Which Marketing Levers Move Sales Most?

Marketing AnalyticsSrinivasan, S.; Vanhuele, M.; Pauwels, K. · 2010Journal of Marketing ResearchEvidence: 62 brands analyzed (of 74) across 4 FMCG categories; 96 four-weekly observations per brand, Jan 1999-May 2006

Topicsmind-set metrics·brand consideration·advertising awareness·brand liking·sales response models·distribution elasticity·leading indicators

At the annual planning meeting, the ad budget and promotion calendar are under review, while distribution is not. That misses the marketing lever associated with the largest sales response in this study. Price is the only other lever associated with a response of a similar order. The ranking compares response size; it does not prescribe budget allocation or show that any lever can be dropped.

For mature consumer-goods brands, distribution is linked to the largest sales response among the four marketing levers.

Data chart

Sales response to a 1% change in each marketing lever

Percent change in brand sales, adding up all effects over time, for a 1% increase in each lever (average of 62 brands).

The ranking shows response size, not where to spend.

How much sales respond differs sharply across the four levers, with distribution and price ahead of promotion and advertising. The chart ranks response size only; the study does not test whether any lever could be dropped.

Total sales response to a 1% increase in each lever, averaged across 62 established French consumer-goods brands, 1999-2006; averages include brand estimates that are not reliably different from zero. Average weighted distribution in the four categories was already 79.6% to 95.0%. Exact values 2.424, -1.734, .277 and .036.

Data chart

Distribution takes longest to reach its peak effect on sales

Average months from a change in each lever until its effect on sales peaks (62 brands).

The lever with the largest sales effect takes longest to peak.
Advertising peaks in under two months, not quarters or years.

Distribution, the lever with the largest sales response, takes the longest of the four to reach its peak effect on sales, and advertising peaks in under two months.

Average time from a change in each lever to the period of its largest effect on brand sales, 62 established French consumer-goods brands, 1999-2006.

Action guide

Protect retail distribution in annual brand plans.Ask sales and trade teams to document how they will protect distribution for each established brand.
Set each lever's review window by how fast its sales effect builds.Promotion and price reach their strongest sales effect soonest, then advertising, then distribution. Use that order to time reviews of an action that looks ineffective; the study sets no stopping rule.
Make sure your sales model includes distribution and mind-set metrics.Marketing-mix models often leave distribution out because it barely changes over the typical three years of weekly data for an established brand. Ask analytics to use a longer history, add advertising awareness, consideration and liking, and report whether forecasts improve.
Turn mind-set declines into response tests.When insights flag a drop in consideration, have brand teams test price or promotion. When liking drops, have them test advertising weight and copy; of the four levers, advertising spending has the smallest association with liking. The study proposes, but does not test, these responses.
Add competitors to tracking and forecasting.Ask insights to collect competitors' advertising awareness, consideration, and liking and give analytics all three series.

Evidence

A 1% increase in distribution (how widely a brand is stocked, weighted by each store's category sales) is associated with about 2.42% higher sales over time, the largest average response of the four marketing levers. A 1% price increase is associated with about 1.73% lower sales, and a 1% increase in promotion or advertising with about 0.28% or 0.04% higher sales.
Distribution's sales effect is also the slowest of the four to peak, at 2.12 months on average. Advertising's effect peaks in 1.83 months, not the several quarters or years advertising agencies sometimes claim. Price and promotion effects peak sooner, at 1.59 and 1.02 months.
Distribution also has the largest association of any lever with consideration, advertising awareness and liking. Advertising spending has the smallest association with liking of the four levers, and a price increase is associated with lower liking.
In the full model, own and competitive mind-set metrics together explain 16.3% of brand-sales forecast variance. The share is nearly even: 8.4% for own metrics versus 7.9% for competitive metrics.
Adding advertising awareness, consideration, and liking raises the average adjusted share of brand sales explained from 53% to 61%.
Mind-set metrics' link with sales peaks after 2.00 to 2.32 months on average, later than price and promotion effects, leaving time to act on a drop in a metric before sales fall.

Key takeaway

For mature consumer-goods brands, distribution is linked to the largest sales response among the four marketing levers.

Limitations

The study covers mature brands in four French packaged-goods categories from 1999 to 2006 and measures sales response, not the financial return from tracking. Its brands were already widely stocked, and the study did not test the distribution result in other countries, durables or services.

Source

Srinivasan, S., Vanhuele, M., & Pauwels, K. (2010). Mind-set metrics in market response models: An integrative approach. Journal of Marketing Research, 47(4), 672–684.

https://doi.org/10.1509/jmkr.47.4.672

Exact values behind the rounded figures: total sales responses of 2.424 for distribution, -1.734 for price, .277 for promotion and .036 for advertising.

Read the paper ↗

Same study, another angle

Brand Metrics as Early WarningBrand metrics add sales information beyond the marketing mix.Read this translation →