Commentary
Digital Ad Marketplaces Don't Work For TV
- by Dave Morgan , Featured Contributor, October 25, 2018
And some folks have done more than just talk. In 2003, a start-up called SpotRunner launched an online marketplace for local TV ads. It was shut down before the decade was out.
In 2007, eBay tried to create a TV ad marketplace, but it never even made it to the launch pad.
Also in 2007, Google created an online marketplace for TV ads, only to shut it down five years later. Microsoft bought a TV ad-tech company, Navic — or its Admira TV ad marketplace — but it didn’t make it much past 2010.
What gives? Why have some of the world’s largest tech companies failed at bringing marketplace technology to TV?
To me, the answer is pretty straightforward. Just as everything is a nail to a hammer, to most digital ad-tech companies, all media is like digital.
advertisement
advertisement
Except, as anyone who truly knows TV knows, the world of TV advertising is nothing like digital. Here’s what I mean:
In TV advertising, demand exceeds supply. TV has a futures market, the upfront, because more advertisers want guaranteed placements on TV than there are placements to be sold. Thus, they must reserve them and lock in pricing many months in advance.
Digital doesn’t work that way. The vast majority of digital ads are not actually bought until a millisecond before they are served, or the availability perishes.
Real-time systems operate totally differently than futures systems do. Virtually all of the core ad technology that powers digital marketplaces — bidding, ad serving, optimization, demand-side platforms and sell-side platforms — operate on a real-time “waterfall” system, and have at their core real-time decisioning engines that optimize against an addressable media unit.
The digital ad-tech systems that can make the most optimal decisions against the largest pool of addressable units (browsers, apps, mobile phones, etc.) in the least amount of time win.
Not so in TV. The value — and optimization potential — in TV advertising is in the future, and the transaction unit is a TV spot. The vast, vast majority of TV spots are not addressable.
Thus, the purchased unit has to be viewed as a portfolio, not a person or browser or app, since every TV ad spot represents thousands or millions of different people that have different duplication and relative cost relationships to every other one of the spots available on TV during a campaign (every day, there are more than 250,000 spots on national TV).
Trying to predict and optimize TV ad campaigns with digital ad-tech systems is like trying to predict tsunamis off the coast of Indonesia six months in advance by sticking your hand out your Manhattan office window 10,000 times per second. Good luck with that.
TV networks have nothing to gain by exposing “unsold” inventory. Why do TV networks have special deals in place with direct-response advertisers, many of whom pay much lower rates than large brands do?
Because the structure of most of those deals permits the networks to preempt those ads at the last second with higher-paying ads, and having them in place guarantees scarcity of available impressions. That dynamic is critical in maintaining TV as a supply-constrained market.
In that world, TV networks gain nothing from exposing the volume or pricing of their unsold, or under-sold, inventory to buyers. In the digital ad world, publishers must push their “avails” information and status directly into the buyers’ interfaces, to be picked off at the last moment.
In the TV world, it is the opposite. Buyers’ demand — the amount that advertisers or agencies are willing to bid and pay for specific TV spots or audiences — needs to be pushed all the way into sellers’ interfaces, with sellers able to unilaterally accept or reject those bids each and every time.
This mismatch between the interests of TV sellers, and the design and operation of digital ad-tech systems built for a marketplace where buyers rule, is one of the most fundamental reasons we’ve seen so many digitally inspired marketplaces fail in the world of TV advertising.
Will all this change? Will we see marketplaces purpose-built for the TV ad ecosystem? For sure. It’s an area I’ve been personally focused on, and we’ve heard from a lot of the large TV ad market participants like AT&T’s Xandr and Comcast that they’re headed there, too. This will be fun to watch.
What do you think?
Good points, Dave. I would add that the underlying assumption in all of the digital tech operations that tried to convert TV to their way of thinking was that every advertiser was in the direct response field. Not so. Indeed, quite the opposite. The vast majority were in the image shaping and/or branding business where metrics that cant be quantified play a key role---like promotional tie-ins and merchandising. Also, many TV advertisers are mass marketers, not niche brands catering to very selective consumer segments. Finally, there was the cost of buying digital--- the so-called "tech tax"--which is approximately ten times the corresponding cost for buying and selling TV now. Who was going to eat those rip-off prices in TV---the sellers? Fat chance.
Great point Ed. For sure, the vast majority of digital leads with a DR mindset. While I'm a believer that TV can do a much better job for performance-focused marketers than it does today, the foundation of TV media should always been the brand advertisers with a longer-term sales and growth focus.
Dave,
Nice piece.
While the road to automation of linear TV buying is going to be gradual, we could be in for some surprises within the next couple of years on the OTT front. vMVPD skinny bundles like Google TV, Sling, Playstation Vue are beginning to amount to more than a hill of beans, currently at 7.1 million homes, expected to more than double by 2020. Right now, the commercials airing on national broadcast and cable networks should be the same in linear as in OTT skinny bundles. It will be interesting to see, however, how this might change as streaming, with addressable capabilities, becomes a more dominant means of viewing. Perhaps networks would offer some kind of addressability during national commercial air time which might suit an advertiser like Unilever who can lock in TV inventory but then dynaically allocate spots against individual brands within its portfolio to individual homes that best match target audience profiles. Under this scenario, you would need some digital-like ad serving technology, which streaming is well suited for. I'm not an engineer but it seems it would be a matter of time before the tv networks would explore this avenue to improve on their inventory yields, which have been languishing.
Great points Gerard. I do think that we're going to see the introuction of targeting into linear TV that will "hybridize" some some buys, delivering bundles of targeted spots (many of which might use different creatives for different brands, or for different households and different brands) blended with linear deliveries that are not addressable. This will require complex inventory systems, since you will have to sychronize addressable deliveries and "bundled" linear deliveries. Net. Net. The technology exists to do it. The big challenge will be the packaging and sales approaches to bring it to market, and make it easy to buy and measure.
What's also needed are equally sophisticated systems that allow the seller to optimize total ad revenue by creating---and resubmitting--- packages of discounted GRPs incorporating off-target fare plus the goodies at such a low cost that the buyer still gets what is wanted but also relieves the seller of less desirable inventory. I fear that we are forgetting the fact that most---not all, but the majority---of targeting specs will focus on younger and upscale consumers when the majority of TV's audience weight is exactly in the opposite direction---older, lowbrow. An ideal system would allow the buyer to create weights for all major segments of the consumer base---some high indexing, others about average, still others, below par---- so the sellers could try to maximize their total ad dollar yield. Many consumers have value---some more than others----and it's a mistake to focus only on a single set of high indexers.
Excellent point Ed. The way that TV sellers today manage "packaging" and "rotators" to fill out the GRP needs of sellers, and dollar-average the pricing of the prime spot investments down, is one of the more clever pricing approaches in media. Critical in a more automated, and marketplace-like, future for TV advertising will be to find a way to replacate this in a digital world.
Excellent information. Great post, bolstered by insightful follow-up comments. Thanks to all for sharing their POV and expertise.
Dave - Don't forget AdAuction.com which became OneMediaPlace! (circa 1998 -2000)
Also, IMHO, a TV network's control over pricing and inventory management is the key to a lot of this. Digital people, as you say, think it is all about the buy side (a la display). The biggest issue in TV has always been on the sales side. Those who can create a process that provides close to the same level of control that linear TV deals offer sales teams now, will end up with the best solution.
The trick is gaining access to the TV inventory and that will require the TV networks to be extremely comfortable with your approach. I believe the solution will lie in making TV even more of a reach vehicle than it is now and a process that syncs up well with traditional TV inventory management processes.
Programmatic TV 3.0 is coming! (Right now we are stuck at 2.0)
Thanks Long. Yes. I remeber OneMediaPlace. Totally agree that the key is an operating model that networks win with, and are comfortable with.