What do price rises do?

A: "I'm worried what will happen to my brand if we raise price. I don't know what we can get away with."

B: "Don't worry, I can (for a fee) estimate your brand's particular pricing power. I'll use powerful statistics to tease out the causal impact of previous price changes on volume."

A: "That sounds really useful. But we haven't changed price much in the past, other than lots of temporary discounts."

B: "Don't you worry, the statistical techniques we use today are like magic".

A: "OK, will this tell me how much I should increase the price, and what will happen if some competitors adopt different pricing strategies?"

B: "It will give great insight"

A: "So you'll stand by your predictions?"

B: "Ah well, of course the models might have to be recalibrated afterwards, all sorts of unknown endogenous things might occur, and... measurement error...and and..."

Marketers fear the consequences of raising prices. Rightly so too. You don't want to be the one who undoes years, even decades, of prior work and hands over market share to competitors.

Yet as costs rise price increases become necessary. Which makes many marketers nervous. It's here that statisticans step in offering to calculate your brand's particular elasticity. This sounds scientific (it isn't). It implies that this will allow you to predict the consequences of a price change (only if you are lucky).

Predicting the future is difficult, and the track record of econometrics (aka, time-series modelling, marketing mix modelling etc) has a particularly miserable track record. There are well documented reasons for this, well known to forecasting scientists but pretty much ignored (or not known?) by many modellers.

Knowing that your brand's price elasticity is say about -2 is kind of useful but wide variation has been detected across different price changes. This is because contextual effects matter enormously. A brand simply doesn't have a fixed price elasticity, how much volume will change depends on the magnitude, how it is signalled, whether it's an increase or a decrease, how many rival brands it "passes", what other brands do, what the retailer does, and so on.

University scientists have been studying this with experiments. This is the correct way to tease out causal impact, not by heroically hoping that modelling a rather random slice of historic data will reveal the future.

More research is underway at the Ehrenberg-Bass Institute. And it's needed....unfortunately many brands in recent times raised prices after being reassured by analysts that they had "pricing power" but then experienced penetration and volume losses.

Remember, if you want to raise your prices relative to competitors without losing volume then improve your mental & physical availability, and/or improve your product. Ideally do both.