AWS is a great example of the opposite, started around ~2006, it was originally deemed an "internal project" so they did not break it out. It was also really really tiny compared to retail revenue (amazon.com). They didn't actually start to break it out until 2015 and that is after wall street was somewhat demanding for it as it was well known it had experienced exponential growth and was generating billions for amazon. Were they trying to had bad numbers? Negative, they were trying to hide how awesome it was doing because it gave them the ability to make further gains w/ first mover advantage before competitors could react.https://www.channelfutures.com/cloud/amazon-com-breaks-out-a...
manquer|4 months ago
By 2015 they were trying to hide how much of the group growth and profits were largely contributed by just AWS , i.e. they were hiding the e-commerce margins .
Without AWS and subscriptions, Amazon is quite an unprofitable company.
Both overall growth and margin driven by AWS(and prime) while E-commerce revenue remains outsized because they count GMV as revenue which is iffy even when they own the merchandise, but being largely a marketplace these days GMV is very misleading metric.
It would be like Stripe decided to count their revenue as $1.4T the amount they processed this year as revenue rather than $10-20B they actually got after paying the banks, merchants , VISA etc . This 20B is not profit either just the pie from which salaries cloud costs etc have to be paid to get to actual profit.
JCM9|4 months ago
bcrosby95|4 months ago
JCM9|4 months ago
The inverse true now with AWS. Lots of press about analysis on how AWS is in “last place” on AI and while AWS leadership has been doing a lot of hand waving to say they’re not, it’s a pretty safe bet this week’s earnings call won’t have any hard financial numbers to counter press that they’re way behind.
JustExAWS|4 months ago
In my experience - and I’ve run comparisons against the various models for projects (consulting) - their in house Nova models usually give me the best results on the spectrum of speed/quality/cost I need for a given project.