What the Published Research Actually Says About Advertising Effectiveness

A quick disclosure before anything else: there are no RateAds numbers in this article. RateAds is a public-beta platform. We do not have a dataset large enough to generalise from, so we are not going to dress one up. What we can usefully do is point at the published work that already exists, say what it found, and link to it so you can check us.

Everything below is sourced to a specific paper. Each one is listed at the end with a DOI. If a claim here is not in the source, that is our error and we would like to know about it.

1. The average advertising effect is smaller than most decks assume

The most-cited synthesis on this is a meta-analysis by Sethuraman, Tellis and Briesch, published in the Journal of Marketing Research in 2011. They pooled 751 short-term and 402 long-term brand-advertising elasticity estimates drawn from 56 studies published between 1960 and 2008.

Their headline: the average short-term advertising elasticity is 0.12, and the average long-term elasticity is 0.24. In plain terms, a 1% increase in advertising is associated on average with roughly a tenth of a percent more sales in the short run. Both figures are substantially below the previous meta-analytic estimates of 0.22 and 0.41 respectively, and the authors report a decline in elasticity over time.

They also find the effect is not uniform: it is larger for durable goods than non-durables, and larger early in a product's life cycle than in maturity.

The honest reading is not "advertising does not work." It is that the typical effect is modest, varies enormously by context, and is small relative to the confidence most campaign post-mortems project.

2. Spending more is rarely the lever. Changing the work is

Lodish and colleagues analysed 389 real-world BehaviorScan split-cable TV advertising experiments — matched households, actual purchase panels, randomised exposure — and published the results in the Journal of Marketing Research in 1995. It remains one of the largest bodies of genuinely experimental advertising evidence in the literature.

Two findings from that abstract are worth quoting almost directly. First, increasing advertising budgets relative to competitors does not increase sales in general. Second, changing brand, copy and media strategy did raise the likelihood of TV advertising positively affecting sales, particularly in categories with many purchase occasions and low in-store merchandising.

There is a third finding that ought to be uncomfortable for a large part of the pre-testing industry: the authors report that their data do not show a strong relationship between standard recall and persuasion copy-test measures and actual sales effectiveness.

3. Measuring the return is genuinely, structurally hard

This is the part that dashboards tend to hide.

Lewis and Rao ran 25 large digital field experiments with major US retailers and brokerages — most reaching millions of customers, together representing $2.8 million of advertising spend — and reported the result in the Quarterly Journal of Economics in 2015. The median confidence interval on return on investment was over 100 percentage points wide. Individual-level sales are so volatile relative to the per-capita cost of advertising that a coefficient of variation of 10 is common, and an informative experiment can easily require more than ten million person-weeks.

Their conclusion is not that measurement is pointless. It is that randomised experiments inject genuinely new information, and that selection bias in observational methods — because advertising is targeted at people already more likely to buy — is, in their words, a crippling concern.

Two later studies put numbers on that.

Blake, Nosko and Tadelis ran large-scale paid-search experiments at eBay and published them in Econometrica in 2015. Returns from paid search came out as a fraction of the conventional non-experimental estimates. In the extreme case, brand-keyword ads had no measurable short-term benefit — the people clicking them were largely people who would have arrived anyway. For non-brand keywords, new and infrequent users were positively influenced, but frequent users who were not influenced accounted for most of the spend, producing negative average returns.

Gordon, Zettelmeyer, Bhargava and Chapsky then compared experimental and observational estimates directly, using 15 US advertising experiments at Facebook covering roughly 500 million user-experiment observations and 1.6 billion ad impressions (Marketing Science, 2019). Their finding, stated plainly in the abstract: observational methods often fail to accurately recover the treatment effects generated by randomised advertising experiments.

If you take one thing from this section: an attribution report is not an experiment, and the gap between the two is not small.

4. The long-and-short balance is the most durable planning finding

Les Binet and Peter Field's The Long and the Short of It, published by the IPA in 2013 on the basis of the IPA Databank case studies, is the source of the widely repeated 60:40 rule — roughly 60% of budget to brand building, 40% to sales activation.

The IPA's own account of the work notes that the follow-up study Effectiveness in Context put the optimum at 62:38, and that across both reports it was the balance between long-term and short-term communication that mattered, with the sweet spot staying broadly stable despite a decade of media change.

Worth being precise about what this is and is not: it is an analysis of submitted, self-selected effectiveness case studies, not a randomised experiment. It is strong evidence about what distinguishes campaigns that reported success, and weaker evidence about causation than the field experiments above. Both facts can be true at once.

5. Attention is won by specific, boring craft decisions

Pieters and Wedel analysed 1,363 print advertisements using infrared eye-tracking with more than 3,600 consumers (Journal of Marketing, 2004). Their findings are unusually concrete:

• The pictorial element captures attention superiorly, and does so largely independent of its size.
• The text element captures attention in direct proportion to its surface size.
• The brand element is the most effective at transferring attention to the other elements.
• Only increasing the text element's surface size produces a net gain in attention to the ad as a whole.

Teixeira, Wedel and Pieters looked at the opposite question — what makes people leave — using zapping data and eye-tracking across 31 commercials with nearly 2,000 participants (Marketing Science, 2010). Simple metrics of attention dispersion strongly predict avoidance. Independently of that, central on-screen brand positions promoted avoidance, while brand size did not. Their model-based recommendation was brand pulsing: distributing the same total brand exposure across the spot rather than parking it in the middle of the frame.

Both papers point the same way. Attention is not bought by being louder. It is lost by being visually scattered, and it is spent by the audience on the elements that reward looking.

6. Humour: helpful, but not where people assume

Eisend's meta-analysis pooled 369 correlations on humour in advertising (Journal of the Academy of Marketing Science, 2009).

Humour significantly enhances attitude toward the ad, attention, and positive affect. It also significantly enhances brand attitude and purchase intention. But two of the findings cut against the received wisdom: there is no evidence that humour affects positive or negative cognitions, nor liking of the advertiser, and humour significantly reduces source credibility. The decline from lower-order to higher-order communication effects is steep — the effect on attitude toward the ad is about twice the size of the effect on brand attitude.

So humour reliably buys you attention and warmth toward the ad. It does not reliably buy you belief, and it costs you some credibility on the way.

What this adds up to

Read together, the literature is more modest and more useful than the industry's usual framing:

• Typical effects are small, and vary hugely by category and life-cycle stage.
• Weight alone rarely moves sales; the creative and strategic choices do more.
• Observational measurement systematically overstates. Experiments are the only reliable arbiter, and they are expensive.
• Balance between brand building and activation beats going all-in on either.
• Attention is a craft problem with specific, testable answers.
• Devices like humour have real but narrower effects than folklore suggests.

How to read the next statistic you are shown

Most of the "trend reports" circulating in this industry are not research. A short filter, which you are welcome to apply to this article too:

1. Is there a paper, or only a number? A figure with no traceable source is not evidence.
2. Who was in the sample, and who selected them? Self-submitted award case studies and self-selected survey panels answer different questions than a randomised experiment.
3. Is it experimental or observational? After Lewis and Rao and Gordon et al., this is the single most important question about any effectiveness claim.
4. Is the effect size reported, or only the direction? "Improves engagement" with no magnitude is unfalsifiable.
5. Does the cited source actually say that? Follow the link. A surprising share of the time it does not.

Sources

All of the following resolve by DOI.

1. Sethuraman, R., Tellis, G. J., & Briesch, R. A. (2011). How Well Does Advertising Work? Generalizations from Meta-Analysis of Brand Advertising Elasticities. Journal of Marketing Research, 48(3), 457–471. doi.org/10.1509/jmkr.48.3.457
2. Lodish, L. M., Abraham, M., Kalmenson, S., Livelsberger, J., Lubetkin, B., Richardson, B., & Stevens, M. E. (1995). How T.V. Advertising Works: A Meta-Analysis of 389 Real World Split Cable T.V. Advertising Experiments. Journal of Marketing Research, 32(2), 125–139. doi.org/10.1177/002224379503200201
3. Lewis, R. A., & Rao, J. M. (2015). The Unfavorable Economics of Measuring the Returns to Advertising. The Quarterly Journal of Economics, 130(4), 1941–1973. doi.org/10.1093/qje/qjv023
4. Blake, T., Nosko, C., & Tadelis, S. (2015). Consumer Heterogeneity and Paid Search Effectiveness: A Large-Scale Field Experiment. Econometrica, 83(1), 155–174. doi.org/10.3982/ECTA12423
5. Gordon, B. R., Zettelmeyer, F., Bhargava, N., & Chapsky, D. (2019). A Comparison of Approaches to Advertising Measurement: Evidence from Big Field Experiments at Facebook. Marketing Science, 38(2), 193–225. doi.org/10.1287/mksc.2018.1135
6. Binet, L., & Field, P. (2013). The Long and the Short of It. IPA. See the IPA's own summary of the 60:40 finding and its 62:38 successor: ipa.co.uk — The next chapter for 'The Long and The Short of It'
7. Pieters, R., & Wedel, M. (2004). Attention Capture and Transfer in Advertising: Brand, Pictorial, and Text-Size Effects. Journal of Marketing, 68(2), 36–50. doi.org/10.1509/jmkg.68.2.36.27794
8. Teixeira, T. S., Wedel, M., & Pieters, R. (2010). Moment-to-Moment Optimal Branding in TV Commercials: Preventing Avoidance by Pulsing. Marketing Science, 29(5), 783–804. doi.org/10.1287/mksc.1100.0567
9. Eisend, M. (2009). A meta-analysis of humor in advertising. Journal of the Academy of Marketing Science, 37(2), 191–203. doi.org/10.1007/s11747-008-0096-y

If we ever publish figures drawn from RateAds itself, they will come with the sample, the method, and the caveats attached — the same standard we just applied to everyone else.