Does TV drive more reach and higher brand recall than YouTube ads?

We need more head-to-head comparisons of YouTube advertising versus TV advertising, using single-source sales as the dependent variable. But in one of the Institute’s Beyond :30 studies YouTube exposures had higher rates of brand recall than linear TV ads and Facebook ads (Bellman et al., 2023). Whether to buy YouTube ads or not depends on their unduplicated reach, compared with TV (Romaniuk et al., 2013). If TV is reaching 80% or more of all category buyers, there would be no need to buy YouTube ads. 

References

• Bellman, S., Rask, A., & Varan, D. (2023). Is news a good context for advertising? Beyond :30 Study 88
• Romaniuk, J., Beal, V., Uncles, M. (2013). Achieving reach in a multi-media environment: How a marketer’s first step provides the direction for the second. Journal of Advertising Research, 53(2), 221-230 

S.B.
14 March 2024

What type of ads decay slower?

If we assume that memory generally decays at the same rate (Rubin & Wenzel, 1996), then the question is: what level of memory does the advertiser want? Visual recognition declines from the highest base, almost 100%, and so after a week, it can still be higher than 90% (Venkatraman et al., 2021). Visual recognition is stronger than verbal recognition (e.g., Haber, 1970; Snodgrass et al., 1972) or any form of recall. If the purpose of the ad is to promote visual pack recognition in the store, you shouldn’t have to advertise as often compared to if the purpose was to have people recall a brand name before visiting the store or searching online.

Do interactive experiences decay slower than passively received advertising?

Again, most evidence suggests memory decays at the same rate, and the important question is what base rate does it decline from. Different types of interactive experiences create different levels of immediate and delayed memory. A click that results in viewing a video ad several minutes long creates a level of day-after recalling equivalent to seeing three repeated exposures to a normal 30-second TV ad (Bellman et al., 2009). On the other hand, a fleeting one-click-and-done interaction creates a level of recall no higher than just seeing the ad, without interacting, or seeing a normal, non-interactive version of the ad (Bellman et al., 2009). In another, unpublished study (Bellman et al., 2005), both day-after and week-after recall were measured to gauge the rate of decay for interactive ads. The decay rate (day-after minus week-after recall) was actually higher for interactive ads in that study, and again depended on whether the interaction was short (impulse) or long (dedicated advertiser location [DAL]) (normal control ads decay = 2% vs impulse = 8% and DAL = 9%). So, the Institute has no evidence that interactive experiences reduce the rate of decay, but a longer interactive experience creates a higher base to decay from.

What can you recommend to create ads or experiences that decay more slowly?

Again, assuming that memory decays at the same rate, the important thing is what level the decay starts from. Memory level depends on node strength and association strength (Anderson et al., 1998), and both of those are likely to be enhanced by rehearsal over longer or repeated experiences. It is hard to get consumers to pay attention for long periods of time, but media (e.g., cinema), creativity, and interactivity can all be used to encourage longer exposure times. Note that the interactive-ad studies above controlled for the self-selection effect of interactivity as opposed to its incremental effect. The self-selection effect selects out the consumers who were already going to buy the brand. Since the job of advertising has already been done for these people, paying for additional advertising that targets these people is a waste of money (Blake et al., 2015). When you control for this self-selection effect to measure the incremental effect of interaction on interactors, the apparent effect of interaction can disappear (Bellman & Varan, 2012).

Do ads/experiences that decay more slowly need less topping up?

If the goal is to maintain greater than 50% visual recognition for the pack (shown in the ad) among all category buyers, then less topping up might be needed (e.g., 1+ frequency per month). To achieve the same level of brand recall (from a category need cue), among all category buyers, would generally need more exposures per month (3+?), as recall is a harder task (Rossiter & Percy, 2017).

References

• Anderson, J. R., Bothell, D., Lebiere, C., & Matessa, M. (1998). An integrated theory of list memory. Journal of Memory and Language, 38(4), 341-380
• Bellman, S. & Varan, D. (2012). Modeling self-selection bias in interactive-communications research. Communication Methods and Measures, 6(3), 163-189
• Bellman, S., Schweda, A., & Varan, D. (2005). Effects of Interactive TV Clutter (Beyond :30 Report: Study 6)
• Bellman, S., Schweda, A., & Varan, D. (2009). A comparison of three interactive television ad formats. Journal of Interactive Advertising, 10(1), 14-34
• Blake, T., Nosko, C., & Tadelis, S. (2015). Consumer heterogeneity and paid search effectiveness: A large-scale field experiment. Econometrica, 83(1), 155-174
• Haber, R. N. (1970). How we remember what we see. Scientific American, 222(5), 104-112
• Rossiter, J. R., & Percy, L. (2017). Methodological guidelines for advertising research. Journal of Advertising, 46(1), 71-82
• Rubin, D. C., & Wenzel, A. E. (1996). One hundred years of forgetting: A quantitative description of retention. Psychological Review, 103(4), 734-760
• Snodgrass, J. G., Volvovitz, R., & Walfish, E R (1972). Recognition memory for words, pictures, and words+ pictures. Psychonomic Science, 27(6), 345-347
• Venkatraman, V., Dimoka, A., Vo, K., & Pavlou, P. A. (2021). Relative effectiveness of print and digital advertising: A memory perspective. Journal of Marketing Research, 58(5), 827-844

S.B.
5 March 2024

What is the best practice on frequency for media touch-points? For something like TV vs social vs programmatic, what kind of frequency capping should we follow.

There is some generalised evidence around the potential impacts of exposure frequency (regardless of whether that is in a single media or across touch points) that will be helpful here.

The idea that there was a need to be exposed to advertising messages with frequency came from the psychology literature originally, suggesting that in order to learn something people need to be exposed to it multiple times.

In media, this was called Effective Frequency and a number of 3+ exposures was thought to be the threshold. However, evidence that buyers have to “learn” in order to buy brands, is lacking. To the contrary, Single Source analysis shows that the greatest uplift in sales impacts of advertising typically occurs from the first exposure in a window (i.e., 1+ reach in a month), with diminishing returns from subsequent exposures (this is shown in the chart below).

As a result, we recommend that you worry less about reaching a potential buyer multiple times across touch-points and more about reaching additional potential buyers. This will give you the greatest potential for aggregate level advertising impacts.

Some additional considerations:

Firstly, in most markets, we don’t know when a particular individual will buy, because buying behaviour is heterogeneous and ad impacts decay, so we can’t just advertise once and never again. Our advertising needs to be continuous (i.e., spread across time), in order to reach as many potential buyers as possible as often as possible. Exposures (frequency) that is spread over time is continuity. Exposures which are spread are more valuable than those that are burst (more chance to intercept as if random buying, get light media users and improve memory effects).

Campaigns inevitably results in some frequency for those who do see our communications (some people are very heavy media users compared to others). Aim is to minimise excess frequency where ever possible. Using big media, popular programs etc. can help.

Secondly, not all reach is equal. As you point out, there are different touch-points, and a touch-point that is closer to the purchase may be more impactful than touch-points that are further away from the purchase, because we do know that recency of exposure helps with impact (Report 57: A Guide to Continuous-Reach Advertising). However, there is an important distinction to be made here between advertising (which helps build and refresh probabilities of purchase (i.e., Mental Availability)) and communications that help make the most of a buyer who has already decided to buy. Activations are more about ‘catching them when they fall’ in that the buyer has already recognised a need to buy and this may just be that final message to reinforce behaviour that was in all likelihood about to happen regardless. You need to be there at these times and touch-points, to the point where you get your fair share.

Our recommendation with your communications is to optimise unique, unduplicated reach as much as possible. This sometimes means adding touch-points to your campaign, advertising across media/platforms, and in turn delivering some exposure frequency as audiences who are exposed at one touch-point, also see the advertising at another point. When planning to maximise overall reach, start with the touchpoint or media that delivers the most potential categories buyers, adding to that with the objective of reaching unique audiences (potential category buyers) as much as possible (references below). Adding digital to a TV campaign, for example, may be better than adding two channels of TV if that gives you access to an audience not available on TV.

So directly to your question, in relevant media cap at 1 OTS if scope to still exists to build unique reach for the relevant window. If budgets allow, more exposures are better than less but with diminishing returns and spreading those exposures is better than bunching them (especially if it means you go dark later).

Achieving Reach in a Multi-Media Environment. How a Marketer’s First Step Provides the Direction for the Second

Cross Platform Frequency (Beyond :30 Report)

In summary, you should be aiming to maximise reach (spread exposures across buyers and across time) for all campaigns, as that will be optimal for impact.

E.R.
7 February 2024

I see and hear a lot of generalist talk about how Gen Y & Z are way more engaged with brands and their content via social media.

Has the Institute conducted research on variations in brand engagement (in terms of loyalty, media consumption, buying behaviour) across different life stages?

Some facts and figures from the Institute itself were published in a Journal of Advertising Research article a few years ago, but are still relevant today (Barwise et al., 2020). In that article, we obtained Nielsen data in 2017 to compare younger television viewers (aged 18 to 34 years) with older viewers (35 to 54 years). Those born from 1981 to 1996 (aged 21 to 35 in 2017) would be Gen Y (Millennials) and those born later (aged 18 to 20 in 2017) would be Gen Alpha. These younger television viewers were different from the older viewers because they watched less live TV and more online TV (via smart TVs and connected-TV devices, or other online devices, such as PCs and smartphones). Compared with 18-34 year olds in 1992, 25 years earlier, the 2017 18-34 year olds watched slightly less TV (live + online) per week (29.1 hours per week vs 30.1 hours in 1992, a 3.6% reduction). In contrast, the older group (35-54) had increased their TV hours (live + online) per week (from 33.3 hours in 1992 to 38.8 in 2017, a 17.5% increase). We can probably assume that 18-34 year olds in 2017 dedicated as much of their week to media consumption as 35-54 year olds, and that the nearly 10 hours difference was absorbed by consuming social media and other alternatives to television.

This suggests that advertisers need to use other media in addition to television to reach Gen Y and Gen Alpha, such as social media, radio, and out-of-home.

Which social media to use to target younger people changes when they migrate to alternatives less likely to be used by their parents. Currently, a Statista report from October 2023 shows that most (68%) of the audience for TikTok was aged 18 to 34. Only 25% was aged 35 to 54.

These numbers are about media usage, and useful for deciding which media to use to reach consumers of different ages with advertising messages. Another of our Institute articles suggests there are no differences between age groups in brand usage (Anesbury et al., 2022), which should be the main metric for brands to consider.

References:

• Anesbury, Z., Bellman, S., Driesener, C., Page, B. and Sharp, B. (2022). Ageism Kills Brands. Australasian Marketing Journal, 30(4), 364-370.
• Barwise, P., Bellman, S. and Beal, V. (2020). Why Do People Watch So Much Television and Video? Implications for the Future of Viewing and Advertising. Journal of Advertising Research, 60(2), 121-134.

S.B.
11 January 2024

Would you have any practical guidance on how to create effective video? (level of branding, testimonial, narrative, problem/solution, etc.)

I wouldn’t say “rules” so much as there are tactics we have found that have increased odds of success (where the outcome is short term sales effectiveness), such as negative framing, humour appeals, speaking characters, sensory experiences, and some form of animation/fantasy. Here is a key article: Creative That Sells: How Advertising Execution Affects Sales.

Regarding branding tactics, execution principles for increasing the likelihood of brand recall are to have visual frequency, being dual mode (visual and verbal mentions), and introducing the brand earlier in the ad, here are some sources:

• Report 103: Harnessing the Power of Audio Branding

• Article The efficacy of brand-execution tactics in TV advertising, brand placements, and internet advertising

N.H.
10 May 2023

What would be Ehrenberg-Bass Institute’s view on the value of product placement?

At the Institute, we have carried out many studies of product placement effects. Product placement is attractive because television ad-avoidance is perceived to be high, and product placement is the only way of advertising to viewers of ad-free subscription video networks. However, like many things, product placement works best in moderation. Product placements in entertainment content need to be subtle, so that they don’t trigger a negative emotional response. With informational content (such as a reality cooking show), the audience is more forgiving, because some of the information conveyed by these types of shows includes recommendations about brands of products. But even with informational shows, a brand can do too much product placement and sponsorship (e.g., combining blatant product placement with sponsorship billboards and commercials). This can backfire on brand choice. One way to have very high levels of product placement is to make a special program all about the product category, but which does not mention the brand. This was effective for the Chipotle brand in the United States, but is also very risky.

Product placement needs to be considered in the context of the brand’s overall media strategy. The most important objective of a media strategy is to reach all category buyers with at least one noticed exposure. If you have product placement in a single movie or series, that puts a limit on the amount of reach you can achieve. On the other hand, product placement may be a good complementary medium, for reaching the category buyers not reached by television and radio commercials. And if the program is a series, the series will accumulate more reach over time. One thing to be wary of is that sometimes product placement is offered as part of a package deal along with TV commercials and sponsorship bumpers. If your brand is new and needs high frequency (during the same program) to be remembered, these package deals might be worth considering. But for established brands, it is better to spend the money for a second exposure on a first exposure to a different consumer. Another thing to consider is the quality of the exposure. The quality doesn’t have to be high, but subtle placements can be too fleeting and not get noticed at all. It would have been better to spend the money on advertising instead of an unnoticed product placement. Subtle placements rely on reminding people about brands they have learned about through other means, mainly advertising. So unless you are continuing to advertise, people won’t be reminded about the product. Double-check your perceptions of how many viewers are avoiding commercial breaks. Television’s research, using cameras in living rooms, shows that commercial zapping is not that high. Avoidance by leaving the room is higher and that affects programs as well. And finally, the size of the audience for ad-free networks is not that large. Advertising still reaches most category buyers.

Here is a selection of our research on this topic:

• Sharp, B., Newstead, K., Beal, V., Tanusondjaja, A., Kennedy, R., & Danenberg, N. (2014). Best Practice: Ehrenberg-Bass Institute Key Media Principles. Ehrenberg-Bass Institute report 66.
• Bellman, S., Rask, A., & Varan, D. (2019). How Chipotle used unbranded content to increase purchase intention by changing beliefs about ethical consumption. Journal of Marketing Communications, 25(7), 763-782.

• Bellman, S., Murphy, J., Arismendez, S. V., & Varan, D. (2019). How TV sponsorship can help television spot advertising. European Journal of Marketing, 53(1), 121-136.
• Corkindale, D., Neale, M., & Bellman, S. (2023). Product placement and integrated marketing communications effects on an informational TV program. Journal of Advertising, 52(1), 75-93.
• Neale, M., & Corkindale, D. (2022). Make product placement work for you: Get less exposure. Business Horizons, 65(2), 149-157.

S.B.
2 May 2023

I was wondering if the Institute had done research / had a perspective on the testing of television commercials (TVCs): IPSOS Copy Effect Index scores a KPI for many companies, but the anomalies in scoring raises many doubts. System 1 is an interesting alternative, but also question marks over veracity of results. With many companies unable to support the gold standard of single source testing, I wonder what the recommendations on testing TVCs are from the Ehrenberg-Bass Institute.

Yes, the Institute has done research on the testing of TVCs. For example, we tested over 100 Mars ads with single-source sales data using a variety of measures including memory, liking, and biometrics, including facial expression (Bellman et al., 2017). The conclusion was that these measures do significantly better, but not substantially better than chance (they are right 2 times out of 3). They may be helpful if a management judgement test might be incorrect, particularly when judging whether consumers will laugh at an ad designed to be funny. But in general, the Institute’s view is that if you are using a very good creative agency, then the agency should be free to develop the ad, within the parameters of the brief about what Category Entry Point(s) to mention, and how much branding there should be throughout the ad. Instead of pretesting, we advocate post-testing, and withdrawing new ads that are not working in post-tests. This policy tends to retain good ads on air until they really wear out, rather than withdrawing them just because the brand has a new CMO. If you do need to do pretesting, and do not have access to facial expression testing using labs or webcams, our research showed that expert coding can be a good substitute for testing with consumers. My understanding is that System 1 uses an expert-coder method, and their philosophy favours humorous ads like they used to do in the UK in 1980s. I’m not sure what goes into the IPSOS copy effect index score, but in general, it is hard to believe that a single number can apply across ads with various strategies (e.g., purchase intention for impulse products, ad liking for fear ads, etc.).

See journal articles:

• What Makes a Television Commercial Sell? Using Biometrics to Identify Successful Ads
• Finding Creative Drivers of Advertising Effectiveness with Modern Data Analysis

S.B.
5 April 2023

Is there any rule or threshold on minimum media weights/GRPs per week or month? I am often told to run minimum GRPs on TV in a particular period, which effectively means, I can’t spread out my media spend.

No. Great ads nudge the propensities of those exposed to buy from a single exposure.

This “minimum GRPs” idea is often supported by agencies who need to be able to demonstrate effectiveness when they are reliant on tracking data. Specifically tracking of GRPs against advertising awareness metrics; where changes in advertising awareness are only able to be seen where there is lots of variation e.g. large bursts (heavy GRPs, short timeframe) then silence.

There is, however, no evidence to suggest that advertising below a certain weight/level of GRPs per week is ineffective where you have access to the likes of single source data. For any given budget, the number of weeks with GRPs is more important than the number of GRPs per week.

K.V, N.H. & R.K.
15 November 2022

For a 15″ ad, is there any difference between branding and message results? For example, an established brand launches something completely new in the category…would the 15″ still be as effective?

We didn’t specifically look at ‘message take out’ in the study looking at 30 vs 15 second ads. Rather it focused on ad recall and correct branding score where 15 second ads were found to have recall and likability scores 80% of the 30 second making them good value if you are paying 80% (or less) the cost of a 30 second slot.

If something is new to market (or the message is new) then it takes more effort to ‘teach’ consumers something new rather than just refreshing existing memory structures.

Here is the link for the full report: In praise of the 15 second advertisement

V.B.
08 June 2021

What are the most and least important metrics for Advertising Tracking on TV?

We are limiting our answer to in-market post-testing, where consumers are exposed to ads in the real-world. The most important measure is purchase/sales, collected via single-source data or in-market experiments with matched markets/regions. We understand this kind of measurement is not feasible for all advertisers in all markets. In terms of survey research, advertisers should be looking to measure reach achieved. That is, exposures rather than opportunities-to-see. Ad recognition is a measure least biased by prior brand usage (see Report 87), so is a simple measure to collect/report. There can be some overclaiming, but there are ways and means to lessen this. Correct brand linkage is particularly important, because if the ad is not linked to the advertised brand in memory, then the ad will have little to no impact on nudging the propensities to buy the brand. From our observations, this is the area where ads tend to perform most poorly. Prior research has signalled that likeability have a positive correlation with performance, so this is another measure worth collecting.

K.V. & N.H.
21 April 2021