5 ms·

A few honest questions for anyone who's an ad buyer for mixed media: I've wondered how advertisers look at broadcast advertising compared to digital. For one t

by redm 9y ago

A few honest questions for anyone who's an ad buyer for mixed media: I've wondered how advertisers look at broadcast advertising compared to digital. For one there's no direct correlation (as in a click) with broadcast television. There's also no direct accounting of who actually saw it, just a best guess based on Neilson numbers. It seems to me that DVR's are akin to ad blockers as well. Why is digital so conversation obsessed (clicks/purchases) and not focused as much on the branding values (CPM)? Is broadcast advertising so much cheaper than digital that it makes sense to just blanket ads and hope for the best?

jackgolding 9y ago
I've worked with buy side teams and a lot of it is ignorance. I think you'd be horrored with some of the slots that your company gets purchased (and that process, in Australia anyway.) Things like having 1/3 of your 5 figure weekly budget spent on one advert.
scottshapiro 9y ago
Broadcast TV also has 70 years of inertia.
teej 9y ago
The way people handle it is to triangulate on the value based on a couple different sources. For example, you may test your TV ad in a handful of markets first and see if you experience lift in those areas. In addition, you may offer a unique coupon/promo code in that ad and measure uptake. Then you survey customers and ask them how they heard about you. So maybe in your test regions you saw 2% promo uptake, 3% lift, and 5% survey responses saying they saw the ad. You'll pick an estimate somewhere from 2-5% depending on how conservative or aggressive you want to be. That's how you approach it at a small to medium sized company that actually cares about how their marketing budget is spent. I've also seen an ad agency attempt to measure incremental website traffic gain in the test regions during the time slots the ad aired. If your traffic is large enough to have a clean baseline or the ad response is massive this is probably okay, but I'm skeptical of this methodology. If you're a bigger company, there are also ways to get better viewthru conversion metrics. Shazam offers something to this effect, where they allow you to Shazam a commercial. Shazam can then tell you of people who viewed the commercial who went on to convert. Nielsen may sell this too. With TV if you can't get the numbers to work on a pure customer acquisition basis, you'll measure brand recognition and recall in your test regions. If your ad makes an impact there you can then justify spending some of your brand budget in addition to the marketing budget to justify the overall expense.
- malyk 9y ago
  In the US you can buy national ad spots on major channels pretty cheaply (couple hundred $ per spot) if you buy "remnant" inventory. Basically that means that the network hasn't sold all of their "run my ad in the 9:22am commercial break" spots and so you can try to by a spot in the "morning" day part and it might run in the 9:22am commercial spot...or it might not. But they will tell you if/when the ad ran at some point so you can go back and check your stats. Attribution is kind of a bitch, but the methodology above tends to work very well given the right traffic size/mix.
throwaway284728 9y ago
I work at a major advertiser who's media budgets measure in the billion dollar zone. > I've wondered how advertisers look at broadcast advertising compared to digital. At our levels of spend, we build econometric models to understand media spend effectiveness and spend a fair bit on tracking and vetting the methodologies of trackers. > For one there's no direct correlation (as in a click) with broadcast television. Correct. Frankly, we're interested in valid human viewable reach, not clicks. Not even Impressions though most our buys are CPM today and we want to buy viewable CPMs. We use in fraud detection providers and generally don't pay for Impressions detected as fraud. > Is broadcast advertising so much cheaper than digital that it makes sense to just blanket ads and hope for the best? In many cases yes... In some cases no - especially when there are demographics who simply don't watch much TV which is increasingly a reality. At our scale, digital - even the biggies struggle to give us reach affordably.
- hyperpape 9y ago
  How many companies have ad budgets in the million dollar zone? And roughly what percent of revenue is that?
- shostack 9y ago
  Fellow senior buy side guy. Used to work at a top search agency and am now client side. This is the correct answer. At that scale focus groups, surveys, econometric modeling, etc. are what determine success. The harder part is for mid size advertisers who don't have the budget to move the needle to make those things viable, but still need to be branding heavy. Beyond that, digital is way more trackable and often more efficient as a result, even on the branding side.
philipodonnell 9y ago
My understanding is that even broadcast television advertising can be pretty well targeted based on the show's demographic profile, the viewership of the station/channel, known DVR usage trends, and the viewer's location. Surveys on brand awareness and recall can be delivered locally to confirm the ad was seen. The larger the company (large scale means more likely to do broadcast TV advertising), the more they can control other advertising channels in order to test whether the advertising is showing lift locally. There is a whole industry and 50+ years of experience using these things to gauge advertising effectiveness, so the limitations and usefulness is pretty well understood. Then the internet comes along and digital advertising is promoted as being better at all types of advertising (which of course its not). And its new (new is risky/scary) and no one knows if exposure to digital advertising has the same "brand awareness" lift as non-digital, or how to quantify that. The results have been mixed. P&G recently announced they would dramatically scale back social advertising because they determined it wasn't effective. So its on digital to prove its value, and the argument becomes "direct correlation/clicks" is proof, which is fine. Some people accept that and some don't. In 50 years, the limitations and usefulness of digital will be pretty well understood, but in the meantime, advertisers need more than "its new" to trust it.
rhizome 9y ago
Why is digital so conversation obsessed (clicks/purchases) and not focused as much on the branding values (CPM)? I'm not in the profession, but my theory is that what can be measured, will be measured. However, and someone please correct me if I'm wrong, that doesn't mean that any digital strategy is more effective than any print/broadcast strategy from 40 years ago. Keep in mind these are the people pushing for autoplay video.
bcoates 9y ago
It's a mix of two things: The first is searching under the streetlight: both attributable digital ads and non-attributable ads are put into a media plan with (mostly fanciful) estimates and projections as to returns and value before the project starts. Then the campaigns are run, and the attributable ads are scored according to actual results, while everything else is either carried forward on the assumptions the estimates are true, assigned an arbitrary fraction of all sales/traffic, or run through some GIGO model. A lot of personal prejudice and tribal knowledge is applied here. The other, smarter instinct is that while traditional/non-digital advertising can't be finely attributed it can be monitored and audited, and this is vastly more important. Mistakes and underperformance are bounded--fraud is not. A billboard in an inappropriate place with shitty creative at least exists and has some non-zero value not that far from the best possible billboard; but fraudulent Internet advertising can just dump all your money down a hole, never be seen by a relevant human, and do nothing but give you false metrics. DVRs aren't that much of a threat -- most of what a TV ad says is that "my brand exists, I can afford TV ads, and your social circle has seen my ads" and that works at 10x speed anyway. Hypertargeting makes Internet ads worthless for that sort of positioning. (I buy internet display only right now but I used to work in a mixed team)

by redm 9y ago

jackgolding 9y ago

I've worked with buy side teams and a lot of it is ignorance. I think you'd be horrored with some of the slots that your company gets purchased (and that process, in Australia anyway.) Things like having 1/3 of your 5 figure weekly budget spent on one advert.

scottshapiro 9y ago

Broadcast TV also has 70 years of inertia.

teej 9y ago

The way people handle it is to triangulate on the value based on a couple different sources. For example, you may test your TV ad in a handful of markets first and see if you experience lift in those areas. In addition, you may offer a unique coupon/promo code in that ad and measure uptake. Then you survey customers and ask them how they heard about you. So maybe in your test regions you saw 2% promo uptake, 3% lift, and 5% survey responses saying they saw the ad. You'll pick an estimate somewhere from 2-5% depending on how conservative or aggressive you want to be. That's how you approach it at a small to medium sized company that actually cares about how their marketing budget is spent. I've also seen an ad agency attempt to measure incremental website traffic gain in the test regions during the time slots the ad aired. If your traffic is large enough to have a clean baseline or the ad response is massive this is probably okay, but I'm skeptical of this methodology. If you're a bigger company, there are also ways to get better viewthru conversion metrics. Shazam offers something to this effect, where they allow you to Shazam a commercial. Shazam can then tell you of people who viewed the commercial who went on to convert. Nielsen may sell this too. With TV if you can't get the numbers to work on a pure customer acquisition basis, you'll measure brand recognition and recall in your test regions. If your ad makes an impact there you can then justify spending some of your brand budget in addition to the marketing budget to justify the overall expense.

malyk 9y ago
In the US you can buy national ad spots on major channels pretty cheaply (couple hundred $ per spot) if you buy "remnant" inventory. Basically that means that the network hasn't sold all of their "run my ad in the 9:22am commercial break" spots and so you can try to by a spot in the "morning" day part and it might run in the 9:22am commercial spot...or it might not. But they will tell you if/when the ad ran at some point so you can go back and check your stats. Attribution is kind of a bitch, but the methodology above tends to work very well given the right traffic size/mix.

throwaway284728 9y ago

I work at a major advertiser who's media budgets measure in the billion dollar zone. > I've wondered how advertisers look at broadcast advertising compared to digital. At our levels of spend, we build econometric models to understand media spend effectiveness and spend a fair bit on tracking and vetting the methodologies of trackers. > For one there's no direct correlation (as in a click) with broadcast television. Correct. Frankly, we're interested in valid human viewable reach, not clicks. Not even Impressions though most our buys are CPM today and we want to buy viewable CPMs. We use in fraud detection providers and generally don't pay for Impressions detected as fraud. > Is broadcast advertising so much cheaper than digital that it makes sense to just blanket ads and hope for the best? In many cases yes... In some cases no - especially when there are demographics who simply don't watch much TV which is increasingly a reality. At our scale, digital - even the biggies struggle to give us reach affordably.

hyperpape 9y ago
How many companies have ad budgets in the million dollar zone? And roughly what percent of revenue is that?
shostack 9y ago
Fellow senior buy side guy. Used to work at a top search agency and am now client side. This is the correct answer. At that scale focus groups, surveys, econometric modeling, etc. are what determine success. The harder part is for mid size advertisers who don't have the budget to move the needle to make those things viable, but still need to be branding heavy. Beyond that, digital is way more trackable and often more efficient as a result, even on the branding side.

philipodonnell 9y ago

My understanding is that even broadcast television advertising can be pretty well targeted based on the show's demographic profile, the viewership of the station/channel, known DVR usage trends, and the viewer's location. Surveys on brand awareness and recall can be delivered locally to confirm the ad was seen. The larger the company (large scale means more likely to do broadcast TV advertising), the more they can control other advertising channels in order to test whether the advertising is showing lift locally. There is a whole industry and 50+ years of experience using these things to gauge advertising effectiveness, so the limitations and usefulness is pretty well understood. Then the internet comes along and digital advertising is promoted as being better at all types of advertising (which of course its not). And its new (new is risky/scary) and no one knows if exposure to digital advertising has the same "brand awareness" lift as non-digital, or how to quantify that. The results have been mixed. P&G recently announced they would dramatically scale back social advertising because they determined it wasn't effective. So its on digital to prove its value, and the argument becomes "direct correlation/clicks" is proof, which is fine. Some people accept that and some don't. In 50 years, the limitations and usefulness of digital will be pretty well understood, but in the meantime, advertisers need more than "its new" to trust it.

rhizome 9y ago

Why is digital so conversation obsessed (clicks/purchases) and not focused as much on the branding values (CPM)? I'm not in the profession, but my theory is that what can be measured, will be measured. However, and someone please correct me if I'm wrong, that doesn't mean that any digital strategy is more effective than any print/broadcast strategy from 40 years ago. Keep in mind these are the people pushing for autoplay video.

bcoates 9y ago

It's a mix of two things: The first is searching under the streetlight: both attributable digital ads and non-attributable ads are put into a media plan with (mostly fanciful) estimates and projections as to returns and value before the project starts. Then the campaigns are run, and the attributable ads are scored according to actual results, while everything else is either carried forward on the assumptions the estimates are true, assigned an arbitrary fraction of all sales/traffic, or run through some GIGO model. A lot of personal prejudice and tribal knowledge is applied here. The other, smarter instinct is that while traditional/non-digital advertising can't be finely attributed it can be monitored and audited, and this is vastly more important. Mistakes and underperformance are bounded--fraud is not. A billboard in an inappropriate place with shitty creative at least exists and has some non-zero value not that far from the best possible billboard; but fraudulent Internet advertising can just dump all your money down a hole, never be seen by a relevant human, and do nothing but give you false metrics. DVRs aren't that much of a threat -- most of what a TV ad says is that "my brand exists, I can afford TV ads, and your social circle has seen my ads" and that works at 10x speed anyway. Hypertargeting makes Internet ads worthless for that sort of positioning. (I buy internet display only right now but I used to work in a mixed team)