Hacker News Reader

Comment by gwern | Hacker News Reader

More detailed discussion of the 4 other datapoints: https://www.gwern.net/Ads#replication (I added McCoy et al 2007 just last night, so now it's 5 total: me, Pandora, Mozilla, LinkedIn, and McCoy et al 2007.)

View on HN

It also lines up pretty well with the fact that a lot of early stage SV-style startups don't monetize before they have a lot of users. You don't want to lose 10% in the early stages of exponential growth. 10% less users later on would not be a big deal however.

I'd assume that the 10% later on is a relatively larger dollar amount in revenue, no?

But you're not sacrificing growth then. When you sacrifice growth for revenue up front, you're sacrificing not just 1 user but that user and all subsequent users they would've brought in via downstream/compounding effects plus the associated economies of scale (which may well be the only thing that will make you eventually profitable, ads or no ads).

Assuming that advertising is going to be a primary source of revenue, than you aren't going to have much revenue to lose before you start advertising.

The idea is that companies don't advertise at first to take advantage of the exponential growth and user retention when they start getting hot, and once they have market share they start to target profit rather than growth.

It depends - if you're going from zero revenue, then, losing 10% of your users to go from zero dollars to N dollars is a great deal.

If you have existing other revenue sources, then you definitely have to factor that into the equation.

I am so glad that I’ve never worked for a non B2B company. It’s so much easier having 10-20 large companies as customers that pay real six and seven figure amounts for your services than chasing after a few pennies.

Some studies you may separately find interesting in connection to this,

- https://www.kellogg.northwestern.edu/faculty/gordon_b/files/...

- https://marketing.wharton.upenn.edu/wp-content/uploads/2017/...

- https://academic.oup.com/qje/article-abstract/130/4/1941/191...

First & third are already linked. Haven't seen #2 before.

It's often said that less than 10% of users click ads. If there would be way to identify those users with reasonably low false positive rate, you could annoy only small fraction of users and keep more of them (if you want to have them).

Yeah, that's the thing. I will never (except possibly accidentally) click on an ad. Every single time I see an add in my Instagram feed, I mark it as "not relevant". And yet these days I see ads every 5-7 photos in my feed. I use Instagram much less now because of the quantity of ads, and Instagram should clearly be able to tell that I don't interact with ads at all except to get rid of them. I've probably reduced my use of the app by 75% or so (both viewing and posting), solely because of ads.

What are you trying to accomplish by clicking "not relevant"?

My (naive) hope was that after a while they'd give up.

Is there a better response, you think, that would possibly decrease the number of ads I see?

For what is worth, this does work on Twitter (by straight blocking the account which posts the ad). I started seeing noticeably fewer ads after doing that.

The "not relevant" button certainly means "please, show me other kinds of ads".

I've tried "I see this too often" for a period of time without any better results.

And in the end, I don't really see much variety in the ads I see. Decent variety in account/company, but very little variety in product/market type.

> Yeah, that's the thing. I will never (except possibly accidentally) click on an ad.

I've noticed more and more dark patterns to trick users into accidentally clicking on an ad. Showing an "x" in the top corner before the ad is allowed to be closed/skipped, for example. I'm not sure whether it's game developers or advertising agencies who are responsible but it's definitely frustrating.

Not all ads need clicking, and the big budgets have never been in generating clicks, but in Branding (like TV ads).

The 10% that click are likely spit somewhere around 50-50 between the least valuable eyeballs (suckers who click ads) and people that actually are interested, depending upon the ad type (Search more valuable, banners less so).

Are we controlling for SEO effect? Having ads on your site probably nerfs your Google ranking a bit.

Wouldn't the incentives go the other way? I'd imagine Google wants people to get to sites that serve ads from Google so they can collect fees.

Ads don't affect SEO. It would allow websites to essentially pay for higher search results.

There are tools that come with using ads that allow for insights into better SEO though - e.g. Adwords keyword planner.

> would allow websites to essentially pay for higher search results. //

Isn't this exactly what Google does? Paid placements put you above organic results in the SERPs, presumably other search sites do that too.

Yes, and they mark them as ads so users know that they're promoted and not organic search results. Allowing websites to buy ads to affect their organic placing would affect the quality of search results.

> Yes, and they mark them as ads so users know that they're promoted and not organic search results.

And over the years they've added more ads on top of the organic results (instead of the sidebar) and made them less distinguishable.

Google's core business is letting people pay to be the top of search result.

For paid ads. SEO is not about search ads, but about organic search results (which is what I'm referencing).

My SEO knowledge is a few years out of date, but when I ran sites with AdSense, it hugely increased the page weight and load time. Google's own tools would ding me PageSpeed Insights score for some of the scripts, even though they were loaded async. GA would tell me I had 8+s average page load; when tested from slow connections, it would never be more than a couple sec.

Reply on news.ycombinator.com