Hacker News Reader

Comment by MyFirstSass | Hacker News Reader

Anyone have a good guesstimate of how much money is "lost" when Youtube/Facebook/Instagram/Whatsapp (more?) are down each minute?

View on HN

I can’t answer your question, but when I was at Google I made a mistake that caused ads serving on Google results to become unclickable. For the postmortem they had me calculate (I don’t think a dollar amount but) the number of ad clicks that would have happened during the time it was down. Of course I looked up average cost per click rates. Not sure if I could share even if I remembered, but it really put things in perspective.

Overall it was a good learning experience. I didn’t get reprimanded; several months later I got a promotion.

I can't believe you typed all that and didn't include a number, what a tease

Something like $X million an hour.

This is rough napkin math, no need to downvote if anyone knows the real number and this is way off :)

Meta 2023 ad revenue was $131 billion. To make it easy, let's assume an even spread for # of users and ad revenue generation per hour/minute of the day and day of the year (which I'm sure is not the case).

This would be:

$358 million per day

$15 million per hour

$249k per minute

This also assume a minute down won't be somewhat or totally offset by a spike in users when it comes back online.

I didn't check your math, but your last number is probably per minute, not per hour.

Thanks, you're right . Edited it

I assume they will run all those campaign views after yhe outage, so very little will be actually lost

A hell of a lot less than when they are up.

Naively, divide ad revenue by time to get a dollars-per-time.

But thats naive because ad serving isn’t totally sold out so they can make up for it by increasing the density of ads in the next time window. If the outage is short, then the impact is small.

But some markets are totally sold out and there’s no making up for lost impressions.

This is something I've thought about a while back. Like Facebook probably has a "maximum number of ads shown to users per post" value. So theoretically, they have a ceiling for how many ads can be bought in a specific time frame before having to increase the ceiling/find new users.

Do they share data about this?

> before having to increase the ceiling/find new users.

Or raise the prices of ads.

> they can make up for it by increasing the density of ads in the next time window

Not only that but the bigger spenders will have more budget so the bidding after a large outage should return higher bids on average leading to increased profit per ad slot.

Possibly, but increasing ad density is usually negative for performance. So it’ll probably be end of bad for Meta as a whole, as people spend more but don’t get more value out.

> they can make up for it by increasing the density of ads in the next time window

no need to do that, for google search, people will come back later to make the search

Yes, but, they’ll need to serve X ads in a smaller amount of time

They also do not refund spend during this time typically so wouldn't be a 1:1

They don't? I've worked in ad tech at some smaller places and we absolutely refunded spend during outages.

From my experience you'll receive a partial refund - and in some instances like inexplicable overspending, etc. - you won't receive anything. This may be an exception given a full sitewide outage, though

How is money even being spent if users aren't viewing the ads?

Not all ad spend is impression based, there are things like takeovers. And, sometimes impression counts arent guaranteed, theyre estimated

That's because they were some smaller places.

When you've more-or-less monopolized a lot of the web's content sharing you get to tell your clients to pound sand. Where else they gonna go? Twitter? The incel white supremacist dollar is not what advertisers call "the good dollar".

Reply on news.ycombinator.com