- The Crosstab
- Posts
- Linear Media Buying, The Times They Are A Changing.
Linear Media Buying, The Times They Are A Changing.
An interview with Adam Wise from National Media
Mark Harris
September 23, 2024
Giphy
Today Adam Wise, Chief Innovation Officer at National Media, answers five questions I sent him about linear media buying and how things are changing, and most importantly what campaigns can do about it.
I hope to do more of these interviews with folks about particular niches of our industry please shoot me an email at mark@coldspark.com if you have any ideas.
How has tv buying changed during your time in the business
Fragmentation is the biggest change. Since I joined National Media, the average TV ratings have fallen by 60% & the dynamics of how people watch have completely changed. My first year at the company in a typical week broadcast could reach 90% of the U.S., now it’s barely at 60%. At the same time the amount of content has increased, it would take someone watching TV all day every day for a year to consume 80% of what was produced last year. That’s not including sports or news which are the main drivers of ratings. Just to put things into context below you can see the decline in FL or 4 match broadcast affiliates & how swing voters are less likely to watch TV then the standard A18+ population.
Affiliate (Florida Markets) | 2024 (A18/Swing Voter) | 2010 (A18/Swing Voter) |
ABC | 39% (31%) | 58% (58%) |
Fox | 38% (29%) | 52% (52%) |
CBS | 39% (30%) | 58% (57%) |
NBC | 35% (24%) | 62% (60%) |
Simply put 10 years ago, it was a lot easier; you have to be smarter & more vigilante than ever today. We used our in-house smart TV panel of 24 million households matched to the voter file to track cross-screen reach & frequency against first person audiences.
What can campaigns do about rising rates
Reserve early, have money escrowed for clearance of highly valuable programming & be ready to make choices informed by data. This will be a recurring theme in our conversation but points aren’t equal. Be ready to pick between high impact programming & point goals. Reserving early isn’t a perfect solution, but it helps the market set & adapt to demand. It’s also important to have buyers that push the stations… the stations play games, and you need someone who can call BS at the right times. We’re fortunate internally to have a lot of great buyers with an average of two decades of experience but also the former Chief Distribution Officer of E.W. Scripps, which managed 60 stations in 40 markets. We know the games they play & how to counter them.
The other issue… is to reserve digital early. Linear will have price rate increases but in the digital and/or advanced advertising space (i.e. addressable) a lot of the inventory is gone when it’s gone. Buy it up early & adjust.
What are common mistakes campaigns make on media buying
On linear the biggest mistake is a cost-centric mindset. Meaning going for low cost per points at the expense of key programming. Sports, in particular, are expensive, but they matter. There was a recent VideoAMP case study showing the only way to reach younger consumers was College Football. Over 80% of reach came not from social or digital… but linear TV on college sports. Not all points are created equally. SambaTV did a study a couple of years ago showing that over 90% of linear TV impressions for the top 10 corporate advertisers in America went to half the U.S. population. This meant half the country received 140 ads per week and the other half 7 per week (in the updated version just released it’s even more dire… over 90% of linear & CTV IMPRESSIONS go to about half the country). We have to build campaigns for reach so we can get 50+1%. That’s very different than the traditional mentality of how cheaply can I deliver the message? Frankly, we’ve lost business for being honest. You can cheaply reach half the country, but to get to the vast majority you have to be smarter & relentlessly creative.
Also, frankly, I think we overbuy news because it’s easy for buyers. It runs multiple times every day & has ratings. But who is a heavy news consumer that hasn’t made up their mind? Malchow has a great study in his recent book (on page 2!) that shows the bulk of people moved by political advertising are habitual voters who can’t name what party controls congress. Is that news? Yet 47% of GOP advertising last year was on … news.
On the digital side we’ve seen a few things, first & foremost a consistent under investment in digital. People have also confused digital with programmatic. The majority of quality digital functions more like TV & cannot be bought through a DSP. Make sure your digital buy includes premium providers. Linear TV has become sports & PrimeTime has become CTV. Digital buying is over-relying on programmatic. The largest CTV supply in the county is YouTube – not available programmatically, Hulu – not available programmatically, Peacock – not available programmatically & many others. Fun scenario in Montana in 2020, we were doing Senate IE in very tight market conditions & our direct placements were running at 1/3 the cost of programmatic inventory & were 3.5x more likely to reach our target than A18+. It also allowed us to be higher up on the waterfall and have a higher share of voice on key platforms, driving significant value for the client.
What should campaigns look for in a media buyer
Honesty, creativity & OCD. But seriously it’s a great & tough question. The industry has seen an influx in vendors promising the moon & slicing rates without clear KPIs for what is a good buy. There are a few things I think you should look for in a buyer:
Nielsen or ComScore License (ideally both): too many buyers depend on others (aka stations & media sellers) to provide ratings info. When buyers use station ratings that means an average 27% increase in ratings from what we internally estimate, which lowers cost per points, but you’re not getting what you think. It’s amazing how often success is defined by a low cost per points (cpps) & people jack up cheap programming, inflate ratings or use alternative measurement solutions to look cheaper. All the buyer has done is hide reality.
Reconciliation: what is your process for reconciliation (and YES IT’S POSSIBLE ON DIGITAL). Recon, is how buyers look at orders & invoices to see what cleared. A good buyer should do this but also should have a very low return rate.
Transparency: there’s been a rush to lower commissions in buying but the result has been hidden fees, media might cost $35 cpm now they’re charging $60 cpm & offering you a 1% commission. Congrats you bought down a point on your mortgage only to pay more than you would have otherwise.
For context here are some normal rates +/- 10% & ignoring some extreme demand situations
OTT/CTV - Premium - $58 in most cases but a few vendors can be a little higher depending on additional targeting added and the geo. Examples would be Max, Peacock, Hulu, etc.
OTT/CTV - Programmatic - $42 in most cases but in some cases when targeting niche or smaller audiences it could be a little higher, but it almost never should be. An example of this would be “quality” programmatic inventory. Much cheaper than this and you’re getting bad stuff. More expensive and you need to ask some questions.
OLV - Programmatic - $22 in most cases but in some cases when targeting niche or smaller audiences it could be a little higher, but it almost never should be. Same rule as above… cheaper, check what you’re getting (always love when someone brags about a CPM then I see outstream video in the logs).
Data: how do you connect the clients goals to media...we’ve never known more about voters than we do today, but on the flip side all our clients’ investments in voter data and targeting can be siloed from the media spend if not executed properly. That’s why we’ve hired folks from the RNC and Data Trust who have experience building voter universes and working with campaign teams and pollsters to make sure everything is working together.
What is one cutting edge thing you're doing that you are most proud of
Great timing! We just launched last week a new data solution that solves for voter file match rate problems. We took historical polling data from one of the largest & most credible pollsters…then we tied it to what people are reading online to build custom audiences.
It’s pretty powerful for a few reasons…but first the problem with voter file matching… I go find “Mark Harris” on the voter file, I upload it on Liveramp, Nuestar, etc. there’s about a 90% chance they’ll recognize his address but only a 50-70% chance it’ll end up on a DSP… aka 63% chance he’s findable at the end of the day. But here’s the REAL problem, signal loss, let’s assume we found Mark, we can only find him within 40% of impressions. What do we do for the 37% missed on the match & the 60% of other impressions?
Frankly, we think we found the answer & rolled it out to the public. The average person reads 540 stories online each month… we’ve turned that into a data set for modeling. The system processes every 30 days 200 billion website visits on more than 100 million websites taking the domain, title of the story & using NLP/AI the top 128 themes of every story. It uses this to build out 70 thousand data points on 380 million devices each month. Simply put… a lot of data & a lot of very sensitive & specific data compared to the one thousand + files on the consumer file.
But we took it a step further… we actually took polling data (real polls… live caller, text to web, etc., not online panels which as you pointed out in a previous post aren’t there yet) and matched it to this data set to build swing audiences. If a voter answers a poll that they’re undecided, we then track each day what they read online to build a semantic model of who is persuadable.
It solves a few problems: 1. 3rd Party Voter segments miss HALF of swing voters & even worse are 64% waste, 2. Scale – we’re scoring every impression, every ID meaning more reach & more impact 3. Cost – it’s available in major DSPs & we can quickly spin up more segments as needed.
We also built a pretty cool ACR panel but that tracks cross-screen (broadcast, cable & CTV) reach & frequency against 1P audiences, but that’s been out since 2019.
Bonus: What's one piece of advice would you give to people just getting started in this business
Be bold, the more you tell the more you sell… it has taken me too long in my career to realize competence alone is only partially rewarded. Just remember push & pull, you’ll attract as many clients as you repel.. that’s totally ok.