How effective is internet advertising?

This new paper by Tom Blake, Steven Tadelis, and Chris Nosko is not entirely reassuring for the future of journalism, but it confirms what I have long suspected:

Internet advertising has been the fastest growing advertising channel in recent years with paid search ads comprising the bulk of this revenue. We present results from a series of large scale field experiments done at eBay that were designed to measure the causal effectiveness of paid search ads. Because search clicks and purchase behavior are correlated, we show that returns from paid search are a fraction of conventional non-experimental estimates. As an extreme case, we show that brand-keyword ads have no measurable short-term benefits. For non-brand keywords we find that new and infrequent users are positively influenced by ads but that more frequent users whose purchasing behavior is not influenced by ads account for most of the advertising expenses, resulting in average returns that are negative.

The NBER version is here, an ungated version is here.


I often wonder if traditional advertising is any more effective than internet advertising. And if it's not, what result if we could actually measure its effects?

If you look at the statistical approach taken in (discussion: ), you'll see that it's almost impossible to show that online advertising works and use it to profit; so unless offline advertising is dramatically, orders of magnitudes, more effective than online advertising, the reduced demographic info & increased measurement error compared to online advertising means that the conclusions will hold doubly for offline advertising too.

And here lies the fundamental contradiction of online advertising:

POSITIVE: You can measure the effectiveness of your ads.
NEGATIVE: You can measure the effectiveness of your ads and discover they're not effective.

But for some reason a black and white ad for some watch brand I don't care about in a random page of a newspaper that I might, just might flip past is worth lots.

Ignorance is bliss.

' Half the money spent on internet advertising is wasted, but no one knows which half '

it seems like it's a lot more than half is wasted...

Amazon's web advertising remains highly effective, I'm sure, with a level of granularity and self-interest that ensures results.

Especially since Amazon only pays for success, creating the sort of win-win synergy that a good economist is certain to favor.

I saw a related paper linked on Robin Hanson's site a couple months ago.

Without necessarily critiquing the technical claims of either paper, I want to make clear that at least certain advertising platforms (in particular the one I work for) have demonstrated (very) positive return on ad spend under extremely controlled conditions, in particular using rigorous intent-to-treat holdouts and end-to-end measurement. Now, perhaps (probably) not every platform creates as salient of an ad experience as my publisher, and certainly not every platform has the resources to work with advertisers to perform such measurement, so my data point may not be typical.

Lots of different models of internet advertising, it depends on what you measure.

Google AdWords focused very heavily on 'results' and had their system shutting down ads that people didn't click on, but the advertising industry pushed back against this, they WANTED ads (or wanted to SELL ads..) that no one ever clicked on, calling it "branding."

But the pay-for-performance model is increasingly prevalent ... revenue-shares, for instance, with the publishers. This isn't a perfect model since there are tradeoffs, people who would have bought anyway but whose purchases get traced back to an individual advertiser. Measurement here is hard.

There's a trend though towards measurement and payments based on metrics and demonstrated results, at least in financial services for sure. Sometimes advertisers may mis-specify the metrics, so may not really get what they're looking for, but that's less an issue with advertising than the buyers of advertising.

In any case, the more measurable the results off an ad the higher the premium the ad space will command. You can buy worthless ads with tons of impressions for next to nothing. Or you can sell display ads at $8 or even $10 per 1000 on a pure impression basis. Move to pay for performance and you can see hundreds of dollars per qualified lead.

Ebay is not at all a typical advertiser. Ebay is totally dominant in the Auction vertical in US (+80% share) with most of that competition being in very specific auction verticals. This means the next-best competitor is often very inferior. When someone searches for "ebay [something]" Ebay dominates the result page as there are very few competitors for auction-type markets.

However this is not at all typical. Even a huge online seller like Amazon faces far more competition in retail. Though a user might be seeking Amazon when they type "amazon digital camera", there are lots of reasonable competitors seeking space in the organic and paid result page. If Amazon stops bidding, they face a real trade-off that users will be exposed to competitors offering almost exactly the same merchandise at almost exactly the same prices. Ebay faces no such competition in the used/auction market.

I suspect that if this research were replicated with another, less vertical dominant brand, the results would be very different.

Al Broadman here, economist. Used to run a blog about economics until I found it to be a huge waste of time and productivity.

I did a piece on advertising a couple years back that showed the same results but included all forms of advertising.

The gist was this.

Basically people ignore it. Its hit or miss, luck of the draw, being in the right place at the right time.

The internet and search has changed the whole arena for people. No longer are they forced to trust little ads competing for their attention to get information about consumption opportunities. If someone has a need, the solution is but a Google away. First page positioning used to be very important but as more of it becomes paid inclusion and money driven and as consumers are getting smarter and more suspicious of big money, that factor is becoming less and less irrelevant.

So if advertising as we have had so far is not working then what will? Personal and Local are the two keys. Local consumption is a huge trend right now. People are concerned about the environment, employment, and ethics so local consumption gives them an opportunity to see how a producer is handling those three issues and make their choices accordingly. Shaming a producer is much more effective if that producer goes to one of the local churches and attended the same town hall meetings and works well as a means to control the more anti-ethical behavior that capitalism can inspire.

I'm a well known advertising hater and will go so far as to turn my head away from it to keep the propaganda from affecting my point of view. I'm true individual and consider my brain to be a private sanctum that no marketing company's offerings are allowed. I make my choices based on need and research, not shiny promotional pieces designed to tug at my emotions while twanging my behavioral responses based on new neurological science. Oh no, we won't have that in my head.

I recently went to a local BBQ gathering in my home town and local farmers as well as backyard BBQ's were showing off their skills for the community. I eat way to much BBQ and sampled way to many BBQ sauces. But what I did do was come home with quite a few bottles of BBQ sauce I would have never given a second thought to. Why you ask?

The interactions were personal and local. The person selling the stuff was right across from me and actually used the stuff and was giving out or selling plates in my home town. I felt like I was supporting a neighbor instead of a transnational corporation who only panders to the consumer elite don't care or notice their anti-value-added nonsense such as brand names, patients, copyrights, CEO wages that are so insane they make you wanna slap yourself must less the board of directors, etc. Someone who works for themselves or for a real person
you or someone you know knows.

I'm not anti-global. Not by a longshot. I believe that globalism is an economic opportunity for each and every one of us that has literally been unheard of in the history of man. What I am against is transnational. The making of something in a remote place and then the assembling of it in another, and then the pandering of it in another. It is the very nature of buy low and sell high that is spoiling the global opportunity for everyone but big business.

But I digress here. Advertising. I've long been a fan of promotional advertising and feel it is one of the only types that does not offend the senses. A single thirty second promotional ad before and after a entertainment segment is non-offensive and gets its point across while leaving the consumer with a vague good feeling about the product or company. Now you put a ad segment in the middle of that same entertainment and all your going to do is piss a lot of people off. Advertising in its right place, and no where else.

These are just thoughts from inside an economist's head. Nothing more, nothing less.

I'm with Josh Sacks. The point of advertising on google search is to show up on the first page for specific products and services that you wouldn't rank for organically. Ebay is not at all a typical advertiser for the reasons Josh stated.

Also want to echo a Michael's comment with a personal example: One time I was watching Hulu and I got the option to answer a short quiz and they would direct the advertising revenue from my session to a charity of my choice. The quiz was a handful of questions about a car, and I couldn't answer one of them. I couldn't remember the make, the class, anything, despite the fact that I had presumably just watched two or three commercials about it. I thought geez, that commercial had zero effect on me. How can tv continue when advertisers can measure the ineffectiveness of their ads?

this was a paid hit piece from ebay, to discredit adwords and thus open the door to ebays new ad network offering. Not sure why this is being passed on as scholarly work...

If true, Google's slogan should be "a sucker is born every minute". But I don't think it's true. And I have no idea why Tyler long suspected that they don't.

In two decades wasting time on the Internet, I've never clicked through to purchase anything but that's just me. My impression from my peer group is people will occasionally click thru to Amazon for a very targeted product. I'm not sure how Facebook justifies its valuation.

Speaking of, I've often wondered how close "The Social Network" was to a true depiction of Zuckerburg's character. That "cool" aesthetic disappeared pretty quick once the IPO talk started.

Comments for this post are closed