate a formula, but the output and the coefficients will represent log values:

In a log-log model, the coefficient (ϐ) directly estimates elasticity. ϐ usually will have a negative sign since demand generally falls when prices rise.  Therefore, if ϐ= -1.5, then that means a 1% price increase is correlated with a 1.5% decline in demand.

True Elasticity Isn’t Just About Price

The skincare example above is greatly simplified. Often, brands will have several independent variables beyond price that they must control for, including:

• Seasonality and holidays
• Promotion activity and spend
• Inventory
• Competitor pricing
• Channel mix

Without controlling for these factors, you risk confusing marketing effects with price effects.

What Price Elasticity Modeling Can Reveal About Your Brand

Elasticity modeling offers several potential insights. In addition to identifying which products can support price increases and determining when discounts help or hurt, elasticity modeling can also help you identify or determine your most price sensitive customers, the SKUs with the greatest pricing power, your optimal promotion frequency, and bundle pricing opportunities.

Had the coffee brand above conducted proper elasticity modeling, it could have developed a pricing and promotion strategy that drove true demand growth: perhaps a one-time discount only to new customers; $10 off of a total purchase of $100 or more; free shipping on certain order amounts; or a free pound of coffee beans with the purchase of a French press. Targeted discounting strategies like these have the greatest potential to build incremental demand.

Elasticity: Here Today Gone Tomorrow

Demand for your products won’t always be elastic or inelastic. Even if your product is unique or superior and demand has remained stable over several price increases, that elasticity won’t last forever.

Economic conditions change. New competitors and products enter the market. Substitutes – products that are different from yours but provide the same benefit – become available. Your customers enter new life stages. Your product matures. Technology evolves. Each of these conditions changes the game, especially demand for your product.

Even if your product’s demand seems inelastic, you should always have an eye on when it will become a commodity. Elasticity should be monitored continuously.

Pricing Should Be Measured – Not Guessed

Elasticity modeling illustrates that pricing should be measured, not guessed. In many cases, the goal is not maximizing units sold but rather maximizing profitable revenue. Discounting without understanding elasticity can damage profitability. And today, even small and midsize DTC brands have access to advanced analytics approaches for price optimization.

Brands that understand elasticity can price with far more confidence than brands relying solely on instinct.