Commentary
Performance Is New Programmatic In TV Advertising
- by Dave Morgan , Featured Contributor, April 20, 2017
Nope, something even more fundamental is happening in media, and it’s going to have its biggest impact on TV.
The future of TV will be about performance. As media legend Alan Cohen proclaimed as he took over as president-CEO of independent agency Quigley-Simpson earlier this year, “Performance media is where it’s at. It’s the new programmatic.”
While Cohen was speaking of media broadly, the rise of performance-based buying and selling will have its biggest impact on TV advertising, I think.
Why? Performance dominates digital advertising, because it can be measured and optimized that way. And performance is what marketers truly want, fundamentally, when they buy.
advertisement
advertisement
This was never really possible on TV historically, not at the spot-person level, but now it is. The rise of massive amounts of direct, second-by-second TV viewing data at the person and household level that can be matched to online and offline purchase data in a privacy-safe way means that marketers no longer have to wonder how TV ads work in driving sales or leads. Now they can know. Better yet, now they can use those insights to better plan and optimize their campaigns going forward, maximizing their performance.
No longer are they hamstrung by only having mix models that take months to conduct, and not getting much deeper than the network or day-part level.
Is this really new? A bit yes and a bit no. Television has always had performance advertising, you might say, since direct-response ads have been with us almost as long as TV has. That’s different, though. You might also argue that TV has always been a performance medium for marketers who sell regularly and track sales, whether it’s retailers like Walmart or packaged-goods brands like Coca-Cola.
Without question, retailers have always known in a general way how their TV spend correlated to store traffic and sales. Movie studios have always seen the impact of TV spend in their box-office numbers. So too have airlines, restaurants and hotels when it comes to butts in seats and heads in beds.
However, TV advertising has never been as predictable, provable and performance-oriented as new digital channels like search and social have become for marketers. On platforms like Facebook, marketers can buy campaigns that are guaranteed to reach a certain group of people as well as deliver their desired business outcomes. Marketers get to have their cake and eat it too.
That is where TV is going fast. The notion of “programmatic” TV was a bit of a head fake. The vast, vast majority of TV is not about to be bought and sold on programmatic exchanges — not for many, many years.
However, the idea that TV will be bought and sold based on performance — that world in unfolding before us right now.
What do you think?
Dave, imagine a world where the majority of "linear TV" buys were based on almost instant advertiser reactions to sales results and day by day juggling of ad buys to acount for dynamic events unfolding in the marketplace, based on knowing exactly what every viewer watched on a second by second basis.It sounds great---in theory---but we have no measurement of what people are "watching" with this degree of precision---especially when it comes to commercials, but also for program content. All we have is device usage---- a very misleading indicator of viewing and, in particular, of attentiveness.
But even if we had all of that and an advertiser attempted to react---in a very timely manner---to data about what people were buying as a result of "seeing" a brand's commercials, how would the ad sellers handle this degree of flexibility? There would be no upfront. The sellers would never know if they were sold out on a given telecast until the last minute---an impossible situation----and the advertiser would have to hire scores of people to assist each brand manager in evaluating---and reacting---to the seeming marketplace variations----or develop a computerized system for making such decisions.
My point is that "performance" -based TV time buying sounds great but in actual application, it can never get nearly as "granular" as some---and I don't include you-- may think. It makes much more sense---where the concept actually applies---for an advertiser to make periodic adjustments--say quarterly----as as awareness and/or sales patterns--- within certain demos, product user groups or mindset segments are spotted. This may involve adding GRP weight via more responsive scatter buys or reallocating an existing upfront buy among the various brands when this is needed, rather than locking their spots all into place on a yearly budget basis.
So, I'm with you in principle---for those advertisers who require this kind of performance mechanism and may benefit by it. But at the same thime we have to approach this in a practical manner, bearing in mind how the various organizations involved---brand management, client media watchdogs, agency media planners and time buyers plus the "creatives" interact and make decisions. Also, the sellers' needs must be given serious consideration as without their cooperation, nothing will get done.
As I've indicated, I don't believe that every branding or image-driven ad campaign needs to be preformance based in the manner you mean---but it's a valid concept for many others. What needs to be developed is a step by step progression that eventually gets such advertisers and their agencies---as well as the sellers ---closer to the ideal than is now the case. This will take time and may never be fully implemented, but it's a worthwhile goal.
I think you're right Dan and we've been advocating for new ad units with direct response via TRE.
With 93% of the viewing audience having their phone nearby non-peformance ads mean leakage to competitors via search and/or never finding what you wanted anyways.
We see these ad units being paid for as 1) a base payment for showing the ad on TV and 2) a performance variable $/KPI such as $/sale or $/email address, etc and this variable component creates an open ended revenue opportunity for the ad carrier.
This payment model helps make sure that the add has strong call-to-action and both the ad owner and the ad carrier are aligned in making sure the ad is placed in front of the right audience versus spray and pray.
Sorry Dave......meant to type Dave and typed Dan instead. My apologies.
Performance based buying is what all media buys/negotiations should be based on regardless of media channel, but you have to look at the correct metrics that correspond to those media buys. In eComm, yes, you can't look at same day metrics to see if TV buys have influence any lifts in transactions, branded search or even direct navigation. You have to judge it based on the average customer transaction lag time. Depending on lag time, performance could be judged on weekly, monthly or even quarterly, but the customer lag time should set that cycle.
If you think about your own purchase decision, when was the last time you saw a product and purchased same day? That purchase decision is sometimes pushed out to weeks, months or maybe even years depending on the price point of the product. Last interaction might be promotional messaging. In which case, that evaluation would be shorter, but that customer is already in-market. They're just waiting on the right time to convert. In order to judge TV appropriately, you have to track where that TV campaign falls within the purchase funnel, find that efficiency and scale those opportunities rather than looking at same day metrics.
We've been purchasing programmatic TV for almost 3 years now when the first exchanges went live, and have built out models to guage performance for branded campaigns that tie to either purchase intent or actual purchases. Since programmtic TV landscape is fairly new to most brands, we've started to educate the industry on the subject starting with the good: http://www.mediatwo.net/programmatic-tv-the-good/.
Trey, I agree with your comments about the difficulty in attributing sales to a particular TV ad "impression"---it's not as easy as the direct-response types think it is when trying to quantify the effects of branding campaigns. I disagree with your contention that all media buying negitiations ---as opposed to media plans---should be performance-based. This places an undue responsibility on the ad seller who has and desires no control over the advertiser's product in terms of quality or desirability and certainly is not responsible for what the advertiser says in the ads. Trying to improve media targeting is one thing but making the ad seller a "partner" in the ensuing ad awareness or sales results is a non-starter. That ball belongs squarely in the advertiser's court.
OH. Where to begin. Is it reasonable to expect the same type of performance results and use the same criteria when promoting a $55,000 family car and a Ginzu Knife? With digital response rates in the low single digit conversion rate versus impressions delivered how will the performance pricing be negotiated and how will impressions be valued if perceived without value? If I am constantly optimziing and reacting on the fly have I also built predictive modeling in a a world evolving faster than I can react? What about tomorrow's prospect? and the influencers? How do I ensure that I have right creative for all this dynamic flying? I am no Luddite, but I do wish we would learn from the mistakes from digital and be realistic about what we can do with media measure and proof of performance in an increasingly fragmenting inchoate media world.
You're right, performance for an auto and a set of knives is different but they all have an ROI.
For the auto I would hope that email addresses have a $ value for the dealer and this could be the KPI. (One can easily see an auto dealer pay for a list of names and emails of people who have indicated they want more info.....now.)
The knives can be purchased instantly via DR so sales are the KPI.
Dave,
I think you are on the right track. However, performance based marketing and advertising must not be limited to the ability to measure and track results, as it’s done in direct response TV. The real shift that must occur is for the sellers of advertising to provide advertisers a guaranteed sales increase, upfront. It's easy to accomplish. For example, you mentioned Coca-Cola, well let’s say Coke wants to sell 1 billion bottles of cola during a two day weekend. The media company providing the ad time can guarantee the advertiser, Coca-Cola, will sell 1 billion bottles of cola during the two days the ads run. The sales must produce net new revenue. Meaning the sales must be above and beyond sales the Coca-Cola company would normally make during that weekend. This guaranteed sales increase deal is made upfront before the advertiser pays for any ad time. And the seller/provider of the ads only gets paid for their ad time when the guaranteed sales increase is met. Now, that's what every advertiser wants! As the pioneer of the guaranteed sales increase algorithm that makes it possible, and the first to advocate for such a change to take place, I am glad to see at least your media company begin to follow along this line of thinking. Let’s hope that many more join in the weeks and months to come.
Mark, no TV ad seller in his/her right mind---except for some desperate bottom dwellers trying anything to get some business---is going to guarantee sales "upfront" as you predicated. Which raises a question for Dave. Dave do you guarantee your clients sales increases along the lines Mark proposes?
All good in theory, but two thoughts.
Burger ad airs. Then it happens to rain - a deluge. Traffic is gridlocked, QSR sales are down and all of a sudden the TV ad didn't work? C'mon ... really?
Second, a broadcaster should be focussing on content (programme and advertising) and trying to bring the best product to the screen rather than working out whether they should be showing luxury car ads or a Sham Wow ad. That is the marketers role ... to know, test, and understand the results of what they are doing.
Purchasing behaviour is multivariate. There are probably more than 100 factors that contribute to a sale. Of them a half-dozen to ten would be critical, It is an absolute vanity to think that programmatic is going to shift the performance dial to the extent that marketers expect.
Ed, if the TV ad seller is a broadcast network, and they had the "knowledge and skill" to provide advertisers guaranteed sales increases, upfront, they would be crazy NOT to do it. Why? Because there is not a global brand advertiser "in their right mind" that would turn down guaranteed sales increases, upfront. And you can take that to the bank!
And Mark, they would be crazy TO DO IT, because as my recent post explains, there are WAY too many factors outside of their control to guarantee sales.
Mark, the national TV ad sellers are in the drivers seat---not the advertisers---because the sellers---especially the broadcast networks---know that the advertisers are wedded to them and have no serious alternative---yes, I mean digital. Guaranteeing an advertiser's sales when you have no control over the quality or appeal of the product nor how it is marketed, is a suicidal approach---unless the guarantees are so limited in scale that they are meaningless.
John, your post does not explain anything about the "knowledge and skill" to provide and deliver advertisers guaranteed sales increases, upfront. You only speak about what you think can't be done. Obvisiously, if the network does not "know how" to do it, then yes, it would not be wise for that network to try. That would be like jumping in the deep end of a pool and not knowing how to swim. But for the one network that does "know how", that network is indeed crazy, crazy like a fox!
Ed, I agree. I don't think that all TV ads will be super granular ad performance based. Branding and sponsorship and roadblocks will be around for a long time. but, I do think a big portio of TV ad inventory will be performance sold and managed.
Totally agree Trey. Even if it's not bought and sold on performance, the performance characteristics of all TV inventory should be measured and monitored that way.
Very good point Dale. The mobile device creates signficantly enhances the feedback loop for some purchase categories.
Dorothy, you raise some really good points. Being successful in this future won't be about just the data, it will be about the management and exploitation of a massive number of complex models across all categories, consumers, competitive and environmental characteristics. it will not be an easy world for most companies to play in.
Right on target Mark. That is exactly what I think will happen. We're starting to do some of that at my company in the tune-in category, guaranteeing TV network marketers incremental audience lift (ratings lift) for shows based upon commited spend. This model will become a key part of many of the most competed for "moveable purchase" categories where TV ads are a big driver of sales - travel, CPG, retail, ecommerce.
John, there is no question that many, many factors impact sales performance of TV ads. However, even with that uncertainly, I think that TV ad inventory is fundamentally and signficantly underpriced relative to performance such that you can be safe with guarantees as long as you have strong enough predictive analytics and enough budget and time to meet the thresholds. The fact that some TV sellers will do it - and find sucess - will cause many in the market to follow suit, whether they like it or not.
Great point Ed. Yes. The national TV ad market is a sellers market with demand outstripping supply. However, once some of the media owners find that they can dramaticallly increase yields of their lower demand inventory - non-prime, over 55 demos, etc. - others will have to follow suit or won't be able to keep up with them on earnings. I expect this to play out in a significant way in 2-3 years.
Ed, offering guaranteed sales increases does not mean giving up control over the qaulity and standards of the ads broadcast on the network. And while it may have been true in the past, that advertisers had no alternative to national broadcast networks, or digital online sites, I have it on good authority that is changing in 2017, with the Guaranteed Sales Upfronts. Global corporate brands will for the first time have serious and exciting alternatives to the oligarchy of broadcast and digital networks, and for the first time be able to choose guranteed sales increases, upfront for every product they advertise. Ultimately the only opinion that will matter is that of the bottom line. And like the old saying goes, "If it makes dollars, it makes sense!"
Mark, my point is that a TV campaign is univariate. The broadcaster has very deep "Knowledge and skill" can (just about) control everything relating to that buy - well their part of the buy.
The thing is that out in the real world or marketing and selling of product, the marketing and sales levers are multivariate. The factors that impact on a sale also go WAY beyond those limited marketing and sales levers.
When you deal with a broadcaster (or indeed any media owner) they focus on delivering the task at hand. Rather gratingly they tend to call is a 'campaign'. Wrong. For starters, broadcast TV may be just one of the media contributing to the campaign. Within the TV component of the campaign, it might be a three broadcaster buy. No single broadcaster can 'guarantee' TV sales delivery, simply because they don't know (until post-event) the campaign - they just know their slice of the TV pie.
In a heavy weight sophisticated campaign it might be a three broadcaster national TV buy, a 10-channel cable buy, a 30-channel local buy, out-of-home on three formats, AM and FM radio, interet and social media, along with press, EDM, catalogues & couponing, letter-box drops and flyers, all supported by deep retailer dsicounts and in-store promotion.
These are the factors that the marketer can in some degree control.
Then there are the things they can't control. Competitor's advertising. Competitor's pricing and in-store marketing support. Weather. Macro-economic factors. Events (both good and bad).
Each one of these inputs 'nudge' a brand to sales success. There is no one input that can claim to be responsible for hitting the sales target ... or conversely, all inputs could claim the be successful.
Further, no one media owner is anywhere near having comprehensive enough data covering competitive media, competitive brands etc. to be in a position to guarantee sales outcomes.
(TBC)
And I haven't even touched on the impact that running a successful campaign strategy (e.g. repeating last year's blockbuster) but receiving poor advertising creative would have (again, outside the media owners control).
It is naive of any media owner to think that they, and they alone, can be in such control that they can guarantee sales, and it is vain to claim solus credit when there were other factors involved.
A media owner SHOULD guarantee what they can control. Basically TV has been doing this for eons - guaranteeing delivery of audience eyeballs that have an opportunity to the see the ad. So have magazines. And newpapers and radio. And lately O-O-H. Each are supported by independent third-party audience measurement systems. Digital delivery data is prolific but far from independent apart from comScore, Nielsen etc.
The client, and to a lesser extent the media agency, is the only person who can see the whole picture and measure its sales effects.
Let's not even discuss that many of such analyses confuse correlation with causation! Yes, your medium or media property may have been actively utilised during the sales period, and yes sales may have increased - but that does NOT mean it was your media property responsible for the success.
I'll shut up now - just a few decades of working on this stuff spewing out.
Mark, I don't know who you are talking to about the TV upfronts, but please try someone else. There are no upfront advertiser sales guarantees being made by the ad sellers nor will this happen in 2017 or later. What you are confusing with sales guarantees is the practice of using product user data melded with set usage ratings to profile Nielsen measured audiences by various sellers for selected accounts. the sellers do this because they know that this methodology creates the illusion that their shows do a better job of targeting product buyers than is really the case----but the actul selling is still done the same old way---with bundled program packages and discounted CPMs that allow the sellers to unload all of their GRP inventory---no matter how bad it is. In short, nothing has changed---it's just a lot of hoopla. There is no rational way for a media ad seller to realistically allow an advertiser to set sales goals for the ads run on a particular media platform, then demand that thye be attained ---or no pay. As John correctly points out, audience duplication and many other fasctors which are completely out of the seller's control as well as being unreadable as to their precise causative effects regarding every "impression", make sales guarantees of the type you are advocating a no no.
Very good points Ed. Performance guarantees fit much better in scatter than in upfronts. While TV today doesn't have anything even close to the dynamic and addressable world of media placement, audience databses, data targeting and real-time creative optimizatoin of the digital ad world, there have been some developments that make it much better suited for sales performance guaarentees than it ever could have before. HH matchable set-top box at scale means that we can now build massive single source data structures underlying TV and link TV ad exposure to sales at the HH level very quickly (across millions and millions of household). Audience fragmentaion has become a friend of targeting here (as Erwin Ephron always told us it would), since the eplosive growth of small episodes (75%+of TV ad inventory now occurs o an episode rrated under 0.5%. Near reali-time ad occurence data means that we know where competitive ads ran, as well as ads under the control of other networks. Set-top bob matchable credit card purchase data means that we can know how competitors ads are performancing, as well as our clients. None of this brings certainlty to sales performance, but applying all of this across thousands of small spots in a massive campaign can bring a lot of performance optimiztion to campaigns - certainly 2-5X - as long as we don't let the pursuit of perfection interfere with just doing better.
Ed, I believe you misread what I wrote. I never said I was talking to anyone about the “TV upfronts". What I did say was...... “I have it on good authority that is changing in 2017, with the Guaranteed Sales Upfronts." Ed, the "Guaranteed Sales Upfronts" is a different meeting and "alternative" process to what you refer to as the “TV upfronts". You most likely would not have heard anything about the "Guaranteed Sales Upfronts" because it has not yet been officially announced. And I am not at liberty to go in to detail here. Now once it is announced, you may or may not want to attend or even be invited, but that does not change the fact it is happening in 2017. Second, you are the one that is confusing the old Nielsen measured audiences, ratings, CPM, GRP, with what I have refereed to as "guaranteed sales increases". Again, it's two different paradigms. One old. One new. Everything you have stated about the old/current paradigm being used by Nielson is obsolete, and irrelevant to the guaranteed sales increases algorithm, that I have pioneered. The guaranteed sales increase algorithm, is a polymathic algorithm ( advanced mathematics that I invented), and has a solid calculus base. It’s been tested for 10 years, certified and it works. If you and John wish we can discuss the math here on this forum. But with all due respect, you are not able to legitimately criticize the algorithm because you have not yet been privileged to learn it.
Interesting, Mark. Can you tell me which of the broadcast TV networks is participating in your guaranteed sales upfront and are they selling "linear TV" time---primetime---on this basis?
Mark, I'm all ears.
I've worked with, used, created and assessed many such systems and algorithms over the past couple of decades. I'd love to know more about your 'secret herbs and spices' - well as much as you can share about your approach.
Math and algorithms cannot solve for collinearity and spurious relationships. More importantly, marketers will have to rise the the challenge of defining performance. Oh how we long for the arrival of clearly defined goals, operational definitions, data clean enough to produce clear signals. Bring it on. Obfuscation be gone.
John, you can start with the foundational math here. https://youtu.be/mAQd-LLLMio
Very interesting, Mark, and I am eagerly awaiting John's reactions to your YouTube presentation. It certainly makes it pretty clear about your thinking process. I hope that my fellow poster, LZ, takes note, as well.
Thank you for the link Mark.
I simply don't know where to begin.
But if I were you I'd start with brushing up on the difference between an algorithm and an equation. And by the way, it is not calculus you are 'using' but it is algebra.
The mathematics you use is roughly Grade 6. It's odd that what you have 'invented I was learning in school when I was a little tacker. While it is all mathematically correct, I haven't made up my mind whether it is naively simple or simply naive.
There wouldn't be a marketer our there that doesn't know that revenue = number of sales * price per unit. Similarly, they would all know that profit = revenue - costs.
In your 'hypothetical example' you appear to have forgotten that the retail price is not the price the manufacturer/marketer gets. Yes, 500 1 litre bottles of cola raises $500 revenue ... for the store. The cola manufacturer might get half of that.
When you talked about profit, the only cost you mentioned was advertising. So you deduced that for the $500 of (retail) sales less the $100 advertising costs was $400 profit. Yes, you did mention that was $500 'new sales' - but in no way explained how the advertising was able to double the existing base line (retail) sales. You didn't include any costs for manufacturing your cola. Or its distribution to the stores. And apparently you have no staff costs. No rent. No overheads. No media production costs.
I find that your claim that GSI is "...for calculating the exact number of sales business will make from advertising a product or service. For the first time it is now possible to guarantee an increase in sales for every product advertised in every media." to not be plausible.
Should you be able to provide empirical evidence for any product that has used GSI, that due to just advertising and using any medium has always increase sales, I would be more than willing to review the data and should it past muster change my opinion.
I also have one small question to ask.
I notice that in your YouTube statistics, that since you posted the video on June 12, 2016 it has had 120 views up until April 16, 2017. In that 309 days there appears to be at least 1 view every day, and for most of September to January it had one and only one view each day over that roughly five month period - which just looks odd. Is it possible that the Y-axis on your graph is not quite right? Anyway, look forward to a little bunp in your views when Youtube cathes up and there is no lag.
You are welcome John. The youtube video is not there to get views, it’s simply a convenient presentation tool. While you have acknowledged that the math presented in the video is correct, you apparently missed the section in which I explained, that not all the math is shown because it is “proprietary.” It is not our objective to publish research or get a Nobel prize. Our sole purpose is to increase the profits of global corporate brands. After all, we are in business, not academia. So unless you are Chairman of the Board of Directors, or own a majority of shares in a publicly traded company that is using our guaranteed sales increase algorithm you will not be privy to more details. It addition, no one other than myself and the officers of our company fully understand how it works. As far as your request to review data from companies using our guaranteed sales increase algorithm, we never discuss or publish a company’s revenue, profits, or sales results. Such information is kept confidential by all parties. You are correct, that in our example, the cost for manufacturing the cola is not included, nor is any other cost of doing business. That’s because the profit is on the advertising investment, not the operations of the company. In our hypothetical example all the company is interested in, is receiving a billion dollars in net new revenue. In real life, I have found the same to be true. It has been my experience so far, that global corporate brands do not care what media professionals think about the plausibility of receiving guaranteed sales increases, upfront. They only care that they receive it. And that’s the bottom line!
As I suspected Mark.
Yes, I said your maths is right, but only because it is simplistic at best. That is in no way an endorsement of the rigour of your algorithm. It's like saying 1+ 1 =2 thefore the rest of the 'secret algorithm' must be correct.
I'm sorry but I come from a background of 40 years in research, statistics and mathematics. I am also a member of the Australian Market and Research Society which has its code of ethics. One area of the code is that empirical evidence and peer review is essential.
However, I understand your point about confidentiality.
The global brands I have worked with have ALWAYS demanded the highest levels of proof and rigour that any analytics system be 'bullet-proof'.
So we will just have to agree to disagree, which given the naive simplicity of the basics of your 'system' and the lack of credible proof but instead a handful of fanciful hypotheticals of a guaranteed billion dollar sales increase purely from advertising is something that I am very comfortable with.
John, I wonder how many advertisers have data that tells them exactly how many units of product they will sell for every dollar spent in every media vehicle---a TV station, a radio station, a broadcast TV network, a magazine, a particular website, etc. ---that can be traced exclusively to the ad "exposures" garnered on that vehicle? I'm really looking forward to hearing more about this "guaranteed sales upfront"---especially which media sellers are involved and how their deals are being fashioned---and monitored. I hope that Media Post will cover this event so the laggards of the advertising and media world will profit by this amazing development.
Easy one to answer Ed.
None.
Jack, on target as always. Managing collinearity and spurious relationships will put a ton of stress on the performance models, even given the enormous power of TV. On top of that, as you point out, marketers need to rise to the challenge of establishing what performance is, something that has been beyond most in this context.
John, you must be confused, it is not, nor was it ever my intention to seek your endorsement, or approval. I only joined this discussion to comment on Dave’s Performance Is New Programmatic In TV Advertising, article. As far as the guaranteed sales increase algorithm and process is concerned, the only approval that matters is from the members of the board of directors of global corporate brands that have a fiduciary duty to their shareholders for maximizing profits and shareholder value. Should you ever rise to that level, we can certainly talk further.
Ed, every advertiser that has received a guaranteed sales increase analysis, has that info and capability. This includes my own companies which use the guaranteed sales increase analysis and algorithm to sell advertising. We know exactly how much ad inventory we will sell this year. (Yes, we practice what we pioneer. ) Speaking of the “Guaranteed Sales Upfronts” that meeting and process is very much a private affair, by invitation only. There are no press invites. And, unlike the hoopla, and glitzy glamour of the TV Upfronts in NYC, the Guaranteed Sales Upfronts are about business, not show business.
Jack, the primary reason for the obfuscation is media/marketing people are all speaking their own languages rather than speaking the one language that matters. And that is the language of business. The board of directors of the global corporate brands that advertise have a fiduciary duty to achieve shareholder wealth maximization. Share holders want to see an increase, in sales, profits, and ultimately stock price. The results of advertising should contribute directly to the figure on the bottom of the free cash flow statement, in the form of net new revenue. Because at the end of the day, the only research statistic the bank will accept is the one with a $ in front of it.
Mark, recognizing the top secret nature of your guaranteed sales upfront and the obvious fact that neither John, Jack or myself will ever rise to the level of corporate responsibility that is required to participate in such a ground breaking undertaking, can you at least tell us approximately how many ad dollars you expect to see placed in this manner for the 2017-18 upfront?Surely you have such figures, just as your advertisers know exactly how many sales they will attain for their ad spending.
Spot on Jack and Dave.
I've found that if you get enough data (e.g. three years of weekly sales data) and you use multivariate time-series micro-econometric modelling, you can probably tease out collinearity effects for slightly better than one-in-two brands. And with that you end up with around a half-a-dozen key drivers of sales that can describe/explain sales movements with around 70% correlation.
Unsurprisingly price (generally expressed as price relativity to the category mean) is usually #1. Distribution is crucial - but if all major players have good distribution it loses explanatory power. Weather (be it temperature thresholds, absence/presence of rain etc) would be a driver for around half of the 100 or so brands/categories I've worked on. Sadly, advertising, or more correctly advertsing share, would factor in sales models probably 2/3 of the time (a very tough - but truthful - result to report to the client when you work in a media agency), but the certainly factor in brand equity modelling with its more long-term nature.
You get some really specific oddities that occur. We modelled McDonald's Happy Meal sales (which won an Effectiveness Award in the early 2000s - woo hoo) and found the 'theme' of the premium was crucial, along with radio presence during afternoon drive time (especially on rainy days but less so in school holiday periods). That completely changed the media planning. We also found 'theme' crucial in advertising 'scartch-and-win' lottery tickets which drive sales way more than the volume of advertising.
But the KEY thing is that ALL of this is done posteriori, coupled with the knowledge that what worked in the past has between a one-in-two to two-in-three chance of moving the incremental sales needle again. NEVER has there been 100% explanation and 100% guarantee.
Mark, Ed asked what my opinion was.
I provided it.
You used the phrase "you have acknowledged that the math presented in the video is correct".
I was merely ensuring that it was 100% clear that acknowleding that the basic maths you use (and remember it is algebra and not calculus - fix that up in your sales deck pronto!) is in no way endorsing your product or service.
Yes Ed. It is sad that we won't scale those dizzy heights.
plkfidjdosp - oops ... sorry about that. My 2014 Hall of Fame award fell on my keyboard.
But just to let you and Jack know that I've found this great deal buying cheap Florida real-estate that apparently has 24/7 water (I think that was a typo - I'm sure they meant waterviews) that you might be interested in. I was also thinking of going off grid with it and using the lightning-in-a-bottle I bought.
John while I don't do the complex math part, I understand what's being done and , I , too, have seen and been involved in numerous attempts to correlate data---past sales, ad spending, GRPs, etc. ----with particular media that were part of a media mix---like daytime versus primetime TV, broadcast vs. cable, TV vs. print, and, in some cases, attempts to cut it even finer---by specific media vehicle---or type of vehicle. Invariably these efforts fail because there is no way to factor in the effects of audience duplication and prior exposures, let alone to account for differences in ad exposure, ad impact, promotional spending, word-of-mouth, distribution problems, the activities of rival brands,etc.So, even if an advertiser at the outset of a new positioning campaign or a new product launch, thought it was possible to predict sales results based on total ad spending or reach x frequency, this could never be applied to a given ad seller, even if that seller was crazy enough to guarfantee sales results.
Ed, I do appreciate the question but find no reason to provide that information to the public.
John, the road to solving a calculus based problem is often paved with algebra. I assumed you would know this since you claimed to have 40 years experience working math problems. And as you have acknowledged, you are not privy to the polymathic algorithm so you have no basis to criticize, correct or offer any advice whatsoever on that part of the process. Might I also suggest that you cease the obvious troll bait and juvenile mockery. It adds nothing of value to the conversation, nor does it enhance your stature.
You all assume any kind of targeting will be used for "good" purposes to change the hearts and minds of individuals to buy products/services/ideaologies. And when these technologies are used to sell and people buy ideaologies to destroy, will you be as sorry as the dude who invented the K-cup ? or worse....
Paula, I do not presume to know what you mean by "any kind of targeting " but I can assure you that unlike Einstein’s E=MC2, the guaranteed sales increase algorithm will only be used for good, and not fall into the hands of those who want to use it to destroy. In my opinion Einstein should have absolutely kept his discoveries a trade secret and licensed it's use only to those companies and organizations than wanted to use it for the betterment of mankind. On the other hand why would the inventor of the k-cup be sorry? And how does that relate to performance media exactly? Just trying to follow the logic and thought process of some of the folks that have commented here, which is very odd and confusing in regards to business and commerce. I do understand and agree with Jack and Dave, that marketers can't agree on the definition of performance. Even though Wall street analyst, investment bankers, shareholders, and the financial team and board of directors of every company in the S&P 500 know exactly what "performance" means. It's also odd when so called marketing people who claim to be award winning experts say something can't be done simply because they can't do it, or didn't invent it. It makes one wonder, is this truly representative of the mind set of marketing/media professionals today? I sincerely hope not. Anyway, thanks for adding a different perspective to the conversation.
Good for you, Mrk. Even though we disagree about the basic concept, it's good to know that your algorithim will not get into the wrong hands, which might cause great harm. Good luck on your upcoming guaranteed sales upfront. Keep us informed on the progress.