No Visitor Log Or Meter Is 100% Accurate

Confusion about Stats, and referer spam inaccuracy, is expressed, weekly.
You never have a real idea as to how many real visitors view your blog, since these referrer spammers are in fact identified by their name, and their number of visits on the Stats "referring URL" list.
Referer spam inaccuracy is possibly the best known - but certainly not the only - Stats problem, discussed daily in Blogger Help Forum: Something Is Broken.

Stats is a controversial Blogger feature, because of various perceived inaccuracies - but even so, is more useful than competing products, because of its unique design.

Not every blog owner realises that all visitor logs / meters provide displays and statistics, which contain known inaccuracies. The fact is, no visitor log / meter can ever be 100% accurate.

Many blog owners see the tabular design of some visitor log / meter displays, and treat the various displays as different views in a balance sheet.

Some blog owners download details and statistics from the visitor log / meter of choice, and save the contents in logs or spreadsheets. Some like to reconcile details in one spreadsheet page, against summary totals in another. Others may use multiple visitor logs / meters.

Comparing aggregated details to totals - or comparing figures between different products, they will discover discrepancies.

  • Cache related deflated counts.
  • Daily count reset time skew.
  • Detail to Totals reconciliation.
  • Main page / post page reconciliation.
  • Owner activity inflated counts.
  • Page load impatience deflated counts.
  • Referer spam inflated counts.
  • Security filter deflated counts.
  • Social sharing inaccuracies.

Without understanding the origin of the various discrepancies, some blog owners will become enraged, and accuse the various log / meter suppliers of providing defective or inaccurate software.

Some of these discrepancies, unfortunately, are inherent in every visitor counter, log, or meter.

Cache deflated counts

Every visitor log or meter which counts visitors, or records visitor activity, based on access to a web server, will be inaccurate.

Some people will use browsers and computers which cache heavily. Others will use browsers or computers which are subject to upstream cache.

In either case, content which is cached - and when cache has not expired - will not require server access, and won't be counted, for spurious server access.

Daily count reset time skew

Every visitor log or meter which provides totals for any delimited time period (day, week, or month) resets the totals at some time during the day, for all blogs / websites. Since most countries subject their inhabitants to a twice yearly clock reset (aka "daylight savings time"), no inhabitant of the world will see the totals, in their visitor log or meter, reset consistently at their own local midnight.

Everybody must accept the reality of seeing the totals reset at some time other than midnight, at some time of the year. Some blog owners will see a skew of up to 24 hours, at some time of the year.

Different products will reset at different times of the day, because not all products reset at midnight GMT. Comparing daily totals, between services, will produce discrepancies.

Detail to Totals reconciliation

Stats provides breakdowns of pageview counts, by browser / computer operating system, by country, and by page / post - but limits the breakdown displays to the 10 most popular entries, in each list.

When trying to reconcile totals of (the 10 most popular) posts, against totals by browser or country, discrepancies are unavoidable. With any blog of any appreciable size (more than 10 posts), or reader population dispersion (more than 10 countries), any attempt to total up any breakdown will be a waste of time.

Main page / post page reconciliation

Many visitor activity logs provide a breakdown by individual post (page). Since many Blogger blogs publish post content in the index pages (archives, label indexes, and home / main page), pageview totals will never equal counts aggregated for all displayed (10 most popular) posts (pages).

Owner activity inflated counts

Every visitor log or meter will have problems with owners of blogs with low visitor activity, who may consider their own activity to generate inaccurate counts. There are workarounds for this inflation - but no workaround is 100% effective, persistent, or transparent.

Page load impatience deflated counts

Every visitor log or meter which uses installed code will contribute to page load delay - and be subject to impatient visitors, who may close a browser tab or window because they get tired of waiting for the complete page to load.

Referer spam inflated counts

Stats is subject to inflated counts, from people who use referer spam to attract visitors to their blogs and websites.

Security filter deflated counts

Every visitor log or meter which uses cookies or JavaScript, in the process of recording visitor activity, will lack some detail. Everybody who reads your blog will use a computer which is protected against malicious code from unknown websites.

Almost every visitor log or meter, other than Stats, will be subject to some security filter (blocking cookies and / or scripts) based deflation.

Social sharing inaccuracies.

FaceBook, Google+, and other social sharing services, have various activity counters which refer to sharing of content, within their services. Blogger displays the Google+ "+1" counter for the posts in your blog, to indicate sharing of Google+ stream posts, which refer to blog posts.

Not all +1 counts, which refer to a specific blog post, will reference actual views of the post. There are many reasons for the discrepancies.

The bottom line here is that there will never be any visitor log or meter which will be 100% accurate, when loaded into a spreadsheet, and rows and columns are tabulated, and totaled. Every blog and website owner has to understand that visitor logs and meters are subject to some inaccuracies, and simply use the log / meter on their blog to identify trends, and to proactively watch for problems.

Comments

Hi, until recently (last 3 weeks) the summary of PageViews has been somewhat close to the top ten posting pageviews for a given period. Suddenly our PageViews have dramatically increased, but I can't see this reflected in any specific blog postings - ideas? Comments? Thanks