My ad network, local pageview counter, Google Analytics and Cloudflare Web Analytics all report pretty different numbers. I use the count to describe traffic to advertisers, so I don't want to overstate it. Which one is most reliable?
Which site analytics numbers should I trust?
Asked by
5d!mapdrawer_273
Answers
There isn't a single most honest number until you decide what you're counting. A page request, a browser pageview event and an ad impression are different things. Even two tools labeled pageviews may count them differently.
Before comparing providers, check that the reports cover the same dates, time zone, pages and event definition. Then compare trends as well as totals. A lower count isn't automatically more accurate, and a higher one isn't automatically wrong.
Reply to this answer
Your own server count gives you a useful record of requests that reach it, but it isn't automatically a count of real people viewing pages. Bots, repeat requests and requests that never become a visible page can inflate it. A JavaScript check helps filter some automated traffic, not all of it.
I'd audit the local counter before using it in a sales pitch. Check when it increments, whether reloads count, and what filtering is in place. Keep a clear description of the method alongside the number.
Reply to this answer
Browser-side analytics can miss visits when scripts or requests are blocked, fail to load, or don't run. That alone can make a big difference compared with a server-side count.
Changing the collection endpoint might affect what gets blocked, but it won't settle which visits are genuine. Also check your privacy and consent obligations before changing how you collect data.
Reply to this answer
For a flat-rate ad sale, I'd be transparent rather than picking whichever total looks best. Share the traffic figure you can explain, name the tool and date range, and say whether it's pageviews or impressions. If you have a separate count for the actual ad placement, include that too.
Reply to this answer
Google Analytics can still be useful for looking at behavior and trends, even if its total misses some traffic. Matomo is another option worth evaluating, but switching products won't by itself prove that the new number is the right one.
Compare how each collects events, handles filtering and reports pageviews before making a move. A dashboard that displays another tool's data more neatly won't change the underlying count.
Reply to this answer
Add your answer
Answers are reviewed before publication.
5d!mapdrawer_273:
For advertisers, would it make more sense to report ad impressions instead of pageviews?
Varlix+4n:
If you can measure impressions for those placements, that's more directly relevant. I'd still label the metric clearly and avoid treating every pageview as an ad view.