TV Has An Attribution Problem

  • by , Featured Contributor, February 22, 2018

The advertising industry is quickly approaching a future where household-level sales attribution becomes table stakes for all media campaigns and channels. This future sets up digital media companies well, but could be disastrous for TV companies.

TV has an attribution problem, and the industry needs to fix it.

Let's be clear: TV advertising does not have an effectiveness problem. No media channel delivers access to more consumers faster, with more impact, than TV. TV’s ability to drive sales predictably and at scale for categories from retail, travel and packaged goods to restaurants, cars and financial services, has been well known and documented for decades.

However, TV companies don’t provide real-time reporting at the household and user level for each and every campaign, as digital, search and social do. And that’s a big problem for TV companies that want to avoid big ad budgets cuts -- especially from agencies and brands under pressures to provide greater accountability for when and where they spend ad dollars.



When marketers buy ads from Google, they get access to Google Analytics to help them attribute sales and conversions to those ads at the user level. When large marketers buy ads from Facebook, they are given access to what is arguably the best marketing science team in the world, as well as a powerful set of software tools to understand and attribute the sales impact of their ads at the person and household level.  When marketers buy ads on Amazon … You get the picture.

Digital is winning bigger and bigger ad budgets because it works, and because digital giants can draw a visible line between ad exposure and a sale. As advertisers become increasingly addicted to campaign-level sales attribution -- and they are -- media sellers that lack this feature will lose budgets, whether or not they are effective. That is the dilemma TV faces.

Fortunately, there are now plenty of tools to link TV ad exposures to sales and conversions at the campaign and household level. TRA, now owned by TiVo, helped pioneer this capability. Neustar has built a powerful product here. Nielsen provides these capabilities through its NBI and NCS tools. ComScore has offerings here. So do another dozen or so start-ups. And sellers like CBS, NBCU, Turner, Fox and Viacom are stepping up in this area, including attribution with some of their campaigns.

However, in spite of a plethora of tools, providing campaign-level sales attribution at the household level still the exception in TV. It needs to become the norm.

What’s holding TV attribution back? First, many in the industry are afraid of it. What if the ROI is low? That's one worry.

Second, many others don’t think that TV should be driving sales, let alone providing attribution for it. TV is a branding medium, they say. Plus, clients and agencies control the creative, so it’s an unfair burden.

Third, it’s expensive. Who will pay for all of this, they ask.

Fourth, putting sales attribution and ROI out front in TV will create a race to the bottom in pricing, they claim.

Fifth, if they just keep doing the same things as last year, and maybe negotiate a bit harder in the upfront, they’ll make it through another year -- or through the sale of their company -- without having to make any big changes.

For all of those worried about what "they say," I point you to TripAdvisor’s Q4 earnings call from last week. On it, the online travel booking company’s CEO and CFO told investors to expect strong revenue and profit growth for the year ahead.

Why? As a result of significant improvement in the company’s cross-channel ad attribution capabilities, TripAdvisor was going to significantly increase its TV ad spend and decrease its digital marketing expenditures -- which would result in more sales of higher quality than it was getting from its previous marketing mix,  more heavily skewed to digital advertising, according to the CEO.

Yes, that's a digital- and mobile-first company making that statement.

The message for TV companies is simple: Trust in your platform. TV ads are underpriced, not overpriced. But you won’t be able to capture the real value of your ad inventory until you know exactly how your spots drive sales for your advertisers, and you make sales attribution table stakes for all your campaigns.

What do you think? Are you ready?
17 comments about "TV Has An Attribution Problem".
Check to receive email when comments are posted.
  1. Henry Blaufox from Dragon360, February 22, 2018 at 2:47 p.m.

    Dave, Your observations on the TV ad sales industry fear of failure and change may be unpleasant to accept, but may have to be faced sooner or later. As different industry players see potential rewards outweighing the risk of loss, resistance to the new meaurement techniques will weaken. It will probably be incremental, at least at first. You've noted a few trying it out already. As successful methods and the value are discovered, the shift will become pronounced. This is the way of technology adoption in all fields. Anyone in accounting still using a TI or HP hand held calculator with an 8 bit chip?

  2. Eric Fischer from HJA Strategic Consulting, February 22, 2018 at 2:56 p.m.

    The one constituency here that is not mentioned is the agency partner.  As long as there is inertia with that part of the equation, nothing will change.  Agencies generally receive higher commissions on digital spending so they have an incentive to not push for this, plus many of the larger holding companies own portions of the digital measurement platforms and use client data to make their trading platforms more robust, so this is another disincentive to move towards providing a true ROI for TV, which we all know will be significant.  Clients will always want to do better as spending is analyzed like never before, but they need to force their agencies into pushing for this.

  3. Ed Papazian from Media Dynamics Inc, February 22, 2018 at 3:07 p.m.

    Dave, while I agree that what you term "attribution" is a potentially good thing in cases where it applies and providing the data is precise enough----isn't the underlying issue that until a buy is made for an individual brand---based on its specific targeting needs---that we are getting nowhere.? In other words, advertisers must decide whether it is better to think of TV as a low CPM, mass tonnage buy, made on a corporate basis, or as a much more selective brand by brand targeting opportunity where attribution, when it applies, should be part of the equation.

    Another part of this problem is the understandable rerluctance of the sellers to allow cherry picking of their program lineups as well as their fear of making GRP guarantees on highly selective metrics.  Last but not least, if we were to shift from a situation where 250-300 large corporate TV buys account for 90% of national TV upfront ad dollars to one where there are 7000-8000 individual brand negotiations, the costs of making such buys and selling the time would probably skyrocket. Instead of paying around 1% for national broadcast TV buying and 3% for cable, advertisers may have to pay much more and the sellers, who expend, perhaps 7-10% of their dollars on sales, Nielsen, etc, might have to spend a lot more---which would be reflected in higher CPMs. So the trade-offs and possible ramifications need to be considered.

    I still believe that the ultimate solution will be two kinds of upfronts---one, the current type featuring corporate tonnage, low CPM buying; the other for individual brands, using more refined metrics and, finally, more opportunistic scatter buys made either way. If this comes to pass,  costs of buying and selling, while rising will still be fairly reasonable, the sellers' fear of being cherry picked and of "weird" guaranteed audience metrics will be minimized and advertisers can decide which way---or ways---they want to go without everyone being forced into a single modus operandi .If you go mass you operate more efficiently but sacrifice on targeting; if you go selective you probably pay a good deal more but, hopefully, it's worth it.

  4. Stephen Pickens from Advertising Advantage, February 22, 2018 at 4:28 p.m.

    Every TV advertiser should be measuring the immediate and medium term impacts of TV, which can be accomplished through a rigorous statistical process that we've developed over the years. Even if your goal is raising long term brand awareness, every spot has the potential to lift website and search traffic and can provide valuable information about where an advertiser should be focusing their efforts and budgets.

  5. Paula Lynn from Who Else Unlimited, February 22, 2018 at 5:20 p.m.

    If you were responsible for a company/corporation/oligopoly's necessity to sell products/services for a certain profitability would you decide to buy television vs stealing money from TV to more digital since you know what you know regardless of "stats" ?

  6. Craig Jaffe from Baruch College, Zicklin School of Business, February 22, 2018 at 5:41 p.m.

    I enjoyed your write-up, Dave. If household-level sales attribution is becoming table stakes, are you sure it will be so for "all media campaigns and channels"? That seems bullish as not many things in life are 100% of anything. I imagine this may be more relevant for some categories and less for others. If this is so, who do you believe will most likely demand attribution as table stakes and what would be your estimate of the dollar value? All my best, Craig

  7. Ed Papazian from Media Dynamics Inc, February 22, 2018 at 7:14 p.m.

    Craig, that's a point that I've been trying to make repeadedly. Many advertisers use TV for reasons other than pure sales generation or with precise targeting as a secondary factor. Often, these are found on TV sports, news and "special" content. In addition there are mass use categories like toothpaste, detergents, etc. which are not as finely segmented as more selective products or services. Still Dave has a point and I believe that too many TV advertisers who go the corporate buying route operate under the false assumption that all that really matters is "creative", namely the attention-getting and motivating power of their commercials. While this may be true 65-75% of the time, that doesn't mean that more precise audience targeting---especially where one can match the mindset of the program viewer with the nature of the appeal the ad is making---can't be fine tuned to give the ads more punch. Unfortunately, too many CMOs still regard "media" as a boring numbers game and won't take the time to look carefully at the trade-offs inherent in the ways they think of TV and buy time.

  8. Dave Morgan from Simulmedia replied, February 22, 2018 at 9:30 p.m.

    I agree Henry. Once a few of the large network groups go in big, it will accelerate accross the industry quickly.

  9. Dave Morgan from Simulmedia replied, February 22, 2018 at 9:32 p.m.

    Eric, I remain hopeful that agencies will eventually do the right thing for their clients here. However, I'm pragmatic enough to know that the change is more liekly to be driven my smaller, independent shops like yours, who make innovation their differentiator.

  10. Dave Morgan from Simulmedia replied, February 22, 2018 at 9:34 p.m.

    Ed, very well said. I agree with you that this won't happen until we see buys isolated to indiviudal brands. I also agree that we are likely to see a two-stage upfront as the most likely path for scaled adoption of audience and outcome-based buying on TV. Your prescription makes for a very good prediction.

  11. Dave Morgan from Simulmedia replied, February 22, 2018 at 9:40 p.m.

    Stephen, I agree. This kind of analysis is a must-do for marketers. However, I think that it it is critical that it goes beyond a lot of historical statistical approaches and that the largest media channels, TV and digital/search/social are measured at the impression to household/user level. Ideally, with both propensity modeling and hold/outs. It is the only way we will be able to truly separate the individual contributions at the channel level.

  12. Dave Morgan from Simulmedia replied, February 22, 2018 at 9:41 p.m.

    Thanks Craig. Certainly, we won't see 100% adoption. My hope is for 40-50%. That is enough to swing the indusry.

  13. Stephen Pickens from Advertising Advantage replied, February 23, 2018 at 3:40 a.m.

    Agreed Dave. It has been a challenge, but there are creative solutions to providing deeper insight despite the limitations imposed by linear delivery. Thats said, my hope is that OTT and Connected TV with their inherit device ID + IP address availability will offer the key to mapping TV exposure along the same lines as digital exposure in the pursuit of more robust multitouch attribution models. Once this form of TV delivery and consumption reaches critical mass, we'll be in a much better position, but I suspect ubiquitous adoption is further off than most anticipate. 

  14. Ed Papazian from Media Dynamics Inc, February 23, 2018 at 7:49 a.m.

    Thanks, Dave. I think that the two upfront concept, once it begins to be used will probably position the mass, corporate buys to go down first, as we are creatures of habit and will go where most of the action is---at first. However, it may dawn on both buyers and sellers that the most effective path is to reverse this and have the brand by brand upfront come first, followed by the mass upfront. This would allow the sellers to maximize their CPMs on those shows that have the most favorable targeting metrics---probably a minority of the shows---while the brands get the largest targeting "lifts". Later, the mass buyers can get their pagkaged aundience tonnage across the sellers' entire schedule of shows, with what is left of the better programs melded in with the more numerous but less selective entries. Eventually, I can see the selective, brand by brand upfront accounting for about 25% of national ad dollars while the mass, corporate buys generate 35-50% and scatter, which can be done either way gets the rest. And many brands  will probably go both ways, laying in low CPM buys in the mass corporate upfronts and scatter as a hedge, while allocating considerable portions of their spending to more selective buys. It will be interesting to see how this develops.

  15. Douglas Ferguson from College of Charleston, February 23, 2018 at 10:13 a.m.

    This reminds me of what Don Schultz was saying twenty years ago: Advertisers will want to pay to reach only their customers. Unspecified reach is less important than accuracy. Advertisers nowadays have a list of their current and prospective customers and they only want to pay to reach the people on that known list, no wasted impressions, but broadcasters don't have a list to match, just GRPs.

  16. Ed Papazian from Media Dynamics Inc, February 23, 2018 at 10:30 a.m.

    Douglas, there are all kinds of advertising situations and all kinds of targeting methods. In my esperience, very few advertisers have lists of who now buys their product and, especially, of who might be swayed to buy the product. The basic strength of TV, with its ability to generate massive reach as well as to focus GRP "weight" on the more desirable segments---when a brand is allowed to do so by the corporate big wigs---- is that you can cover all of the bases. There is no such thing as "waste-free" advertising. If you had a way to target only that 20% of toothpaste users who accounted for 35% of all toothpaste buys, but ignored all of the others, your brand would probably suffer a debacle of huge proportions----as rival brands who go the mass route would steal many of your "other" buyers---the ones you regard as "waste" and no longer advertise to.

  17. Dave Morgan from Simulmedia replied, February 24, 2018 at 8:15 a.m.

    Douglas, while I agree that we're moving much closer to the marketing world that Don Schulz envisioned, I don't think that many marketers today - outside of Business-to-Business marketers - have lists of target customers that they want to reach. I continue to be surprised with how few marketers really know and understand "who" they should be reaching in broad ways. So many advertisers' target audience segments are aspriational and not always very data informed. This is why I believe very much in the advertising future laid out by Byron Sharp in how brands grow - fundamentally by reaching non-customers (typically with wide-reaching media channels first) and acquiring them and learning from them and looking for more. So many times thsese new customers don't look like the aspirational ones the marketers thought that they wanted. For example, so many folks who want to target "millenials' find out that their real opportunity is in "booomers."

Next story loading loading..