I’ve been part of some of the critical innovations that power programmatic today, from the development of the dynamic ad server to the introduction of cross-site behavioral targeting, to the establishment of audience data-centric cooperative models.
I believe in automation. I believe in auctions. I believe in targeting and addressability. I believe in real-time decisioning. However, as much as I believe in those things, I believe more in the need for the advertising industry to fix many of the problems with programmatic advertising.
Programmatic has provided cover for a massive amount of fraud. It has created faux precision and faux control. It has caused consumers to get lots of ads that they don’t like — ads that irritate people. And, it has helped destroy a lot of great content-driven premium publishers who weren’t able to compete on similar faux business models.
advertisement
advertisement
I know programmatic folks see the world of TV and premium OTT as the next market to be conquered, hoping to do to TV what they did to digital display. Why not? To programmatic zealots, TV advertising and premium OTT audiences are just more nails to be pounded down by their hammers.
But it won’t work that way. Digital display supply outnumbers demand and can be created instantly by adding more ad units to more pages, and moving audiences to additional pages and more units. TV and premium OTT aren’t likely to be manipulated that way.
In TV and premium OTT, advertising units that command an audience’s attention are finite. Audiences are measured by ratings and sold on the basis of time. There are only so many people. They only have so much time, and there are only so many units that can deliver portions of that time.
In fact, the massive growth of ad-free, subscription-based streaming video products means the availability of access to that finite time and audience attention is shrinking, not growing.
In that world, why would owners of TV and premium OTT media let unknown media buyers wait until the last 10 milliseconds to try to buy perishable inventory cheap, selling and reselling those units like commodity pork bellies with no regard for the viewer’s ad experience?
Actually, they won’t. They will continue to sell most of it in upfront auctions for premium prices to known buyers and brands, and will sell their scatter availabilities for even higher prices.
Yes, we will see the entire system automated with software. Yes, we will see the entire measurement system upgraded from small panels of 10s of thousands of people to massive hybrid systems of panels and direct measurements of 10s of millions of people. Yes, we will see some ad decisioning in TV made seconds before delivery, not just weeks. Finally,we will see a lot of targeting and addressability, and digital-like performance-based attribution.
Maybe you will want to call all of that “programmatic,” but I don’t. It carries too much baggage from enabling and hiding many digital bad practices.
What do you think?
Dave, instead of "programmatic," should RTB be substituted when explaining the issue you describe? Or is it enough to use "automation" as you do, to distinguish between processes that are automated to achieve greater efficiency as opposed to the programmatic snake oil you dislike?
Henry, actually, I believe that the real-time bidding feature is one of the key enablers for the bad that programmatic delivers. If a media company truly has precious and scarce audiene attention to sell, there is ery little if any incentive to wait until anything close to "real-time" to execute a media transaction. And, I find no reason to let the advertiser be the final one to determine what ad the person should see. Shouldn't that be the decision of the publisher/newtwork? Fundamentally, racing against real-time creates a lot of messiness.
Right, Dave. And, let's not forget the likely cost of programmatic TV time buying were it done the digital media way. Like it or not, advertisers would pay up to 10% for the use of a programmatic buying platform and a similar amount for "data" processing used for targeting purposes, plus who knows what for getting their commercials disrtibuted--5%, 10%? and since the sellers would have to use SSPs-selling side platforms--- to deal with the buying computers, that cost---say 5-10% ---would also be footed by the advertiser via higher CPMs. Net, net, advertisers who now pay 1% for broadcast network TV time buys and 2.5-4% for cable and spot TV buys would be faced with execution costs 20-30 times higher, maybe more. Is that an attractive scenario? Can the programmatic folks guarantee improved targeting efficiencies of 30%, year after year to justify their costs? And what if the sellers notice that certain programs command more interest than others---as they surely will? Wont they raise their CPMs on those shows accordingly---which will nullify the assumed targeting efficiencies edge that are promised? The primary difference between TV and digital is this. In TV land the ad sellers will not be helpless pawns of the computerized buying systems and ad distributors as is so often the case with digital "publishers". Rather, they, the TV ad sellers, will chart their own course and few but some desperate players will subordinate their interests to those of cherrypicking computer buying systems.
Dave, this is really well stated from someone who knows. I hope you are right, but fear that the fraudsters and others will find ways to make advertisers think they are getting the quality inventory at bargain basement prices wiping out the costs Ed correctly foresees. But all these middle men will find the fake or bad inventory and advertisers will buy it because that is what they know how to do.
Ed, spot on. The taxes of programmatic alone make it super unattractive for TV.
Thanks Jack. As much of an optimist as I am, I'm with you that the fraudsters will try to find ways to penetrate any automated systems that enter TV. However, I'm aslo hopeful that strong, hybridized panel based measurement can go a long in helping us sniff out the attempts.
Good piece Dave - the other point is TV is regulated, in presentation and minutage.
Glad I caught this post Dave. You are of course 100% correct. The key difference between online advertising and any other media that has become more "digital" is that online advertising transacts on ad impressions which created an instananeous and permanent state of over supply for publishers to manage -- so "programmatic" was like a vacuum sucking up all of these unsold impressions -- TV, Billboards, Radio, Print even if presented in a digitized format sells based on audience reached so limited supply removes the need for programmatic vacuums.
"TV Audiences are measured by ratings and sold on the basis of time"
Dave - Do you think the TV model could replace CPM/CPC/CPA to help publishers get out of this RTB "race to the bottom" on their inventory?
Pretty bang on Dave. But....we are talking about labels. Yes programmatic as a label has baggage, but that's the dull, rusty edge of a double edged sword. Lots of good things can also be said about the automation and granular targetting and messaging that programmatic has delivered....not to mention performance.
Will TV be automated, yes .
Will it offer granular data driven behavioral and contextual targetting, yes.
So in that respect it's programmatic.
Both Seafood.
Let's call it Lobster, and the other stuff Shrimp.
I like both, Lobster more.
Mark, as you say, it's important to go beyond the labels. You are right---TV time buying and selling will be "automated" -----but does that mean that the sellers will actually allow the computers to make their sales decisions for them? Not likely. More likely is that automated systems---"programmatic systems"---will, increasingly handle tha paper work and data processing and, perhaps, the scheduling of ads. And they may also be used for emergency, sudden cancellation sales and, maybe all of the sales for very small, "long tail", channels--if these still exist. But I doubt that the big money buys will be turned over to the machines anytime soon which means that people will make the final decisions regarding most of the dollars placed on national TV---unlike the current "programmatic" practice on digital media. And I also doubt that advertisers will be willing to eat the kinds of costs that the programmatic platforms and their attendant allies, the data chompers, need to make a big killing. It's simply not a reasonable cost for return trade-off.
Totally agree Ed.
Just saying, It will be different, but the same in many positive ways. Some inventory will be sold as you describe, other less desirable inventory in ways not dissimilar from display.
If, or I should say when, frequency capping comes to TV, and wasted over impressions eliminated, the inventory will not be as finite and more remnant will appear as waste is eliminated......(unless of course humans intervene to maintain the illusion of scarcity, killing the real promise of addressable tv)
For me, the promise of addressable is as much about reducing the waste of over exposure, as it is about reducing the waste of mass targetting.
In any case since Lobster feeds on waste, I think we should call it Lobster and not programmatic.
Mark, I totally agree. I am hopeful that addressable systems in TV will be used most to manage over frequency and also reduce wasted delivery to those well outside the target. The ad experience will be better for all.
Dave - Not to sound obnoxious, but if you talk to anyone who has been selling linear TV, they could have told you 15 years ago that "programmatic" linear TV buying will never work. It would need to provide the kind of control that sellers have always had and as far as I know, there is no automated platform that attempts to achieve this except for Mass Exchange.
Auctions will not work. Negotiations will still need to exist, whether on the phone or in a program manual format. Been there.
Long, I totally agree. However, you are one of the dozen or so people that were there over those years. There are at least 10,000 others in digital advertising and pure linear TV that don't know that what's in the trades and press releases can't work.
Guys, you are right about the lack of knowledge about the pitfalls of "programmatic" TV time buying if done the digital way among many TV people and, of course, the dgital folk who have never worked in "legacy media". There is hope, however, as the larger media agencies have looked carefully into the matter as well as some major---- and smart ----advertisers and they, too, see why it doesn't work. I am hearing this from a growing number of contacts. Hopefully word will spread to the sellers and eventually to the trades. This is a far cry from the prevailing sentiment a few years ago.
Great points Dave, but will agencies, who have gotten a taste of essentially controlling the Ad Ops and optimization (and look the other way on some of the fraud) and who apply blunt inaccurate and lunatics crude tools on the basis of "we don't pay for it if it doesn't pass our muster" basis, be willing to cede that control?
All I know is that there have been hundreds of attempts to automate linear TV transactions and unless TV sellers are involved, or people who are in the weeds of TV pricing and inventory management, they will continue to fail.
This is the problem....Pure digital people thinking they can quickly transition an indutry to a digital model. eg demand side people acting like they understand the supply side of TV. IMHO, there are only two ways linear automation of some kind can work:
1) Build a platform that can handle the complexities of pricing and negotiating thousands of selling titles in close to real time, so TV networks can confidently break away from selling strict daypart mixes and start selling audiences
2) Get advertisers to start grading agencies on cost per reach point, not CPM - this will open the door to a reach optimization approach that will benefit both sides - an advertiser's CPM may go up, but their cost per reach point will decline. Linear TV sales teams are underselling their reach potential by selling daypart mixes that keep reaching the same people and create too much frequency
Well said, Dave. Agree on all points. It's a question of when, not if.
However, practically speaking, for automation, increased panel sizes, improved targeting and addressability to occur, at scale, a lot has to first change within the industry. And if past is prologue, scaled solutions that work for a majority of clients, sellers, agencies and researchers are a VERY long way off.
Even though the video industry collectively understands the challanges and opportunities, it apparently lacks the kind of collaborative leadership required to achieve agreement on new business rules, principles and standards that allow everyone to prosper in a fraud-free enviornment.
Where that leaves many is in the familiar routine of "incrementalism". For some, this is a happy place, with enough opportunity to squeeze out improved YOY results. However, in light of profound changes in video consumption patterns, I seriously doubt that the broader commercial ad marketplace can continue to afford the luxury and pace of incremental change.
Thanks Barry. Great point. Without question, the biggest barrier to automation in the TV ad world is incrementalism.
Long, as you and Dave know, the media planners decide the daypart mix for many TV time buys---in conjunction with their clients---- based on exactly what you are talking about. I think that it's asking too much of the buyers, within a daypart and network type---broadcast or cable---to try to calculate incremental reach for each possible buy. Certainly not an upfront buy where a large number of the shows don't exist yet while many holdovers are in new timeslots. What can be done is an education process --probably by the planners to the buyers---about the pitfalls of buying so many spots in a fixed set of Monday to Friday shows---say on a cable channel----whereby the same message appears on the same show day after day. Or worse, the same commercial appears in the same one hour episode---as so often happens----which means that the second redundent announcement has little impact. But will the buyers listen? And will clients back the planners if they try? Or, more to the point, do many clients brand people and CMOs. even care?
Sadly I have to agree with Barry ---who fought the same battle as Long is talking about at Turner about twenty years ago. Despite what they say for public consumption, most CMOs---who should be the real drivers of new ways to make for ad relevant TV time buys---and plans for that matter----simply can't be bothered. Until that changes---if it ever will it's up to the media sellers to effect changes that are acceptable to buyers and their clients and that's a long way off. Why? For one thing the buyers don't talk to the real client---the brand marketing people--- nor do the ad sellers. Instead, they talk mainly to each other and under this setup there is no incentive for the sellers or the buyers to rock their own boats. So it's audience tonnage and as low a CPM as possible---even though better ideas---like NBCU's short, high impact, breaks of a few years ago---make a great deal of sense.
Ed, you've certainly identified the Achilles Heel in all of the opportunities to make TV advertising better. Too few CMO's are incented to push their TV media buyers to seek media effectivness over lower media unit costs. Nothing holds us back more.