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?

#111

Post by buboard » Wed, Nov 06, 2019, 9:37 PM 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…

> Digital direct response advertising allows very accurate ROI calculations

How many ads are direct response? How many are simply clicked because they are just above the same website's entry in search results (i.e. users were already heading to that site, the ad was just there as google tax)? How many legitimate clicks on an ad that did not led to sale are discarded by google as "invalid clicks" (my adsense report has 20% invalid clicks) ? It's kind of shocking that the internet's most lucrative market does not answer those

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

#112

Post by graeme » Wed, Nov 06, 2019, 9:37 PM UTC

jessefrederik wrote: post #100

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…

Hi this is the author of the article. You're just making the exact mistake this article is about: you assume people who click on your ad and then buy something, bought something BECAUSE of your ad. From the article: "Suppose Luigi’s Pizzeria hires three teenagers to hand out coupons to passersby. After a few weeks of flyering, one of the three turns out to be a marketing genius. Customers keep showing up with coupons…

This only applies if the advertiser could be better expected to get most of the business from those searches.

If someone searches "nike shoes" they might well get them on someone else's site. Whereas someone in the pizzeria is already buying at the pizzeria.

I agree it's important to measure, but I think your concern is overblown for most companies and most non-branded searches. Assuming those companies are measuring.

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

#113

Post by jameslk » Wed, Nov 06, 2019, 9:40 PM UTC

jessefrederik wrote: post #100

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…

Hi this is the author of the article. You're just making the exact mistake this article is about: you assume people who click on your ad and then buy something, bought something BECAUSE of your ad. From the article: "Suppose Luigi’s Pizzeria hires three teenagers to hand out coupons to passersby. After a few weeks of flyering, one of the three turns out to be a marketing genius. Customers keep showing up with coupons…

The effects may be low but what is the alternative? You suggest there's a digital advertising bubble but to have a bubble with the assumption it will burst, you need to show that advertisers can achieve a better return by doing something else.

There's no digital marketing bubble more than there is a traditional advertising bubble. There IS however a conflict of interest in advertising that's caused by the agency which performs the advertising being tasked with showing a return on investment.

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

#114

Post by graycat » Wed, Nov 06, 2019, 9:41 PM UTC

jessefrederik wrote: post #76

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.

Hi, I'm the author of the article, and I don't assume all online advertising is Google search. I also cite research on display advertising (meta-study of 432 display experiments on Google): https://www.ssrn.com/abstract=2701578 And on Facebook advertising: https://www.kellogg.northwestern.edu/faculty/gordon_b/files/... The problem with all online advertising is that there are huge selection effects, which are hard to…

> using conventional statistical methods.

Experimental design, analysis of variance are such. E.g., for the farmers and from the corn fields and hog pens of Iowa:

George W. Snedecor and William G. Cochran, Statistical Methods, The Iowa State University Press, Ames, Iowa.

These methods have been widely used in the social sciences -- e.g., my wife, Ph.D. in mathematical sociology from Hopkins, got quite good with that material. The field is quite serious and mature and goes well beyond just A/B testing.

For the practical challenges of the article, academic fields closer than economics include statistics and optimization.

For the Lagrange multipliers in the article, those likely would be from the Kuhn-Tucker (Karush-Kuhn-Tucker) conditions. There without some special assumptions, e.g., having to do with cases of convexity, the conditions are only necessary for optimality and not sufficient. Generally in practice, it is more difficult to get sufficient conditions.

Yes, correlation does not necessarily mean causality. Usually showing causality needs a mechanism; in practice showing causality just from data and/or statistical methods is difficult and rare. But in practice, correlation can be powerful enough to take money to the bank.

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

#115

Post by nopriorarrests » Wed, Nov 06, 2019, 9:41 PM UTC

jessefrederik wrote: post #100

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…

Hi this is the author of the article. You're just making the exact mistake this article is about: you assume people who click on your ad and then buy something, bought something BECAUSE of your ad. From the article: "Suppose Luigi’s Pizzeria hires three teenagers to hand out coupons to passersby. After a few weeks of flyering, one of the three turns out to be a marketing genius. Customers keep showing up with coupons…

>It's the same in online marketing, you often target people who are searching for say a pair of shoes.

This is funny, because on HN you often see complaints about online advertising being horrifically bad because "it suggests me the products I have zero interest in", hence "online advertising sucks".

But that's the whole point! These people are complaining that they are not getting pizza coupons when they are entering pizzeria!

If it will suggest you something you are interested in already, chances are, you will buy it anyway. And if you will be suggested something strange and you will buy it, it will be incremental customer and incremental sale, and you might even become a regular client.

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

#116

Post by chokma » Wed, Nov 06, 2019, 9:43 PM UTC

Coming from the other end (after someone clicked an ad), I got to say: ads work well. The metric used @work is: cost of sale.

The merchant pays for each click on the ad that is redirected to his product page. What seems like a recipe for wasting incredible amounts of money on an army of click bots is working - because the merchant defines a target cost of sale for those clicks, meaning for example he expects a sale worth 100$ when spending 10$ on ads (a cost of sale of 10%).

And the merchant will set a budget and won't repeat/continue a campaign beyond those costs, so it is in the ad publisher's best interest to deliver real paying customers instead of an army of bots.

Cost of sale is of course different for each product category - clothes and perfume have huge margins (so the merchant may be willing to pay 30% of the final sale price for advertising costs), computer parts less so.

In the case of ebay and amazon and other large retailers it's common to not pay per click but for a percentage of an order's total volume (x% of shopping basket's value once an order is placed). In that case, we have to trust the retailer to deliver correct sales data to us.

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

#117

Post by jessefrederik » Wed, Nov 06, 2019, 9:47 PM UTC

graeme wrote: post #112

jessefrederik wrote: post #100

Earlier quoted context omitted.

Hi this is the author of the article. You're just making the exact mistake this article is about: you assume people who click on your ad and then buy something, bought something BECAUSE of your ad. From the article: "Suppose Luigi’s Pizzeria hires three teenagers to hand out coupons to passersby. After a few weeks of flyering, one of the three turns out to be a marketing genius. Customers keep showing up with coupons…

This only applies if the advertiser could be better expected to get most of the business from those searches. If someone searches "nike shoes" they might well get them on someone else's site. Whereas someone in the pizzeria is already buying at the pizzeria. I agree it's important to measure, but I think your concern is overblown for most companies and most non-branded searches. Assuming those companies are measuring…

This is an empirical question. I suggest you read the research I cite. For Facebook: https://www.kellogg.northwestern.edu/faculty/gordon_b/files/... For display ads: https://www.aeaweb.org/conference/2014/retrieve.php?pdfid=98... For search: https://pubsonline.informs.org/doi/10.1287/mksc.2017.1065

The problem is: yes, there's only a tiny chance somebody who searches 'nike shoes' ends up in your store. But the probability that someone ends up in your store BECAUSE he saw your ad is even smaller. So if you don't correct for selection effects your going to overstate ad effects by orders of magnitude.

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

#118

Post by bryanmgreen » Wed, Nov 06, 2019, 9:49 PM UTC

Roritharr wrote: post #83

bryanmgreen wrote: post #74

Earlier quoted context omitted.

I think that's a self-preservation tactic for many that often backfires. Most non-marketing management people want immediate and huge results... which are both hard to quantify and impossible to get without overspending or having an insanely lucky idea that goes viral. Naturally this leads to asking for a bigger budget and promising bigger results, then KPIs aren't met for X, Y, Z reasons, then they're forced to just…

What do marketing management people want?

I mean, immediate huge results are great if they happen.

Really, they want what everyone else wants - ability to test ideas and executions before committing to a budget or unreasonable KPIs.

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

#119

Post by tempestn » Wed, Nov 06, 2019, 9:50 PM UTC

chokma wrote: post #116

Coming from the other end (after someone clicked an ad), I got to say: ads work well. The metric used @work is: cost of sale. The merchant pays for each click on the ad that is redirected to his product page. What seems like a recipe for wasting incredible amounts of money on an army of click bots is working - because the merchant defines a target cost of sale for those clicks, meaning for example he expects a sale w…

But doesn't this ignore the selection effect discussed in the article? IE to use your numbers say you get a $100 sale for every $10 spent on ads. And say your profit margin (outside of ads) on that sale is 20%, so you earn $20-$10 = $10 in profit.

However, what if 3/4 of the people who clicked on your ads would have found you anyway, through organic search results or other means? Then that $10 spend only actually generated $5 in net income before ads, or a $5 loss after the cost of the ads.

It's tough to say what that number is without doing some testing, but the point is that you can't simply treat all the business coming in via ads as attributable to the ad spend.

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

#120

Post by buboard » Wed, Nov 06, 2019, 9:50 PM UTC

ACow_Adonis wrote: post #30

If you think about it, the very line that the smartest minds of our generation are working on making people click on things is itself marketing designed by marketers to sell the prestige/status of working for a big advertising company. That actually is genius in its own way... but just as the quants in finance during the early booms were "the smartest guys in the room, until they weren't and you realised they all wer…

> smartest minds of our generation are working on making

Considering the deterioration in software functionality that i m experiencing the past years across the board (from superheavy web apps to mobile apps that for no reason remove features to desktop apps that function worse than even mobile apps) - i m not even sure they re employing the " smartest minds of our generation"