From $0 to a couple of millions. It totally depends of the kind of site/content, the profile of users, etc...

ummmm thanks everyone says adsense is best paying but I am finding it really low earnings. I don't get it. 1 million usa traffic hardly makes anything.

Then apparently it's the type of traffic that very few advertisers want.

Without counting:

- ad blockers, incognito/privacy mode of some browsers (the privacy mode Firefox prevents ads to show for example),

- mobile traffic

- robots (is it 1 million impressions from human visitors?)

and as i said all depend of the content, and the reason people are visiting your site.

Adsense impressions I mean, USA mobile traffic.

Most of Your Traffic is Not Human [webmasterworld.com]

As others have already mentioned: it depends. And it depends on a great many different things. Including:

* you ask: earn for 1 million impressions a month, which implies CPM rather than CTR, CPA, what have you. The average CPM I've seen quoted over the past few years is ~2.00USD or 0.002USD (a fifth of a cent) per impression, which on a million impressions would be 2,000USD. The problem with an average, of course, is that half of the values calculated from are below the average.
Since 2015 whether the million number is ads served or ads viewed is critical. The first, now, is only important as a check on delivery, the second is the money number.
With the increase in mobile often higher in content ad blocks are being passed over before they render.

* you say: USA mobile traffic. Mobile ads often pay (30-50%) less than desktop ads.

* as keyplr already mentioned an often unrealised unaccounted for reason for ad revenue problems is the percentage of traffic that is bot rather than human. Generally the smaller the site the greater the impact.

* while ad retargeting means that niche is less important, i.e. ads shown are less about on page content than about visitor prior interest - and retargeting ads usually pay less, site/page topic still plays a role where again some niches attract higher paying ads some lower.

* site audience makes a difference.
Generally, the better educated an audience the less interaction with ads; however, they tend to favour ads for higher priced products and services.
A study done a decade or so ago (I'm not aware of a similar recent one, however human behaviour is slow to change) showed that ~8% of searchers accounted for ~80% of ad clicks. If a site draws it's audience from that 8% it will likely do better than one that draws from the other 92%.

* with the current amount of programmatic ad fraud (~10/sec) and the recent counter of ads.txt whether you have incorporated it correctly and had that confirmed by various media ad bots may be critical in determining which ads you receive if any.

* percentage of device traffic can make a difference. iOS (Apple) traffic is typically worth more than Android.

* etc.

@iamlost
I agree with what you said but I feel a few important clarifications are necessary.

The problem with an average, of course, is that half of the values calculated from are below the average.

No, what you describe is the median. The median is the value for which 50% of the distribution is above and below. While it is often true that the median and and the mean (average) are equal, there are many situation where this is not the case. This is what describes a skewed distributions. When discussing Ad revenue the distribution is definitely skewed. That is few publishers earn a high RPM and many earn low RPM. The distribution has a long-tail (don't get me started on long-tails!).

The implication of misconception he is as follows, for example that a few publishers are earning $20RPM and we are told that the average is $2RPM be I earn $1.75RPM and thus feel like a chump because most publishers are doing better than me. This because I assume that the average and median are the same but they are not. If one looks at the median it is a $1.50 so I am actually in 75th percentile and thus doing better then most.

I realize that this distinction does not really change the gist of your argument but I feel it is an important distinction to highlight because many people are often confused and mislead by these types nuances. My long-tail pet peeve is a perfect example of this.

while ad retargeting means that niche is less important,

I think what you are describing here is interest based ad targeting and not specifically retargeting. Retargeting is when a user has visited a site and then left and then that website uses ads in combination with cookies to serve ads to that user in hope of luring them back to website to complete a transaction. My understanding was that these types of ads can have a high payout given that the advertiser has a big incentive to regain the customer (I may be wrong). Interest based targeting is more general, users are simply targeted based on browsing history and other indicators. In this case earning are likely lower than contextual targeting. Retargeting tends to occur less frequently. Whereas interest based targeting can consume a large portion of all impressions. This is specially true if there is low demand for contextual ads on you site.

You can see how your site fares in AdSense by going to Performance Reports => Advanced Reports => Targeting Type. The "Personalized" category covers contextual ads and I assume retargeting. As I do not see any specific reporting for retargeting type ads.

I personally expect 0.2~0.3 per 1000 impression. So 1 million impressions nets $200~$300 for me. My niche is highly non commercial.

My goal for myself is to eventually traffic 1 billion impressions per year. And I am pretty far from my goal! I spend less time in trying to "optimize" or "remove ads" or hunting down bots, but more time in content creation. So my figures can be highly distorted in terms of bots traffic.

@NickMNS: and I tried so hard not to be as pedantic as I often am :) as you mention mean, median, mode each usually return a different value; I went for simplicity of point as I don't remember, if it was actually stated, which was used to determine the $2.00 number. You are also quite correct that the earnings graph is probably quite skewed. Again, I simplified too greatly. As you say it is an important distinction and the WebmasterWorld audience is largely more receptive to and understanding of such detail. Mea culpa.

If it's follow on ad delivery, however general or specific, it's still retargeting and given that the average (median!) retargeting display ad has a CTR of less than 0.5% (5 per thousand) in both NA and EU it's far more irritating (over half who see same retargeting ad two or more times make a conscious decision NOT to buy the advertised product/service) than converting. Plus attribution is a killer. IMO retargeting is the ad networks/agencies current 'impressions served' smoke and mirrors.

---------------------------

Webdev fora are full of advice (good, bad, ugly) on creating content for SE benefit. What is largely missing is advice on creating content to attract the best most profitable ads available in one's niche. Very few webdevs seem to have even the vaguest clue about the ad spend in their and complimentary niches or might be the biggest spenders and what they like to splurge on. So much effort into optimising for search traffic, so little, mostly none, into optimising the revenue side.