Home / article

The Effectiveness Canon: What the Evidence Actually Says

Every primary source the marketing-effectiveness argument rests on: what each study found, what it does not claim, and where its authority stops.

TL;DR. Almost every argument about marketing effectiveness traces back to a handful of primary sources: the IPA Effectiveness Databank, the Ehrenberg-Bass Institute's panel research, WARC's benchmarks, and a small number of measurement studies. This page lists them, says what each one actually found, and — more usefully — what each one does not claim. Most citation lists stop at the first part.

Marketing effectiveness has an unusually good evidence base, and an unusually bad habit of quoting it badly. A number gets lifted from a report, repeated without its conditions, and within a year it is being used to justify the opposite of what the research found. The 60:40 split is not a law. Double jeopardy is not a rule about loyalty programmes. A multiplier from one study is not a multiplier from another.

What follows is the canon: the sources serious practitioners keep returning to, what each contributes, and where its authority stops. Every entry links to the primary source where one is publicly available, and to the page on this site that goes into more detail.

The IPA Databank and Binet & Field

The Institute of Practitioners in Advertising runs the Effectiveness Awards. Entrants must prove commercial return with real data — econometrics, controlled testing, or robust tracking — which makes the resulting Databank the largest archive of proven campaigns in existence. Les Binet and Peter Field spent two decades mining it.

WorkWhat it foundWhat it does not claim
Marketing in the Era of Accountability (2007)Separates short-term activation from long-term brand building, and shows the two produce different effects on different timescales.That every category splits its budget the same way, or that the split predicts individual campaign outcomes.
The Long and the Short of It (2013)Derives the widely quoted 60:40 benchmark, links emotional creative to long-term profit, and shows share of voice predicting share of market through excess share of voice.That 60:40 is optimal everywhere. It is an average across the Databank, and the authors say so.
Effectiveness in Context (2018)Revisits the split under promotional pressure, and argues the balance should move toward brand when short-term promotions erode pricing power.That the earlier findings were wrong; it refines them.

Related pages: the 60:40 rule · the full review of the 2013 report · excess share of voice.

The Ehrenberg-Bass Institute

Ehrenberg-Bass (University of South Australia) works from consumer panel data rather than award entries: who bought, how often, in which categories. That different source produces a different, largely complementary set of findings.

WorkWhat it foundWhat it does not claim
How Brands Grow — Sharp (2010)Growth comes from penetration, not loyalty. Double jeopardy holds across categories: smaller brands have fewer buyers who are also slightly less loyal.That loyalty is worthless. It is roughly constant across brands, which is why it is a weak growth lever, not a reason to ignore customers.
How Brands Grow 2 — Romaniuk & Sharp (2016)Category entry points and mental availability: brands grow by being linked to more buying situations, cued by distinctive assets.That positioning is useless. It argues distinctiveness beats differentiation as the mechanism of being chosen.
The Dirichlet model — Goodhardt, Ehrenberg & Chatfield (1984)Purchase patterns are predictable from market share: penetration, frequency and loyalty follow from brand size.That marketing cannot change anything. It describes the equilibrium that marketing pushes against.

Related pages: the How Brands Grow guide · double jeopardy · mental availability · the Dirichlet model.

WARC and the benchmark layer

WARC aggregates effectiveness case data globally and publishes benchmarks that let a practitioner compare a result to category norms — and stress-test the IPA findings outside the UK-centric Databank. Its Anatomy of Effectiveness (2019) is the clearest short account of the creative-effectiveness decline, and it is covered on this site separately: what the Anatomy of Effectiveness actually found.

The benchmark layer matters because the IPA data is British, award-entered and therefore self-selected. WARC's material is broader and less curated, which is exactly what you want when checking whether a pattern travels.

Measurement evidence

Effectiveness claims are only as good as the measurement behind them, and the three main methods answer different questions. Attribution describes which touchpoints preceded a conversion. Marketing mix modelling regresses aggregate sales on aggregate inputs. Incrementality testing holds a control group back and measures the difference.

Only the last one establishes causation, and the gap is not academic: a holdout can show that a fraction of the conversions your attribution dashboard claimed were genuinely incremental. MMM sees effects attribution cannot see, and it is slow and expensive. Attribution vs MMM vs incrementality sets out the trade-offs; incrementality testing covers the gold standard in detail.

What the canon does not settle

  • How the split varies by category. The 60:40 average is robust; the correct split for a low-penetration B2B brand or a declining category is a judgement the research does not make for you.
  • How much of the effect is creative versus media. The evidence says both matter and interact. Anyone quoting a single multiplier has usually collapsed two different comparisons — the Creative Dividend numbers are a case study in why that goes wrong.
  • What happens in zero-reach channels. The canon is built on mass-reach media. Its application to narrow, high-frequency digital buying is an extrapolation, not a finding.
  • How to weight short-term measurement when the long term is unmeasurable in-quarter. That is an organisational problem the evidence describes but cannot solve.

How to use this page

When someone quotes a number at you, ask which source, which comparison, and which conditions. If the answer is a blog post quoting a keynote quoting a report, the number is probably three removes from anything measurable. The sources above are all one remove, and each one states its own limits.

Where this site cites a figure, it follows the source's own framing and says so. Where it does not, the claim is marked as an argument rather than a finding.

Sources

  • Binet, L. & Field, P. (2007). Marketing in the Era of Accountability. IPA.
  • Binet, L. & Field, P. (2013). The Long and the Short of It: Balancing Short and Long-Term Marketing Strategies. IPA.
  • Binet, L. & Field, P. (2018). Effectiveness in Context. IPA.
  • Sharp, B. (2010). How Brands Grow. Oxford University Press.
  • Romaniuk, J. & Sharp, B. (2016). How Brands Grow 2. Oxford University Press.
  • Goodhardt, G., Ehrenberg, A. & Chatfield, C. (1984). The Dirichlet model of buying behaviour. Journal of the Royal Statistical Society.
  • WARC. Anatomy of Effectiveness (2019), with contributions from Peter Field, System1, the Ehrenberg-Bass Institute, Neuro-Insight and Kantar.
  • IPA Effectiveness Databank — the archive of award entries with proven commercial return.

Primary sources: IPA Effectiveness Databank · WARC · Ehrenberg-Bass Institute

Where the canon becomes practice: the complete guide to marketing effectiveness applies these sources to budgets, measurement and the mistakes that waste both.

Stay Boringly Effective

Get new episodes, articles, and marketing insights delivered to your inbox. No spam, just signal.

Prefer finding answers in Google? Add Boringly Effective as a Preferred Source in AI Mode — it takes about twenty seconds.