When it comes to activating media, how do we balance buying to a yearly reach (through targeting the market and focusing on high-reach media) versus ensuring we have continuous reach advertising (where Ehrenberg-Bass Institute I talk about minimising the gap between exposure and purchase)?

1. This is the article we are referring to when talking to Ehrenberg-Bass Institute as well – A Guide to Continuous-Reach Advertising 

2. Or is it more about Media Planning that maximises as much reach as possible and trying to translate that plan to real media buying practices? For example running planning scenarios for a TV budget that will get as much yearly reach as possible, but converting to Monthly Reach Goals to ensure the TV buying team have an even budget across each month?

Great question!!! How do you have your cake and eat it too? Let us start by commending you on appreciating that both reach, and continuity are desirable objectives for evidence-based media planning (as many marketers don’t). Even seeking to maximise reach and continuity will get you a long way towards the most effective media plans possible. Inherent in your question is an assumption that it is not possible to be both maximising reach and doing so continuously. I wonder, however, whether there are many scenarios where a continuous schedule is not also producing the best annual reach outcome? I’d be very interested in any examples of a lay-down you can provide that shows this outcome.

The only way that I can see this as a possibility is if there is some seasonality in either the availability of high reaching media, or in your sales. For instance, if for reach reasons, you chose to advertise in the AFL Grand Final, thus delivering a lot of reach at one point in time. Given you can’t buy AFL Grand Final level reach across the rest of the year, this inevitably means not being continuous (unless a combination of other media could deliver similar reach). But there are few such examples of media that cannot be bought more continuously.

Similarly, if there is considerable seasonality in the buying of the brand/product that you are advertising, then it may be sensible to allocate more advertising weight when more potential buyers are in the market. That said, remember too that advertising is about building memories to impact future buying behaviour (rather than looking to activate those already in the market to buy). And most categories are far less seasonal than we think.

To some extent what you are asking about here is also about the extent to which advertising memories decay.

Our work on what happens when brands go dark with advertising…When Brands Go Dark Extending the Evidence to 22 Categories in the United States

…shows that advertising memories (and even associated short term sales impacts) don’t immediately decay. For this reason, if I were forced to choose between reach and continuity, I’d favour reach maximisation, because the size of the nudge from advertising will likely outweigh the decline in propensity to buy that comes from being off-air. Obviously, you can’t stay dark forever and expect the brand to survive, but how long ad memories will last is dependent on many things (not least of which the size of the nudge to buying propensities that comes from a particular execution). This answer to another sponsor question may also help here: What is the maximum amount of time we should be off-air?

In short, your idea of running scenarios and testing your options for the best balance between reach and continuity effects is a fantastic one! We advocate experimentation wherever you can to determine the tipping point at which being more continuous is a more effective choice than to reach another potential buyer.

N.H and E.R
16 DECEMBER 2024

Do you have any meta or evidence-based learning that illustrates how the incremental sales amount changes when the media investment of a new brand launch is incremental vs. stealing from current portfolio media investment?

Firstly, we’ve attached a paper highlighting the challenges associated with using Media Mix Modelling (MMM) to inform media decision making.

Forecasting advertising and media effects on sales: Econometrics and alternatives

This is because, generally this is how managers look at the relationship between their media investment and their sales, and yet there can be some real issues associated with doing this. Notably, the typical outcome of an MMM is media ROI. Striving to optimise media ROI can lead to media decision-making focussed on efficiency, rather than effectiveness, that is decisions that run counter to what works for building brands.

At least superficially, any increased media investment should be beneficial and so arguably your proposal to spend incrementally on the new brand launch, rather than stealing investment from other brands in the portfolio, is the most sensible recommendation if you want the best outcome across the entire portfolio. But remember, spending more also doesn’t necessarily guarantee extra sales (or market share gains), especially in the short term as advertising effects are often felt in long term (Attached is a paper that investigates the relationship between budgeting and sales implications).

Advertising budgeting: a re-investigation of the evidence on brand size and spend

The work of these colleagues shows that while overspending (relative to your market share) doesn’t guarantee brand growth, brands that do grow are more likely to have spent over alignment with their share (SoV>SoM). Because advertising builds memories, we also wouldn’t recommend advertising the newly launched brand with heavy spend at launch if that means reducing spend significantly (or going dark) later.  Further information about our scheduling recommendations (for existing and new brands) can be found in our Key Media Principles report.

In terms of budget setting across the portfolio, the Advertising Intensiveness Curve a nice starting point for deciding on what you might spend on a new brand relative to established brands. It tells us that large brands can afford to spend more than their SoM and still maintain their market share, while smaller brands have to spend in excess of their market share in order to do so. I would recommend considering the existing circumstances of your wider portfolio of brands before deciding to “steal” their investment, as there will inevitably be some detrimental effects for those existing brands. Further, if your portfolio contains many SKUs that are competing in the same category (i.e. serving the same broad buyer need), a new entrant will inevitably cannibalise sales from other brands it competes with in line with their size (i.e. the Duplication of Purchase Law). If your own brands are these competitors, stealing media investment from them for the new brand will have an even more detrimental impact on the existing brands in your portfolio.

The Institute conducts research projects both in Budget Estimation and to Review Media decision-making more broadly if you are interested in digging deeper on this question of how to invest in the different brands in the portfolio including new brand launches.

E.R.

22 November 2024

What was the institutes position on in vitro testing of marketing mix elements. Packaging, Advertising, Pricing etc. are there any great new digital, artificial intelligence methods to gauge elasticities we should now about?

We are big fans of controlled experiments. Digital environments do allow for more testing (e.g. within single online retailers for pricing or pack and in digital media platforms for copy or media tests.

But not all experiments are equal. Ideal to have independent measurement. Multiple tests encouraged e.g. across retailers, media, time etc. Some issues in likes of social where you may be wanting to test one element e.g. copy or a target audience but the algorithms change and it alters factors that you think are controlled across your test cells. (There is a recent reading on this.. not sure of authors).

AI has some amazing applications especially related to efficiency. Not sure it is the tool that is going to deliver in this space where you really want the in-market response to the differences you are testing.

Learnings are very useful from controlled tests. Elasticities may not be the priority measure to focus on (potentially impacted by other factors unrelated to the test).

M.N, R.K, J.D & W.C

22 November 2024

I was wondering if you have seen any research on ‘in market’ ad recall testing? And what you would suggest to replace it?

I’ve seen lots of work on ad recall not being a valid measure because ads can successfully build memory structures without the ad itself being remembered. And I’ve seen a lot of work around the how brand size skews recall measures

I know EBI does Category Entry Point analysis and measurement. Is that a better approach to quantifying post impression impact?

Yes ad recall is not a great metric, mainly because most ad recall measures are not very specific and involve recall of the brand, not the ad (e.g., Have you seen advertising for any of the following brands in the last 3 months?  KOT). And yes because it often involves retrieval of the brand, there is a brand user bias, but it is important to note that this bias occurs even when there is no brand name retrieval involved (e.g., execution only retrieval), albeit at a lower level.  It’s actually about our selective attention, so something in the real world, and also in our measurement.

Recognition measures are better, funnily enough we were talking about these yesterday.  They are better than recall, but there will be instances where they might be problematic – e.g., if you have a very distinctive ad style, might be difficult for people to separate out executions etc.  But the issue is why do you want to check if people remember the ad?  Is it to check the reach (e.g., did if reach 45% of category buyers)?  Is it to check if a particular medium in reaching people (e.g., advertising on YouTube)?  And if people remember the creative, do they also remember the brand? (To get correct branding scores).

So these are more initial measures of the advertising having the underlying conditions for success.  They are useful to assess early to avoid spending too much money on a dud.  We are doing more R&D on this.

This is different from checking if the advertising did successfully build memories/mental availability.  For that you need to know what people took away from the ad (the brand, the message) and then was that processed in long term memory (which is what CEP measurement does).

My thoughts are to first check you have the conditions for success: branding, creative and messaging.  Then the distribution to get scale of success (reach).  This should be done before launch or just after, so you can adjust if below par.

Then you can measure the impact via brand memories (e.g., CEP measurement if your aim is to shift CEPs or build Mental Availability generally).

But I only recommend using post hoc ad memorability measures in brand health tracking as a check on longer term reach and branding quality across all major activities to set up the parameters for potential change in brand memories, not to assess the specific executions.  I talk about this a bit in Better Brand Health, in the chapter on measuring marketing activity.

J.R

11 Oct 2024

Is there academic evidence to show that ads (or stimulus generally) that includes sound, video, etc are more impactful than still images or without sound? We are a default sound off environment but occasionally viewers turn the sound on for our ads. Should we assume these viewers are ‘processing more intensely’ the ad than those who don’t listen to them?

The industry seems to assume that the more senses the better, often this is used to make a case for video since it’s more active than still image ads. Same rationale for video without vs with sound…?

I would say the Institute’s point of view on adding sound to video is that it usually helps, but is not a substitute for visual branding.

An academic journal article found that category-prompted brand recall was higher (55% vs 45%) for ads that used both audio and visual branding, as opposed to just visual branding (Simmonds et al., 2020). Furthermore, this additional recall was explained by attention to the audio messaging, measured by heart rate, as clearly eye-tracking can’t measure audio attention. Heart rate indicates cognitive processing that is more intense than automatic monitoring of the environment, which is the low level of attention given to most advertising. Adding audio can generate attention even when people are paying high visual attention, which is why aircraft cockpits have audio alarms in addition to visual displays (Wickens, 2008, cited in Simmonds et al.).

More recently, an Institute report (Bond et al., 2021), based on 14 MediaScience studies, found that even just one audio mention in an ad improved unaided brand recall (18% vs 16%) and brand recognition (69% vs 52%). More audio mentions, particularly when they occurred early (during the first third of the ad), further improved brand recall. But when audio and visual branding were compared in the same logistic regression model, visual branding was still important, and was the only variable related to brand recognition.

Finally, another Institute academic paper highlights that while motion captures attention, what really captures attention is rarity (Wooley et al., 2022). On a made-for-advertising site with 10 or more video ads on the same page, video is less likely to attract attention. Lack-of-rarity would likely reduce the effectiveness of audio branding as well. In the LinkedIn environment, however, with only one ad on the screen at a time, video and audio should have positive effects on attention and memory.

References

Bond, A., Hartnett, N., Driesener, C., Trinh, G., Sharp, A., & Kennedy, R. (2021). Harnessing the power of audio branding. Ehrenberg-Bass Institute Report 103, June. (https://sponsors.marketingscience.info/reports/harnessing-the-power-of-audio-branding/)

Simmonds, L., Bogomolova, S., Kennedy, R., Nenycz-Thiel, M., & Bellman, S. (2020). A dual‐process model of how incorporating audio‐visual sensory cues in video advertising promotes active attention. Psychology & Marketing, 37(8), 1057-1067. (https://sponsors.marketingscience.info/wp-content/uploads/sites/2/_pda/ac_pub/2020/03/Sensory-cues-and-active-attention-Psychol-Mkt.pdf)

Wooley, B., Bellman, S., Hartnett, N., Rask, A., & Varan, D. (2022). Influence of dynamic content on visual attention during video advertisements. European Journal of Marketing, 56(13), 137-166. (https://sponsors.marketingscience.info/wp-content/uploads/sites/2/_pda/ac_pub/2022/05/2022-05-Influence-of-dynamic-content-on-visual-attention-during-video-advertisements.pdf)

S.B

11 Oct 2024

For our Food & Beverages brands – is there any guidance on recommended splits of budgets between Working media and other working (Point of Sale/Shopper Marketing). We are also interested in understanding how the ratio splits based on the maturity and size of market share of the brand?

Essentially the question is, how we should invest in advertising relative to other marketing activity that reaches the audience at the point of purchase. This question has been raised by others mostly in response to Binet and Field’s calls for marketers to allocate more of this split to what we call “advertising” (they speak of it as “long term brand building”). They proposed a ratio of 40/60 in favour of advertising. There have been a range of issues and challenges raised in regards to their study (which I won’t go into in detail here), but suffice to say there is a lack of evidence to support the use of such a specific ratio. Our view is that brands need to be both thought of (i.e. Mentally Available) and easy to buy (i.e. Physically Available) when a purchase occasion arises in order to be successful. Given that you need both, we recommend aiming to build both simultaneously, as a lack of either will be detrimental to the brand’s performance. A brand’s individual circumstances may mean that they need to place more investment in Mental Availability or Physical Availability initiatives at different times (and perhaps brand/market maturity or brand size might play a role in that), but there is no consistent, evidence-based rule of thumb defining where to allocate the investment based on such factors. The other thing to consider is that point of sale/shopper marketing/retail marketing, etc largely involves speaking to potential buyers who have already made the decision to buy. You indeed need to be there to “catch these buyers when they fall”, but it is the advertising that potential buyers have experienced (often over many decades) that they bring to that purchase occasion. Advertising works by creating and building memories, which in turn nudge buying propensities across the whole potential market. Point of sale marketing, by its very nature is limited to reaching just those who are in the process of buying (i.e. its audience is potentially skewed to those who were buying your brand already). The following comments and papers from the website might also help: 

• When do the (advertising encouraged) sales occur? 

• Should marketing planning be an optimum split of spend between short term sales driving and long term brand building? 

• What is your view on Short Term vs Long Term investment? 

Finally, outside of sponsorship, we undertake media budget-setting projects at the Institute, which may be of interest if you want to dig deeper on how much to invest in advertising. These projects are focussed on just the advertising spend, rather than apportioning investment to either advertising or point of sale activity. However, in undertaking one of these projects, there are some evidence-based patterns that we employ that do involve brand and market context.

E.R
2 Sept. 2024

Is there academic evidence to show that ads (or stimulus generally) that includes sound, video, etc are more impactful than still images or without sound? We are a default sound-off environment but occasionally viewers turn the sound on for our ads. Should we assume these viewers are ‘processing more intensely’ the ad than those who don’t listen to them?

The industry seems to assume that the more senses the better, often this is used to make a case for video since it’s more active than still image ads. Is it the same rationale for video without vs with sound?

The Institute’s point of view on adding sound to video is that it usually helps, but is not a substitute for visual branding.

An academic journal article found that category-prompted brand recall was higher (55% vs 45%) for ads that used both audio and visual branding, as opposed to just visual branding (Simmonds et al., 2020). Furthermore, this additional recall was explained by attention to the audio messaging, measured by heart rate, as clearly eye-tracking can’t measure audio attention. Heart rate indicates cognitive processing that is more intense than automatic monitoring of the environment, which is the low level of attention given to most advertising. Adding audio can generate attention even when people are paying high visual attention, which is why aircraft cockpits have audio alarms in addition to visual displays (Wickens, 2008, cited in Simmonds et al.).

More recently, an Institute report (Bond et al., 2021), based on 14 MediaScience studies, found that even just one audio mention in an ad improved unaided brand recall (18% vs 16%) and brand recognition (69% vs 52%). More audio mentions, particularly when they occurred early (during the first third of the ad), further improved brand recall. But when audio and visual branding were compared in the same logistic regression model, visual branding was still important, and was the only variable related to brand recognition.

Finally, another Institute academic paper highlights that while motion captures attention, what really captures attention is rarity (Wooley et al., 2022). On a made-for-advertising site with 10 or more video ads on the same page, video is less likely to attract attention. Lack-of-rarity would likely reduce the effectiveness of audio branding as well. In the LinkedIn environment, however, with only one ad on the screen at a time, video and audio should have positive effects on attention and memory.

S.B
22 Aug 2024

I am scoping some work related to measuring Category Associations in mass (TV, radio, print) as well as non-mass (social, direct mail, email) channels. We are often asked if specific creatives are likely to drive this perception.  

We have been piloting several approaches (e.g., lift, aided/unaided message recall, etc.) in mass and are looking for how to roll this out to other non-mass testing. Would love to get your perspective on whether this is likely to be fruitful and/or if there are any considerations we should keep in mind as we set out to measure Category Associations in creatives.

Also, given the scale of social media options, formats, platforms, impressions, and other complexities involved in social testing, do you have any thoughts on methodologies/tools to measure social performance at the creative level? Any best practices to consider?”

Thanks for the question, this is something that has been on our mind and we are currently looking at ways to assess this in advertising.

Our belief at this stage is that:

1. A key metric should be about message clarity which is the degree to which everyone who sees the ads gets the same message and therefore builds the same brand associations.
2. An unprompted message capture approach is important so we get what are the associations likely to be built and not just put words into people’s mouths. The danger with prompted approaches is that they will also revert to the general impression of the brand in cases when the message is not clear, and so obscure our ability to detect this.

But we propose doing this at pre-test stage rather than in market as messaging is something you want to get right beforehand. So just showing to a smaller sample of category buyers and asking what they think the ad is saying to them.  We are currently working on testing different wording to see how much that matters.

We are also currently testing the types of execution of the message that lead to greater message clarity (is spoken better than written, are images better than text, is one message better than multiple, how much do people infer messages). No findings on this yet as results are very preliminary.

We also have some work coming out soon showing that the clearer the message the more likely you are to drive the link to specific associations.

In regards to social media testing, do you have any general guidance? We produce thousands of pieces across the various lines of businesses, some produced by our ad agency and some internally. We also have organic media. The whole thing is very overwhelming. I think the best place to start are those that mirror our mass-media led creatives with the most impressions, but curious if/how you all have been thinking about social creative testing/measurement.

I would first ask the question:

Why are we producing so many different pieces with so many different messages? Is that an efficient use of resources?

Next I would try to get some guidelines in place, by testing as series of different creative styles to work out what works (or you can wait for our R&D findings), so there was less chance of an ad being a messaging dud.

Overtime there will be some that you don’t need to test because they use the ingredients that work, it will just be when you deviate from those you need to check you have not gone too far.

You don’t need a bit sample for message testing, 100-200 is likely enough. So each piece is not that onerous, but yes it could easily mount up.

J.R

21 August 2024

I have just read the updated summary of findings on Attention and I am keen to have more details on it. Is there a more detailed report that I can have access to? I’d like to better understand the methodology and see actual numbers behind this conclusion. It’s going against the “trend” which is not a bad thing but requires even more data point if that’s the position we should have on the topic.

Thank you for reaching out about this Attention paper.  This is indeed a complex topic, and we absolutely appreciate that a lot of what we say here runs counter to current claims, particularly of proponents of Attention measurement in industry.

But in the meantime, a couple of considerations for you:  

 Firstly, as an independent university-based research Institute, our job is to critique such trends in industry, and call out claims that are not justified by the data where we find them.  Sampling theory is of course crucial in determining the validity of individual studies, but we place a lot more emphasis on building generalised evidence across a range of conditions (studies, researchers, markets, metrics, etc).  As you note, the more data points, the better!  When we review a topic (such as attention in this instance) we aim to consider as many studies, data sets, journal articles, industry claims, points of view, etc on the topic as possible in order to draw our conclusions.  The scientific philosophy that we bring to the research means we are less enamoured by a single study (even with a massive sample size, or great R2).  For many reasons, such studies can be anomalies, and we respect our sponsors too much to have them jumping at shadows, so we want to make sure that we are very confident in what we are sending out. 

With this in mind, we have drawn on a range of materials to arrive at this white paper on Attention.  For instance, as you noted we refer to the ARF study. This too is a meta analysis of 32 different studies.  One of the important conclusions, the authors draw is the great elephant in the room amongst proponents of attention measurement – that is, few (actually none that I know of) can prove a causal link between paying attention to ads and a sales response.  The ARF has had a working committee looking into this for some time, (I was actually a member a couple of years ago).  This piece draws on many (32) data sets that were made available to the ARF across many providers, using different metrics, etc.  

We also draw on a broader body of knowledge from blind reviewed journal articles and published studies going back over 70 years for our white paper.  Key to this literature is the timeless fact that low attention processing is (and remains) normal, and is the context in which ads are experienced.  As we acknowledge, some attention is necessary for an effect, but beyond this there are not levels of attention (i.e. active being better than passive, etc).  This levelled interpretation is necessary for media to justify that the currency should be based on attention.  Certainly you can code for more or less attention in most attention data, but there is little evidence that the level of attention matters for advertising effectiveness. 

We also have multiple data sets that we have examined to draw the conclusions in our white paper.  These studies use a range of different operationalisations of the Attention construct.  These studies vary from a large data set (~70,000 ads) from a global proprietor of eye-gaze attention data, to, multiple smaller studies, that operationalise attention by way of biometric and neuro (and others in between).  From this range of studies, we have been able to identify that the way you measure “Attention” makes a big difference (also what the ARF concluded).  For instance, eye gaze (as the most commonly available product on the market), is not a less sensitive measure of the same underlying construct, but rather it measures something different from the processing that neuro measures capture.  It is often claimed by those selling gaze, that it is good enough if you can’t afford neuro, and that it is a more realistic measure than lab settings which are the typical setting for neuro measurement.  However, the evidence suggests that gaze is not measuring processing, at least not like neuro can.  This difference in what is being measured is reflected in the contradictory results that have been observed across attention studies that use different measures. 

We also note that there is a push by proponents of attention measurement for attention to be incorporated in the media currency.  Aside from the clear vested interest these parties have in pushing for its inclusion in the currency, we can see clearly in the data that attention is a concept that is brought to the advertising environment by the audience (that is, individuals pay more attention if they are older, are less educated, are from certain cultural backgrounds, at given times of the day).  It is the audience’s circumstance that determines the attention they will pay to an ad, not the media in which that ad is placed.  The media doesn’t control the attention, making it difficult for them to justify charging based on it.

E.R

19 August 2024

Our Marketing team had a question on digital ads and we were wondering if Ehrenberg Bass has a point of view or any research on the below. I saw on the website info on 15 second ads vs 30 but didn’t see anything on digital. Understanding the effectiveness of 6s vs 10s digital ads. Currently I think 6s are a higher % of our buy.

You are correct in that much of the research to date has been on TV advertising and on ads lengths of more than 10 seconds.  That said, there is little indication that implications from these studies wouldn’t also apply in the shorter digital context. The two possible variables that might mean that previous findings do not hold in this situation are: 1) whether the digital platform itself makes an ad more/less effective and 2) whether the length of the ad makes a difference to its effectiveness.

Firstly, on the medium.

Despite there being little evidence that paying more attention to ads results in a sales impact, attention is in the current zeitgeist.  If we were to assume attention matters, there has been a suggestion in industry that audiences pay more attention to some platforms/media than to others. In theory, digital audiences might be more engaged/attentive and view for longer resulting in more ad impact than would have occurred had the ad been aired on a different platform. Unfortunately, the evidence that this is actually the case is also sorely lacking.

Attention is something the audience brings to the media, not something the media provides for the advertiser. Reported differences between platforms or types of media in the attention they are able to deliver are likely a reflection of slight differences in the make-up of the audience, or more likely a result of how the researcher determines attention. For example, cinema will have most audience members looking at the screen, but that doesn’t mean they are processing the ad any more than a TV audience. Similarly, for digital ads in comparison to TV, even if the ad is on screen for longer, this doesn’t mean the platform is responsible for any additional effectiveness.

But are digital ads even on screen for longer than ads on other platforms?  It is widely accepted that the vast majority of planned digital placements don’t make the viewability standard (of 2 seconds at 50% pixels on screen). As an advertiser, you don’t pay for these planned placements, but that the standard is only 2 seconds (not 6 seconds, or 10 or even the full ad length), is meaningful. It reflects that even in an apparently engaging medium like digital, very few ads actually appear on screen much past 2 seconds.

This recent paper, while also largely based on Attention to TV advertising, does speak to attention differences being minimal based on ad length.

What is known about attention (to advertising) Revised

What about differences in the length of the execution?

In many ways, digital video ads display very similar characteristics to TV ads. In fact, some advertisers use the very same creative executions for both media, and we have increasingly been seeing shorter TV ad lengths as well. While the evidence we have is (as you note) about 30 sec ads compared to 15 sec ads (paper included below), we saw there that the shorter ads are not necessarily at a disadvantage. In fact, the abbreviated length seemed to ensure the advertisers make sure they include the elements that matter when they have less time to work with.

In Praise of the 15-second Advertisement

The question is whether it is still possible to include the elements that matter when you get to 6 or 10 second ads. We can’t see why this would be especially difficult.  We know that effective executions need to cut-through, be clearly linked to your brand (through the use of Distinctive Assets), build and maintain memories, and this is entirely possible in short executions. This article goes into more details about tactical tools that have a role (or not) in producing good quality creative.

Creative That Sells: How Advertising Execution Affect Sales

So then why a 6 sec over a 10 sec execution? The main advantage that shorter ads can deliver for advertisers over longer ones (especially when we are talking about 6 or 10 seconds) is the ability to place executions for continuity and maximising reach. Assuming shorter spots cost less, then having more of them allows you to spread them out over time and audiences. And this we know does lead to more effective advertising campaigns.

1 August 2024

E.R