A halo effect, in advertising measurement, is the share of a campaign's impact that lands somewhere the campaign was not aimed: on other products from the same brand, on other media channels, or on sales at other shops and tills. An ad for one shampoo that lifts sales of the same brand's conditioner is a halo. So is a television campaign that sends people to type the brand name into a search engine, where a paid search ad collects the credit. The idea exists because most reporting gives each campaign credit only for the outcome it was built to drive, and anything that leaks sideways goes uncounted or is claimed by another channel.

Three kinds of spillover

Measurement practice recognises three broad forms. The first is the product or brand halo, where advertising one item lifts sales of related items. Retail media networks have turned this into a reporting line. Amazon Ads' attribution guide defines brand halo metrics as measuring "the impact of advertising on the overall brand by tracking conversions", and states that they "exclude the promoted products". Detail page views, add-to-cart events and purchases each carry a separate brand halo count, abbreviated BH, which is added to the promoted-product figure to give a total.

What counts as related is a choice, not a fact. IAB Europe's Retail Media Measurement Standards, published as V1 in April 2024, set the default: "Same Brand, Same Category", with category defined by the retailer's own catalogue. The document also anticipates brands one day setting their own definitions with each retailer. The US guidelines from the Interactive Advertising Bureau (IAB) and the Media Rating Council (MRC), dated January 2024, go no further than saying that reporting of halo products "is encouraged to be transparent".

The second form is the channel halo. Here one medium raises the measured performance of another, most often upper-funnel video or television lifting branded search, social or direct visits. The third, the retailer or offline halo, covers purchases that follow an ad on one retailer's site but happen at a competing retailer or in a physical store, where the network that ran the ad cannot see them.

How halos are measured

Each form calls for a different tool. Product halos are usually measured by attribution: the retailer counts purchases of any same-brand, same-category item by exposed shoppers within a set lookback window. That approach is cheap but correlational: it shows what exposed shoppers bought, not what the ad caused.

Channel and retailer halos need causal designs. In an incrementality test, a control group or region is withheld from the campaign, and the analyst compares outcomes in every channel or store that data can reach, not only the one the ad pointed to.

Marketing mix models take a third route, estimating each channel's contribution from aggregate time series. Halos complicate them. If television drives search queries and search ads then capture those queries, a model that treats search spend as independent will over-credit search. Google's Meridian documentation calls Google Query Volume (GQV) "an important confounder between media and sales" and warns that failing to control for it "can lead to overestimation of the causal effect of paid search". A newer two-stage "full-funnel" approach, described on a Meridian page last updated on September 30, 2026, first models how brand media builds brand equity, then carries that "indirect effect" into return and budget calculations. Modellers can also add interaction terms between channels, though Recast, a modelling vendor, warned in February 2024 that modelling every possible interaction produces a "combinatorial explosion" of parameters.

From rating error to marketing metric

The word comes from psychology. Edward Thorndike, of Teachers College, Columbia University, described "the constant error of the 'halo'" in the Journal of Applied Psychology in 1920, after finding that a flight commander's ratings of 137 aviation cadets on intelligence, physique, leadership and character correlated far more closely than the traits plausibly could. Marketers borrowed the image.

Business usage hardened in the 2000s. Subramanian Balachander and Sanjoy Ghose, writing in the Journal of Marketing in 2003, showed that advertising a brand extension could raise choice of the parent brand, a "reciprocal spillover". Apple supplied the decade's best-known example. In April 2005, after Mac revenue rose 29% year on year, Merrill Lynch analyst Steven Milunovich claimed "quantitative proof of the halo effect" from the iPod, according to MacTech.

Field experiments then showed that halos do not always fall on the advertiser. Randall Lewis and Dan Nguyen, studying randomised display ads across roughly 500 million Yahoo homepage visits, found in Quantitative Marketing and Economics in 2015 that searches for advertised brands rose by about 30% to 45%, while incremental searches for competitors totalled two to eight times that increase. Retail media formalised the metric from 2024, when the IAB and MRC guidelines and IAB Europe's standards addressed halo reporting for the first time.

Why marketers argue about it

Halos move budgets. A channel judged only on its direct return may look weak while it feeds sales recorded elsewhere, and cutting it can depress results in the channels that appeared to perform. The stakes are clearest in retail media. A vendor study of more than 150,000 campaigns and $350 million in spend across four networks, released on June 15, 2026, estimated that siloed attribution misses 36% to 53% of retail media impact. The shortfall rose to 67% to 80% for off-site video, and 62% of that format's impact occurred away from the advertising retailer.

Platforms competing for budget have adopted the term as a sales argument. Reddit said in May 2026 that 13% of its performance impact improved the effectiveness of other channels, citing a TransUnion mix-model meta-analysis it commissioned. Snap cited a Measured study whose title says Snapchat lifts search and social incremental return on ad spend (iROAS) by 12.9%, which PPC Land read as a halo claim. A Lifesight report written with four connected TV (CTV) sellers put the lift in paid search conversions from paired CTV exposure at 22.3%, without disclosing sample or method.

The concept also connects to longer-run thinking. Brand advertising's slow effect on brand salience, and the brand equity it builds, is often the mechanism behind channel halos, which is why evidence that brand marketing drives up to six times the long-term sales impact of short-term activity is cited in the same debates.

Limitations and disputes

The first problem is definition. Same brand, same category is a convention. A retailer that widens it to the whole brand catalogue will report a larger halo from the same campaign, and the IAB and MRC guidelines encourage, rather than require, disclosure of which products were counted.

Second, attributed halo sales are not incremental. A shopper who saw an ad for one yoghurt and later bought the same brand's milk may have bought the milk anyway. Halo totals built on exposure-based attribution inherit every weakness of view-through crediting.

Third, spillover can help rivals. Navdeep Sahni of Stanford reported in the Journal of Marketing Research in 2016, from 11 randomised experiments covering 189,650 users of an Indian restaurant-search site, that one restaurant's ads raised individual competitors' sales leads by an average of 4%, and that the combined gain to competitors was about five times the advertiser's own. "To convert an impression into a sale, you have to beat the competitors that your ad reminds people of," Sahni said, according to Stanford Graduate School of Business. Most halo reports count only the advertiser's own brand, so this leakage is invisible.

Fourth, halos can be negative, can be double counted when two channels each claim the same sale, and can be imagined where none exists. A halo estimated by a mix model inherits the model's identification problems, including the bias that leads such models to overstate paid search returns by about 2.5 times in one simulation. And many published halo figures come from sellers or their commissioned partners, a conflict PPC Land has examined in platform-built mix models.

Not to be confused with

  • The psychological halo effect. Thorndike's 1920 term for a cognitive bias in which one positive trait colours judgements of unrelated traits. Richard Nisbett and Timothy Wilson showed in 1977 that people are often unaware of it. The reverse bias is sometimes called the horns effect. Marketing borrowed the metaphor; the measurement sense concerns sales and channels, not perception.
  • Halo, the measurement framework. The World Federation of Advertisers' open-source cross-media framework, which underpins ISBA's Origin programme, shares the name but measures deduplicated reach, not spillover.
  • Brand lift. Brand lift measures shifts in awareness or consideration through surveys. A halo concerns outcomes outside the campaign's target, whatever the method.
  • Cannibalisation. The opposite movement: promoting one product takes sales from a sibling product or from organic demand that would have converted anyway.

Recent developments

Measurement products have moved towards counting off-platform effects. On July 7, 2026, Amazon added retailer-level breakouts to its Omnichannel Metrics, so that Amazon DSP advertisers can see which named retailers received offline purchases after exposure, a step from a single "offline" figure towards the retailer halo. LiveRamp, when it extended its clean room in October 2025 to let retail media networks measure Meta campaigns against their own sales data, listed halo measurement among planned features.

The counter-pressure comes from causal measurement. The IAB argued in April 2026 that legacy mix models undervalue retail media and that controlled experiments should lead. Experiments, by design, catch halos only where outcome data reaches, which is why cross-retailer and in-store coverage remains, as of October 2026, the main gap in halo measurement.

Timeline

  • 1920: Edward Thorndike names the "halo" error in ratings in the Journal of Applied Psychology
  • 1977: Richard Nisbett and Timothy Wilson publish evidence that the halo effect operates without awareness
  • 2003: Subramanian Balachander and Sanjoy Ghose report reciprocal spillover from brand extension advertising to parent brands
  • April 2005: Analysts credit an iPod "halo effect" for a 29% rise in Apple Mac revenue
  • 2015: Randall Lewis and Dan Nguyen show Yahoo display ads lift competitor searches by two to eight times the advertiser's gain
  • 2016: Navdeep Sahni publishes restaurant-platform experiments showing ad spillovers to competitors
  • January 2024: IAB and MRC Retail Media Measurement Guidelines encourage transparent reporting of halo products
  • February 2024: Recast warns that modelling every channel interaction causes a combinatorial explosion of parameters
  • April 2024: IAB Europe's standards define halo as same brand, same category
  • October 2025: LiveRamp lists halo measurement among planned clean room features for retail media networks
  • May 27, 2026: Reddit says 13% of its performance impact improves other channels
  • June 15, 2026: A vendor study estimates siloed retail media attribution misses 36% to 53% of impact
  • July 7, 2026: Amazon adds retailer-level offline breakouts to Omnichannel Metrics
  • August 12, 2026: Lifesight reports a 22.3% paid search conversion lift from paired CTV exposure
  • September 30, 2026: Snap cites a Measured study on Snapchat lifting search and social iROAS; Google's Meridian full-funnel modelling page is updated

Summary

Who: Advertisers and agencies that allocate budgets across channels; retail media networks such as Amazon that report brand halo sales; measurement vendors, modellers and platforms such as Reddit and Snap that cite halo figures; standards bodies including the IAB, MRC and IAB Europe.

What: The spillover of a campaign's effect onto products, channels or retailers it did not directly target, reported through attribution, estimated through experiments, or modelled in marketing mix models.

When: The term dates to Thorndike's 1920 psychology paper; marketing research on spillover developed from the 2000s, retail media standards addressed halo reporting in 2024, and halo claims became a common platform sales argument through 2026.

Where: In retail media reporting consoles, clean rooms, incrementality tests and mix models, wherever outcomes from more than one product, channel or store can be observed together.

Why: Channels judged only on their direct return can be undervalued or cut, but halo figures depend on definitions, attribution windows and models that the sellers reporting them often control, and spillover can benefit competitors as much as the advertiser.