How much of your website traffic is bots?
Problem
- About half of all web traffic is not people - crawlers, scrapers, monitors and bad bots.
- A raw count that does not filter them overstates your audience, and every plan built on it - ads, sponsors, the salary sum - is built on air.
Solution
- Take your own bot share off your WebCounter report, which filters bots before it counts and shows them as "bots filtered".
- Put your raw count into the free calculator below and it shows the real people and what they are actually worth.
Key points
- Imperva measured 49.6% of internet traffic as bots in 2023 - 32% of it from "bad" bots.
- At 50% bots, a raw 10,000 visits is 5,000 people - the count is 100% too high.
- Those people earn about R820 a month at $10 per 1,000 visits, not the R1,650 the raw count suggests.
- At 70% bots the raw count is more than three times the real audience - 3,000 people out of 10,000.
- Ad networks do not pay for bot traffic, so only the people ever earn.
Every post on this blog so far - from visitors for $1,000 a month to traffic to replace a salary - starts from your visits. This one asks how many of those visits are real people, because a plan built on the wrong number is wrong from the first line.
Why a raw count is too high
A raw count - server logs, your hosting control panel, a counter with no filter - adds one for every request. A large share of requests are automated, so real people = raw count x (1 - bot share).
- Bots are about half the web. Imperva's 2024 Bad Bot Report measured 49.6% of all internet traffic as bots in 2023.
- Not all bots are bad - search engine crawlers are how Google finds you - but none of them reads your posts or clicks an ad.
- Small sites are hit hardest. The bots come whatever your audience, so a new site with few readers can see more bots than people.
- Ad networks screen out invalid traffic and do not pay for it, so the gap between the count and the cheque is the bots.
Work out your real visitors
Type the visits your raw count shows for a month, the share that is bots, what 1,000 real visits earn you and your traffic growth. The calculator shows the real people a month and a day, how far the count overstates them and what they are actually worth - with a chart of the next year.
Prefer a full page? Open the bot traffic calculator on CalcSA. Once you have the real number, the visitors for $1,000 a month calculator shows how far you are from a target, counted in people.
Worked example: the calculator's defaults
A raw count of 10,000 visits a month, half of them bots, earning $10 per 1,000 real visits, with traffic growing 3% a month. Dollar earnings are converted at the European Central Bank reference rate the calculator stores (14 October 2026: $1 = R16.47).
- 5,000 of the 10,000 are real people - about 164 a day.
- The raw count overstates them by 100% - planning on 10,000 assumes 5,000 visitors who are not there.
- The real people earn about R820 a month ($50), not the R1,650 ($100) the raw count suggests.
- In 12 months the raw count shows 14,258, and the real people are 7,129.
Three more cases, same defaults except the one shown:
- 30% bots: 7,000 people, the count 43% too high, about R1,150 a month.
- 70% bots: 3,000 people, the count 233% too high, about R490 a month.
- No bots at all: the raw count is right - which is only true of a count that filters them first.
Estimates. The 10,000 visits, 50% bot share, $10 per 1,000 visits and 3% growth are assumptions, as in the calculator; your own report shows your real bot share. The live calculator uses the latest ECB rate, so its rand numbers may differ slightly from these.
The raw count against real people
Same site over a year, growing 3% a month: the raw count, and the real people in it at 50% and 70% bots:
- Raw count
- Real people at 50% bots
- Real people at 70% bots
Show the numbers as a table
| Month | Raw count | Real people at 50% bots | Real people at 70% bots |
|---|---|---|---|
| Now | 10,000 | 5,000 | 3,000 |
| 3 | 10,927 | 5,464 | 3,278 |
| 6 | 11,941 | 5,970 | 3,582 |
| 9 | 13,048 | 6,524 | 3,914 |
| 12 | 14,258 | 7,129 | 4,277 |
- The gap grows with the count - at 50% bots it is 5,000 now and over 7,000 in a year.
- Growth in a raw count can be growth in bots - a new scraper looks exactly like new readers.
- Only the people line earns, so it is the only line to plan from.
Reading your counter for this
Your WebCounter counter screens every request before it counts: no browser, an implausible user agent or a known crawler, and it is filed under bots, not visitors. No counter yet? Get your free WebCounter counter - it takes a minute and costs nothing.
- Your bot share is "bots filtered" divided by unique visitors plus bots filtered, on your WebCounter report.
- Compare it with your hosting stats: the difference between their number and WebCounter's is mostly bots.
- A sudden jump with no new post behind it is worth a look at the bots column before you celebrate.
- The filter is honest about its limit: a real browser driven by a determined bot can pass, so treat the visitor count as close, not perfect.
Sources
- Imperva, "2024 Bad Bot Report" - 49.6% of all internet traffic was bots in 2023, 32% bad bots. Not re-checked at the original on 18 October 2026.
- Aditude, "RPM for Publishers" (23 July 2025) - $5 page RPM midpoint for travel, food and home; x 2 pages a visit = $10 per 1,000 visits. As cited in the CalcSA calculator.
- European Central Bank euro reference rates, 14 October 2026 (stored copy in the calculator) - $1 = R16.47.
- Calculator: CalcSA, Bot traffic: real visitors.
Estimates only. Your bot share depends on your site, your host and who is crawling this month; your own report beats any average.
Count your page views - free
WebCounter gives your site a free counter and a 30-day report of page views and real visitors, with bots filtered out. No cookies on your visitors.
Get your free counter