Lynne Kiesling

There was an interesting little blurb in Tuesday’s Wall Street Journal on the increasing use of econometrics in advertising (subscription required).

Econometrics uses statistical analysis to measure the relationship between different sets of events, such as the effect of educational qualifications on wage levels. To determine an advertisement’s effectiveness, econometricians write an equation to measure the effect on sales of different factors, including the weather, price cuts and advertising. For the advertiser, the purpose is to help decide which ads to run.

I’ve often thought that such an approach would improve the returns to advertising, by providing a systematic tool to evaluate the effect of advertising on sales revenue, controlling for other factors that influence sales.

I wonder how they account for price cuts as an independent variable; the dependent variable is P*Q, so if you include anything to do with P on the right-hand side you have to take a fancier approach. Perhaps they measure price cuts as a percentage discount, or a dummy variable …

Not everyone is convinced this is a good thing:

For an ad industry that prides itself on its creative spirit, the idea of being judged by mathematical models is upsetting, if not heretical, to some. Some creative staff argue that it is impossible to measure the buzz generated by good advertising.

No it’s not: the buzz that matters is the buzz that leads to increased sales revenue. Easy to quantify. A firm’s objective is to maximize profit, not to maximize the creativity of advertising. The point is to use more systematic tools to figure out what resonates with consumers.

## 12 thoughts on “Econometrics In Advertising”

1. Peter says:

Firms in the real world (unlike those in economic text books) have all sorts of objectives other than maximizing profits. An important objective is to maximize the share price of the company, and this may require creating a buzz among analysts and investors. Creative advertizing can play a large part in achieving such a buzz. It is not only consumers who are the targets of advertising.

2. That’s a good point. But is there an effect of buzz on share prices *separate from* the manifestation of that effect in sales revenue? I can see this idea of buzz leading to, say, bubble phenomena. But how persistent is it? Isn’t the effect of advertising on profit going to be the effect on share prices with the higher magnitude?

3. Joshua Wright says:

Using price discounts on the right hand side, or any other measure related to the actual price for that matter, would not resolve the endogeneity issue between the price variable and revenues (or profits). Some sort of instrumental variable approach would be required. The other option, which might be feasible since the firms control which markets receive and don’t receive discounts, is some sort of natural experiment approach.

Econometric strategies aside, it is hard to imagine a sensible reason that a firm would not be better off with this type of information than without.

4. Joshua Wright says:

Using price discounts on the right hand side, or any other measure related to the actual price for that matter, would not resolve the endogeneity issue between the price variable and revenues (or profits). Some sort of instrumental variable approach would be required. The other option, which might be feasible since the firms control which markets receive and don’t receive discounts, is some sort of natural experiment approach.

Econometric strategies aside, it is hard to imagine a sensible reason that a firm would not be better off with this type of information than without.

5. Peter says:

“But is there an effect of buzz on share prices *separate from* the manifestation of that effect in sales revenue?”

Well, of course there is! Lots of companies are listed on stock exchanges geographically distinct from their sales operations! Why do Indonesian or Korean companies advertise in “Business Week” or “The Economist”? To influence stock analysts and investors, not their target customers.

6. Peter says:

“But is there an effect of buzz on share prices *separate from* the manifestation of that effect in sales revenue?”

Well, of course there is! Lots of companies are listed on stock exchanges geographically distinct from their sales operations! Why do Indonesian or Korean companies advertise in “Business Week” or “The Economist”? To influence stock analysts and investors, not their target customers.

7. But my point is that analysts may be lured by buzz, but the buzz will only have a long-term effect on share price if it’s matched by performance.

8. But my point is that analysts may be lured by buzz, but the buzz will only have a long-term effect on share price if it’s matched by performance.

9. Peter says:

Lynne — You have more faith in the impersonal decision-making abilities of stock analysts than I do. I think lots of analysts, a lot of the time, are making decisions on the basis of personal feelings, impressions, rumour, and their estimations of the likely opinions of others, rather than any non-emotional hard-edged calculations about company worth. Accordingly, I see no logical reason for the enthusiasms (or disparagements, if that’s the word) which analysts have for particular companies to be misaligned with a company’s sales figures only in the short-term — a misalignment may go one for a long time.

As an example, over the last two decades we’ve seen stock analysts in most western countries favour electricity and gas companies who enter deregulated consumer telecoms markets. But ask any analyst if they would favour a telco entering a consumer energy market? No one of my acquaintance thinks this would be sensible. Why is that? When I ask, no one can give me a reason. The only rational explanation I can see is the presence of an emotional bias by analysts: cross-investment in one direction is favoured, while in the other direction it is not.

10. Peter says:

Lynne — You have more faith in the impersonal decision-making abilities of stock analysts than I do. I think lots of analysts, a lot of the time, are making decisions on the basis of personal feelings, impressions, rumour, and their estimations of the likely opinions of others, rather than any non-emotional hard-edged calculations about company worth. Accordingly, I see no logical reason for the enthusiasms (or disparagements, if that’s the word) which analysts have for particular companies to be misaligned with a company’s sales figures only in the short-term — a misalignment may go one for a long time.

As an example, over the last two decades we’ve seen stock analysts in most western countries favour electricity and gas companies who enter deregulated consumer telecoms markets. But ask any analyst if they would favour a telco entering a consumer energy market? No one of my acquaintance thinks this would be sensible. Why is that? When I ask, no one can give me a reason. The only rational explanation I can see is the presence of an emotional bias by analysts: cross-investment in one direction is favoured, while in the other direction it is not.

11. Numbers vs. Buzz

It looks like the advertising industry is having it’s own Moneyball vs. Scouts battle….

12. Numbers vs. Buzz

It looks like the advertising industry is having it’s own Moneyball vs. Scouts battle….