Gmail is a product people use to exchange information. The product is free as far as I know. However, ads are shown in Gmail. When the user is using Gmail, they are signed in and they are ads invarious parts of the experience. These ads I assume are personalized when google has information about preferences and interests about the user, contextual based on the email content. The types of ads that are shown can be broken up into following varieties; untargeted, targeted based on demographics, based on interests, content based. The type of ad content is usually text assumption.
Ad revenue = price x volume, where volume is the average number of times an ad unit was sold.
Volume = number of times a Gmail experience was shown to a user x average rev earned
Number of times and content shown (let's calculate daily) = function of number of users x average ad content views per session
Number of users visiting Gmail every day = not sure of exact number but let us start from MAU, Gmail is dominant afaik except perhaps China. Let's say dominance is around 70% penetration for North America, Europe, India - North America roughly 500m population, let us assume that 80 percent are eligible adults, let's assume Europe is 250m people, again assume 80 percent eligible, in India assume 400 m people have access and 60 percent are eligible.
So, users in developed countries around 600m with Gmail accounts, India and rest of the world - another 600m.
Consider that habits are same, people check email daily - almost all of the monthly users. Let's say 80 percent uses check daily.
So roughly 480m users or round to 500m check daily from each bucket
Consider, mobile is dominant say 80 percent of visits and 20 percent on desktop on developed countries.
Consider 5 visits a day per user, with 3 ad content experience on mobile, 1 on desktop.
So if ads are sold based on cost per impression, then there are roughly
500x5x3,=7.5 billion impressions a day
1.5 billion impressions a day on desktop
Let us assume, that we have 50 cents per 1000 impressions = 3.75 b / 1000 = 3.75 m dollars per day, assume higher rate for desktop but less audience so 20 percent audience but 30 percent higher rate. So, 3.75m + 1m = 5m dollars per day.
So, 150m dollars / month, so 150x12= 1.8 billion dollars a year. Assume ad rates are lower in not developed countries, say 50 percent, then 0.9 billion or roughly 3 billion dollars.
We have possibly considered a much lower baseline rate appropriate for non targeted ads, higher quality targeting mix will push this higher 2-5 times. So revenue could be that much higher.
Other factors, users opt out in Europe and audience could be smaller.