11 ms·

One of the interesting things about advertising is that nobody really knows if it works, and there's a growing body of evidence that it largely doesn't work. W

One of the interesting things about advertising is that nobody really knows if it works, and there's a growing body of evidence that it largely doesn't work. We are sold on the idea that they work, because someone clicks this ad, we track them, they buy the product, and we attribute the sale to that ad. But we don't know whether the customer would have bought anyway. And some research suggests that, for the most part, they would have. Steve Tadelis at Berkeley wrote a paper in 2013 demonstrating little to no efficacy of nine figures of ad spend at eBay. http://conference.nber.org/confer/2013/EoDs13/Tadelis.pdf http://conference.nber.org/confer/2013/EoDs13/Tadelis.pdf This may not be true of all ads - ads for little-known brands, for example, may reach people who would have never heard of it. But the bulk of advertising bought by large brands may largely be a waste. Advertisers still spend money because almost nobody is getting paid to rain on the parade. There's an endless stream of ad agencies, paid search marketers, marketing department staff themselves, CMOs, media properties, ad nauseam who are all invested in continuing marketing spend. A CMO will almost never declare that his department is useless. So they continue to make the case for their continued relevance.

mc32 5y ago
Organic mentions are the best ads; aside from that dedicated niche review sites which I’ll hit to look at the review and the comments. On the other hand, brand recognition I do believe has value but is hard to quantify. I also never click on an internet ad EXCEPT if it links to a price -not to buy.
- bscphil 5y ago
  I partially disagree: fake (paid) "organic" mentions and fake reviews are the most effective form of marketing. And, obviously, they're much easier for a company to achieve than the real thing. How unsurprising is it that the most effective class of ad spend is the most scummy? The reason (fake) organic marketing is effective is exactly because it fools people into thinking that it isn't marketing. Take paid Amazon reviews for example. This is advertising, just one of the most scummy possible forms of it. It's effective because people like me look to reviews to provide a third-party take on whether or not a product is any good. If I have a choice between two otherwise equivalent products, I'll pick the one with better reviews. But usually this just means rewarding the company that paid more to "advertise" on Amazon. I realize you're probably talking about "real" organic mentions, but those are growing increasingly harder to find. A good example is the last time I bought headphones. I got suckered by a bunch of small review sites that had either been paid or gotten free samples in exchange for reviews. I bought two pairs, and when they arrived the sound from both was distorted. One pair broke within a few months. It was clearly trash that had been pumped using fake organic marketing. When I went back to find the sites that had provided the reviews in the first place, a few of them did mention (in very tiny print) that the reviews were paid. Me, suckered by advertising?! More likely than you'd think. Back in the day, Johnny Carson reading an ad was convincing to the median consumer because they could be led to believe that the ad really was an endorsement of the product. That doesn't work any more, or not nearly as much at any rate. The basic strategy hasn't changed: it's still an attempt to find someone the consumer will trust. This is why, in my opinion, ads have always been and always will be bad, an attempt to subvert human rationality. Johnny Carson endorsing a product didn't mean that the product was good, it meant he was paid to endorse it. A review site endorsing a product nowadays doesn't mean that the product is good, it means they're paid to endorse it. In both cases the breakdown between what the consumer believes is happening and what is actually happening is the very thing advertising is trying to create, its raison d'être.
  - mc32 5y ago
    I generally agree with your gist; however, I would point out that there are niche review sites run by enthusiasts (whether computer parts, cameras, tools, gadgets, clothing, and so on.) and there are the SEO types that are near-worthless. Enthusiast sites have a range of community-only and not ads, some have affiliate links and some have paid reviews and so on. So you have to know how the site operates and take that into account. The niche ones have mods which allow genuine dissent as well as other recommendations from competing solutions.
    - shkkmo 5y ago
      So you happen to have a list of these niche sites or a link to a site that reviews the review sites?
      - mc32 5y ago
        I am thinking about the car enthusiast sites/forums (where you can search for information how to fix minor issues), also places like cpubenchmark, tomshardware, anadtech, dpreview, shoptoolreview and so on. You have to get to know the forum and how they make money, whether it's a labor of love or main source of income, how transparent they are about sponsorship, etc.
- katzgrau 5y ago
  Genuine social proof is always the best. Smart marketers don't try to get you to buy from an ad click. They want to move you one step down the funnel, which is usually getting your email/contact info if it's the type of product that has a longer nurturing process and sales cycle.
katzgrau 5y ago
Digital marketing performance is increasingly measurable and there's more data than ever proving that it can be extremely effective. You can certainly waste a lot of money doing it too. Getting the right message in front of the right user at the right time is a lot harder than people think it is. The above is where there's a lot of smoke and mirrors in ad tech. But for advertising as a whole, it's not a massive industry because it doesn't work.
- cm2012 5y ago
  Yeah, it's incredibly easy to prove it works. Incrementally can get hard at scale but that's a different question.
- meowface 5y ago
  >Getting the right message in front of the right user at the right time is a lot harder than people think it is. Guess that's why we need those brightest minds of our generation on the problem, as the saying goes, right? That sentence gives me extreme Matthew McConaughey-in-Wolf of Wall Street vibes. After seeing you wrote that, I immediately assumed you very likely work in advertising. Sure enough... I have no clue whether anything you've written there is accurate or not. Either way, this whole thing and this whole industry just looks, sounds, and feels gross and intrinsically empty. I get a much better and more wholesome feeling from the cryptocurrency industry, even.
  - ipaddr 5y ago
    If the greatest minds of society are going for the easiest payday perhaps they are not really our greatest minds. Perhaps they are the smart ones by cashing in but the great minds with the great ideas they are not.
    - katzgrau 5y ago
      Ad-tech involves big data, lots of volume, and tons of innovation and interesting challenges. There's usually a good paycheck too, because there are lots of companies that need smart people to work on the above. Greatest minds? Maybe there are some, you'll find them everywhere. But I'd worry that the greatest minds are toiling away in academia chasing grants or working on things that they don't want to work on. In the paraphrased words of Don Draper (Mad Men character), nobody grows up wanting to be in advertising. It just kind of happens.
manigandham 5y ago
We definitely know that it works. We generate petabytes everyday proving it does. That being said there’s plenty of waste due to politics, bureaucracy, bad tech and general mismanagement. It’s important to separate the concept from the implementation.
- meowface 5y ago
  As a consumer, why does it very strongly feel like in my case, it never, ever works, then? And why does pretty much everyone else I've ever heard talk about this say the exact same thing? Are we being 5D-chess'd and it's working on all of us subliminally and subconsciously and we just think we're immune? Or is there some other huge population of people out there it's working on who actually somehow are real manifestations of the unfathomably-cringeworthy-to-utter "sheeple" / "NPC" trope or something? One area I can see is more subtle word-of-mouth kinds of influence (organic or astroturfed), where someone may see a lot of their friends or people they admire using and recommending something. I can definitely be influenced in that way in some cases - that's kind of how interest in almost anything works, isn't it? (For a non-profit example, would I have become interested in Rust if I didn't see it positively mentioned thousands of times on websites I use for the past decade? Probably not.) But for traditional ads like banners and commercials and billboards, I perceive zero efficacy in my own case.
  - dougmwne 5y ago
    I used to think like you, but now I think it's 5D chess. Marketing is like air, so ever present you can't even see it. If you don't buy anything and live off your homesteading, then marketing probably doesn't have any effect on you. If you live a life anything like the average westerner, you spend hundreds of thousands of dollars on things you had to be taught to desire. It's in your TV shows, your news articles, your Instagram feed, your word of mouth, your everything unless you are in the forest (but not looking at anyone else's hiking gear brands). Our economy is based on you consuming many many times more goods and services than your great grandparents, and it wasn't cheap to train you to do that.
    - 5y ago
55555 5y ago
oh my god discussions about advertising on hackernews are like a bunch of plumbers talking about software engineering. these threads should honestly just be used as Dunning-Kruger honeypots. we should post about advertising once a week and then just ban everyone who says "there is no evidence that advertising works." I have a pixel on my thank-you page. I manually delete placements and optimize creatives and angles and funnels and demographic targeting until it goes from making no sales to ludicrously profitable. This is an entire industry you know nothing about. Big brands can be dumb but they do mostly have the tools to not be.
sfifs 5y ago
Not really correct. Advertising incrementality is pretty easy to determine by in-market tests and statistical modeling or causal inference techniques. Feature Engineering for effective model construction does require a body of significant domain knowledge and so your average academic Post-Doc or FAANG data science type tends to struggle in this space unless they learn quickly (I've hired from both before). The space is dominated by large research vendors, specialist consultancies or in-house teams with accumulated institutional knowledge. All this costs though and pays out only if your marketing budgets are in the tens to hundreds of millions of dollars zone. In a lot of Tech industry firms, the in-house expertise just isn't there to do this stuff and that possibly breeds skepticism. I've had fairly junior members of my teams head out to FAANG who've been shocked at how limited the depth of knowledge is.
magicroot75 5y ago
There's an extremely well defined industry surrounding conversion lift. Which is exactly what you're talking about. Advertisers use holdout groups to test what % lift was correlated to an ad. Using large enough datasets, it's possible to have strong statistical confidence in the effects of ads on user behavior.

mc32 5y ago

Organic mentions are the best ads; aside from that dedicated niche review sites which I’ll hit to look at the review and the comments. On the other hand, brand recognition I do believe has value but is hard to quantify. I also never click on an internet ad EXCEPT if it links to a price -not to buy.

bscphil 5y ago
I partially disagree: fake (paid) "organic" mentions and fake reviews are the most effective form of marketing. And, obviously, they're much easier for a company to achieve than the real thing. How unsurprising is it that the most effective class of ad spend is the most scummy? The reason (fake) organic marketing is effective is exactly because it fools people into thinking that it isn't marketing. Take paid Amazon reviews for example. This is advertising, just one of the most scummy possible forms of it. It's effective because people like me look to reviews to provide a third-party take on whether or not a product is any good. If I have a choice between two otherwise equivalent products, I'll pick the one with better reviews. But usually this just means rewarding the company that paid more to "advertise" on Amazon. I realize you're probably talking about "real" organic mentions, but those are growing increasingly harder to find. A good example is the last time I bought headphones. I got suckered by a bunch of small review sites that had either been paid or gotten free samples in exchange for reviews. I bought two pairs, and when they arrived the sound from both was distorted. One pair broke within a few months. It was clearly trash that had been pumped using fake organic marketing. When I went back to find the sites that had provided the reviews in the first place, a few of them did mention (in very tiny print) that the reviews were paid. Me, suckered by advertising?! More likely than you'd think. Back in the day, Johnny Carson reading an ad was convincing to the median consumer because they could be led to believe that the ad really was an endorsement of the product. That doesn't work any more, or not nearly as much at any rate. The basic strategy hasn't changed: it's still an attempt to find someone the consumer will trust. This is why, in my opinion, ads have always been and always will be bad, an attempt to subvert human rationality. Johnny Carson endorsing a product didn't mean that the product was good, it meant he was paid to endorse it. A review site endorsing a product nowadays doesn't mean that the product is good, it means they're paid to endorse it. In both cases the breakdown between what the consumer believes is happening and what is actually happening is the very thing advertising is trying to create, its raison d'être.
- mc32 5y ago
  I generally agree with your gist; however, I would point out that there are niche review sites run by enthusiasts (whether computer parts, cameras, tools, gadgets, clothing, and so on.) and there are the SEO types that are near-worthless. Enthusiast sites have a range of community-only and not ads, some have affiliate links and some have paid reviews and so on. So you have to know how the site operates and take that into account. The niche ones have mods which allow genuine dissent as well as other recommendations from competing solutions.
  - shkkmo 5y ago
    So you happen to have a list of these niche sites or a link to a site that reviews the review sites?
    - mc32 5y ago
      I am thinking about the car enthusiast sites/forums (where you can search for information how to fix minor issues), also places like cpubenchmark, tomshardware, anadtech, dpreview, shoptoolreview and so on. You have to get to know the forum and how they make money, whether it's a labor of love or main source of income, how transparent they are about sponsorship, etc.
katzgrau 5y ago
Genuine social proof is always the best. Smart marketers don't try to get you to buy from an ad click. They want to move you one step down the funnel, which is usually getting your email/contact info if it's the type of product that has a longer nurturing process and sales cycle.

katzgrau 5y ago

Digital marketing performance is increasingly measurable and there's more data than ever proving that it can be extremely effective. You can certainly waste a lot of money doing it too. Getting the right message in front of the right user at the right time is a lot harder than people think it is. The above is where there's a lot of smoke and mirrors in ad tech. But for advertising as a whole, it's not a massive industry because it doesn't work.

cm2012 5y ago
Yeah, it's incredibly easy to prove it works. Incrementally can get hard at scale but that's a different question.
meowface 5y ago
>Getting the right message in front of the right user at the right time is a lot harder than people think it is. Guess that's why we need those brightest minds of our generation on the problem, as the saying goes, right? That sentence gives me extreme Matthew McConaughey-in-Wolf of Wall Street vibes. After seeing you wrote that, I immediately assumed you very likely work in advertising. Sure enough... I have no clue whether anything you've written there is accurate or not. Either way, this whole thing and this whole industry just looks, sounds, and feels gross and intrinsically empty. I get a much better and more wholesome feeling from the cryptocurrency industry, even.
- ipaddr 5y ago
  If the greatest minds of society are going for the easiest payday perhaps they are not really our greatest minds. Perhaps they are the smart ones by cashing in but the great minds with the great ideas they are not.
  - katzgrau 5y ago
    Ad-tech involves big data, lots of volume, and tons of innovation and interesting challenges. There's usually a good paycheck too, because there are lots of companies that need smart people to work on the above. Greatest minds? Maybe there are some, you'll find them everywhere. But I'd worry that the greatest minds are toiling away in academia chasing grants or working on things that they don't want to work on. In the paraphrased words of Don Draper (Mad Men character), nobody grows up wanting to be in advertising. It just kind of happens.

manigandham 5y ago

We definitely know that it works. We generate petabytes everyday proving it does. That being said there’s plenty of waste due to politics, bureaucracy, bad tech and general mismanagement. It’s important to separate the concept from the implementation.

meowface 5y ago
As a consumer, why does it very strongly feel like in my case, it never, ever works, then? And why does pretty much everyone else I've ever heard talk about this say the exact same thing? Are we being 5D-chess'd and it's working on all of us subliminally and subconsciously and we just think we're immune? Or is there some other huge population of people out there it's working on who actually somehow are real manifestations of the unfathomably-cringeworthy-to-utter "sheeple" / "NPC" trope or something? One area I can see is more subtle word-of-mouth kinds of influence (organic or astroturfed), where someone may see a lot of their friends or people they admire using and recommending something. I can definitely be influenced in that way in some cases - that's kind of how interest in almost anything works, isn't it? (For a non-profit example, would I have become interested in Rust if I didn't see it positively mentioned thousands of times on websites I use for the past decade? Probably not.) But for traditional ads like banners and commercials and billboards, I perceive zero efficacy in my own case.
- dougmwne 5y ago
  I used to think like you, but now I think it's 5D chess. Marketing is like air, so ever present you can't even see it. If you don't buy anything and live off your homesteading, then marketing probably doesn't have any effect on you. If you live a life anything like the average westerner, you spend hundreds of thousands of dollars on things you had to be taught to desire. It's in your TV shows, your news articles, your Instagram feed, your word of mouth, your everything unless you are in the forest (but not looking at anyone else's hiking gear brands). Our economy is based on you consuming many many times more goods and services than your great grandparents, and it wasn't cheap to train you to do that.
  - 5y ago

55555 5y ago

oh my god discussions about advertising on hackernews are like a bunch of plumbers talking about software engineering. these threads should honestly just be used as Dunning-Kruger honeypots. we should post about advertising once a week and then just ban everyone who says "there is no evidence that advertising works." I have a pixel on my thank-you page. I manually delete placements and optimize creatives and angles and funnels and demographic targeting until it goes from making no sales to ludicrously profitable. This is an entire industry you know nothing about. Big brands can be dumb but they do mostly have the tools to not be.

sfifs 5y ago

Not really correct. Advertising incrementality is pretty easy to determine by in-market tests and statistical modeling or causal inference techniques. Feature Engineering for effective model construction does require a body of significant domain knowledge and so your average academic Post-Doc or FAANG data science type tends to struggle in this space unless they learn quickly (I've hired from both before). The space is dominated by large research vendors, specialist consultancies or in-house teams with accumulated institutional knowledge. All this costs though and pays out only if your marketing budgets are in the tens to hundreds of millions of dollars zone. In a lot of Tech industry firms, the in-house expertise just isn't there to do this stuff and that possibly breeds skepticism. I've had fairly junior members of my teams head out to FAANG who've been shocked at how limited the depth of knowledge is.

magicroot75 5y ago

There's an extremely well defined industry surrounding conversion lift. Which is exactly what you're talking about. Advertisers use holdout groups to test what % lift was correlated to an ad. Using large enough datasets, it's possible to have strong statistical confidence in the effects of ads on user behavior.