What do we really know about the effectiveness of digital advertising? • Hacker News Board

Re: What do we really know about the effectiveness of digital advertising?

#281

Post by davidedicillo » Thu, Nov 07, 2019, 4:59 AM UTC

tomrod wrote: post #230

davidedicillo wrote: post #17

Overstated. This article assumes that: - All online advertising is Google search - That brand search advertising is representative of the entire industry There are plenty of companies making double-digit ROAS (return over ad spend), especially direct response advertisers.

As the other commentator identified, you work @ FB -- could you please provide the following three scatter plots to support or disprove your claims?: (1) Size of Company vs. ROA (2) Advertising Expenditure by Company vs. ROA (3) [Advert %] vs. ROA I hypothesize you'll see a rapidly decreasing return on advertising due to size, and would love to be proven wrong.

Not at FB anymore but I would assume your hypothesis isn't too far from the truth. It's also true that there are hundreds of thousands of small businesses that thrive thanks to online advertising.

Re: What do we really know about the effectiveness of digital advertising?

#282

Post by mochomocha » Thu, Nov 07, 2019, 5:01 AM UTC

tacon wrote: post #99

This is a rather unimpressive article. Yes, brand advertising is a disaster, just like it has always been. But to say that all digital advertising has unknown ROI is ridiculous. Digital direct response advertising allows very accurate ROI calculations. (Admittedly, a lot of advertisers are not applying the tools to their advantage. They call that the "Google stupid tax". Ditto these days on Facebook.) And direct resp…

Google removes targeting knobs because despite what marketers think, humans consistently underperform algorithms when it comes to targeting. Typically, the people tweaking the knobs have reverse incentives because they don't want their jobs to be automated so they love these knobs. But they are just constraints on the bidding systems that end up delivering poorer performance of the campaigns.

YMMV but for the vast majority of campaigns, the algorithm does know better.

Disclaimer: I've built bidding systems for multiple ad companies, so I'm surely biased towards "the algorithm knows better".

Re: What do we really know about the effectiveness of digital advertising?

#283

Post by countryqt30 » Thu, Nov 07, 2019, 5:09 AM UTC

"The experiment continued for another eight weeks. What was the effect of pulling the ads? Almost none. For every dollar eBay spent on search advertising, they lost roughly 63 cents , according to Tadelis’s calculations." - that issue isn't related to a problem with digital advertising, but to horrible financial management

Re: What do we really know about the effectiveness of digital advertising?

#284

Post by countryqt30 » Thu, Nov 07, 2019, 5:13 AM UTC

How successful companies do ads: ONLY invest in the ad if you have at least 3X ROI to cut out all the "bullshit tiny 1% improvement" ideas, and you don't run into any problems the article mentions.

eBay ran into millions of problems, because they focussed on "how could we improve sales by 0.0001%" and is the ROI of this ad 0.001% or 0.002%?

Re: What do we really know about the effectiveness of digital advertising?

#285

Post by sifar » Thu, Nov 07, 2019, 5:37 AM UTC

austincheney wrote: post #245

buboard wrote: post #127

Great to see an article like this. This subject is almost suspiciously under-researched by journalists. There 's definitely something rotten in the promises that internet advertising has become somehow more effective. If that were true, wouldn't businesses notice? Wouldn't they spend a bigger part of their revenue on ads ? Wouldn't that return even bigger ROIs? Yet businesses spend roughly the same for over a century…

> Wouldn't they spend a bigger part of their revenue on ads ? That isn't the best question. Businesses spend stupid money on advertising in the mere hope and faith of grabbing attention. Sites like Google (the actual search engine) are really good at measuring effectiveness because their entire revenue model is built on micro-auctions measured by engagement. The better question is: If digital advertising were so effe…

This is the best articulation I have read explaining the phenomenon. Thank you. Can I steal this ?

If only business leaders can wrap their head around recursive feedback and have the courage to account for it.

Re: What do we really know about the effectiveness of digital advertising?

#286

Post by nl » Thu, Nov 07, 2019, 6:05 AM UTC

DataWorker wrote: post #199

nopriorarrests wrote: post #133

Earlier quoted context omitted.

Hey Jesse, I spent 6 years working in ad-tech, and I was researching the incrementality effect of ads, selection bias and stuff like that for last 3 years. I want to let you know that your article is good. Online advertising is not a bubble, but your article perfectly explains why it is much harder than it seems.

The thing about bubbles is that nobody knows they exist. So an insider such as yourself saying, “there is no bubble,” is actually strong evidence of said bubble. Necessary but not sufficient.

This isn't really the case. Bubbles usually have plenty of people claiming they are bubbles, and people on the inside privately acknowledging it.

In online advertising there are few people on the inside who think that non-brand advertising is a bubble. It just works too well, in too many places, and too many people are making money using it. It's reliable and effective, and those just aren't attributes associated with bubbles.

Yeah, attributing everything to the last-click isn't the best measurement in the world. But as pointed out elsewhere there are techniques to avoid it.

And there are enough direct-to-consumer brands which launch on Instagram and do nothing else to be able to see how effective it is.

Re: What do we really know about the effectiveness of digital advertising?

#287

Post by dehrmann » Thu, Nov 07, 2019, 6:45 AM UTC

vmurthy wrote: post #219

Somewhere near the very bottom of the article is this gem. Emphasis mine. "And it wasn't because it was working for them. ... But there was one group of advertisers with a valid reason to purchase own brand name ads. Not that they were any more effective for this group, but because their competitors were "crowding them out" by buying ads that targeted the brand owner's name (so you search for "Bestbuy" but a sponsore…

This is one of my gripes with Google and not being evil. Allowing competitors to outrank you for your brand if they pay and you don't is racketeering. Allowing this is so blatantly wrong I'm not sure why it doesn't get more attention.

Re: What do we really know about the effectiveness of digital advertising?

#288

Post by asdff » Thu, Nov 07, 2019, 6:48 AM UTC

JohnFen wrote: post #189

Spooky23 wrote: post #162

Earlier quoted context omitted.

I disagree. Most of the people I grew up with today know that it’s Miller time, that sometimes you feel like a nut, sometimes you don’t, or that coke is the real thing. They learned that from TV. Online ads? They don’t trigger anything for me. They either need to deliver information contextually, deliver preferred placement for something I’m already looking for, or be deceptive.

> I grew up with today know that it’s Miller time, that sometimes you feel like a nut, sometimes you don’t, or that coke is the real thing. True. I acknowledge here that anecdote is not a synonym for data, but I will offer that despite those phrases having been seared into my brain for most of my life, I don't actually drink Miller, eat Almond Joy or Mounds, or drink Coke.

If you were the kind of person who drank light beer, candy, and soda, I bet that jingle would spark something.

Ad quality these days has gone down for sure. Everything is formulaic and established and safe, yet nearly every major ad over the course of advertisement history was major because it was the opposite of these. I guess quantity has gone up to compensate.

Re: What do we really know about the effectiveness of digital advertising?

#289

Post by asdff » Thu, Nov 07, 2019, 6:52 AM UTC

bostik wrote: post #65

ngngngng wrote: post #8

> a Super Bowl ad cost three million dollars. Why? Because that’s how much it cost. What does it yield? Who knows. This article is awesome. In my free time I'm building an advertising platform that topic matches ads off the content of the web pages. No tracking of users. I'm hoping to allow people to host and run their own all inclusive self contained advertising platform for just a licensing fee. I also have a way t…

> I also have a way to beat adblockers that I'm afraid of bringing to light Proxying ads through as first-party content? If you really are invested in making them work ethically, consider these ground rules: https://bostik.iki.fi/aivoituksia/random/no-stalking.html

Even doing that you can block ads trivially. I can set up a filter right now to block any element on the hn page forever if I wanted. You can always serve an ad to most people, but anyone willing to slip down the rabbit hole and end up with u block origin will be able to foil any advert with a right click.

Re: What do we really know about the effectiveness of digital advertising?

#290

Post by quixoticaxolotl » Thu, Nov 07, 2019, 7:02 AM UTC

I'm still dubious about the implication that digital advertisement is fundamentally broken.

My issue is that the article doesn't demonstrate selection effects dominating over advertising effects for companies with no major brand identity. The examples cited to back the case - eBay (which is huge), "large" retailers in the case of Facebook ads, "major U.S. retailers and brokerages with millions of customers" per Lewis and Rao 2015 - are all established companies that have already received substantial presence. Where are the experiments that demonstrate selection effects dominating over advertising effects for companies with no pre-existing brand identity e.g. The Correspondent (sorry, I couldn't resist :P)?

Second issue: the article demonstrates convincingly that brand-keyword ads are broken, which is great. But at the same time brand-keyword ads are not (to my knowledge as a former engineer at an SEO company, so it could be imperfect) the primary kind of keyword ads. A brand-keyword ad is when someone types in "eBay" and you get paid eBay links - but the majority of keyword ads (again, this is from memory) are targeted keyword ads e.g. someone types in "fur coat" and Macy pays Google to put a link to its "fur coats" pages. Equally interesting, multiple companies can pay for the same targeted keywords, meaning you can have multiple links pointing off to different sites on the same page.

The pizzeria analogy explains why brand-keyword ads make no sense monetarily (someone searching for your brand specifically will find you anyway), but it falls apart for the targeted keyword space, where people are searching by items but not by brand.

Given that much of SEO and digital marketing are target-based rather than brand-based, I'm not sold yet on the ineffectiveness of digital advertising overall.