Much lower visits counted as with clicky?

floripaoliver · July 7, 2015, 1:27pm

Hi piwik community,

I am using since a couple of month now Piwik and all seems working fine. I made of course sure that the piwik code is in all my pages.
How ever I keep getting much lower visit counts as with clicky parallel counting. I try to find out what is the reason for this but I cant find anything.

Of course I checked already a couple of times that all pages which include clicky also include piwik but still the results are complete wrong. Now of course I dont know if piwik counts to less or clicky to much. We are not talking about a small site with a couple of hundred visits. I do have in clicky about 50.000 visits a day but in piwik only about 40.000 so there is missing a huge part.

I would be happy if somebody can give me a tip on how I find what might be missing or why my piwik results are so much lower.

thank you all in advance,

matthieu · July 7, 2015, 3:46pm

Hi there,

If you compare Actions / Pageview do you find a difference between Piwik and Clicky?

floripaoliver · July 7, 2015, 4:28pm

Hi, thx for your reply…

Lets see June 2015 for example.
Piwik Visits: 1.16 Million / PIs: 2.34 Million
Clicky Visits: 1,46 Million / Pis: 2,31 Million

With other words… pretty big difference in Visit counts but actually almost equal Pis
Which means… now I am even more confused…

matthieu · July 7, 2015, 8:09pm

What would be really great is if you could compare with Google Analytics - any chance you could setup GA for a few days just to test the numbers? Ideally Piwik would look like GA if not, then this could be an issue for us at Piwik. Thanks for your help!

floripaoliver · July 7, 2015, 11:19pm

Hi Matt,

LOL, I installed Piwik to get rid of google because I am one of those who think that google has way to much influence and I rather don’t depend only on one firm

But I do have a analytics account and I can add for a couple of days both but I need a moment to do this. I guess I add the code tomorrow, than test it clean for 3 days complete and we compare the results. I keep you posted…

matthieu · July 8, 2015, 9:18am

Thanks! it will be useful.

I usually never ask here anyone to use GA, it’s very exceptional
We try to implement similar algorithm to google analytics in how we track visits.

floripaoliver · July 8, 2015, 1:40pm

Hi Matt,

Thanks for your support. I did already go through my entire portal and made sure that all 3 codes are in the same pages now.
In detail I did the following.

Scanned all sources for analytics code, piwik code and clicky code. Once I got all pages I deleted all codes first.
Scanned again and came up clean (no more codes)
Added again in all relevant pages the code of all 3.
All codes are loaded at the bottom before the closing tags
Always load first Piwik, than Clicky and last Analytics

So we should get clean results to compare starting from midnight tonight. I guess we wait 2-3 days and compare to get some data. In 3 days I should have something like 100.000 unique visits which gives us some date already.

If you contact me (I guess you can see in my profile) via email I can also give you access to my analytics and clicky accounts as well as my Piwik of course.

Thanks for your help

floripaoliver · July 14, 2015, 12:29am

Hi Matt,

Here are the results and they are pretty strange. I did compare now for the last 4 complete days the stats from GA, Piwik and Clicky. And as I wrote before in my topic I guarantee that all three codes are installed side by side in all my pages without a doubt. How ever the differences are huge. According to GA I have approx. 20% more traffic as Piwik stats show me.

How ever also Clicky shows me about 10% less than GA. Now since we deal on my site not with only a couple of users but approx. 200.000 visits in this 4 days this is pretty significant difference in stats. Now the question is if we believe more GA or more Clicky or Piwik

Check out the results. Send me a private message for more details and as well I can give you full access to all stats without a problem (or to my site). For the example calculation I set GA stats as 100% (which does not mean it is the correct one)

Thanks for your support in advance…

matthieu · July 14, 2015, 7:48am

Hi there,
Thanks for the numbers, the difference is quite big!

Could you prepare the same table, for “pageviews” as well as “visits” ?

nomeaning25 · July 14, 2015, 8:00am

Hello

Besides Piwik we use 2 other tools (Google Analytics and IBM Coremetrics) and we are seeing huge gaps between the data.

Analisyis from last 30 days (from 13.6. til 12.7.):

Piwik tracked for around 65 000 users
Google and Coremetrics tracked around 95 000 users.
Piwik tracked around 290 000 pageviews
Google Analytics tracked around 340 000 pageviews
Coremetrics tracked aroudn 375 000 pagewies.

More info in: forum.piwik.org

Good day

floripaoliver · July 14, 2015, 3:21pm

Hi Matt,

Okay I did the sheet for impressions now as well and the results are in deed interesting.
The results between clicky and google Analytics remain about 7-8% of the google results as well for impressions as for visitors which means that in the end clicky is about 7-8% of with their tracking code for whatever reason. Maybe bug delay for their tracking etc.

Now the results from analytics in compare with PIWIK are interesting because it looks like the impressions are actually correct (+/-1%) during the visits are +20% off. Check out the table below.

PS: I use latest version of Piwik.

quba · July 14, 2015, 3:43pm

Did you disable DNT support in Privacy tab?

Please also see: Configure Privacy Settings in Piwik - Analytics Platform - Matomo

quba · July 15, 2015, 10:40am

Please also keep in mind that Piwik excludes many visits because visitors are rocognized as bots, etc.

floripaoliver · July 15, 2015, 12:17pm

Well, I did not disable DNT yet but I can do for a test, how ever I highly doubt that this affects 20% of the visitors since only a tiny portion of people will use this at all and as well this is not going to make a difference because if DNT is involved the Page Impressions would be discounted as well which are not. Either way I give it a shot.

Now the information that Piwik excludes google bot visits etc. is complete new to me and could explain a lot of traffic I guess. Otherwise These bots done come with thousands of IPs (which are missing) .

quba · July 15, 2015, 12:21pm

For reference here is the list of excluded bots:

github.com

matomo-org/device-detector/blob/master/regexes/bots.yml

###############
# Device Detector - The Universal Device Detection library for parsing User Agents
#
# @link http://piwik.org
# @license http://www.gnu.org/licenses/lgpl.html LGPL v3 or later
###############

- regex: '360Spider(-Image|-Video)?'
  name: '360Spider'
  category: 'Search bot'
  url: 'http://www.so.com/help/help_3_2.html'
  producer:
    name: 'Online Media Group, Inc.'
    url: ''

- regex: 'Aboundex'
  name: 'Aboundexbot'
  category: 'Search bot'
  url: 'http://www.aboundex.com/crawler/'
  producer:

This file has been truncated. show original

It’s based on useragent.

matthieu · July 15, 2015, 2:44pm

Hi guys,

whether Bots are tracked or not could indeed explain some of the differences. Piwik by default works hard to exclude all bots of all kinds (including referrer spam bots)

Can you please try the following:

In Google Analytics, exclude all “Bots” from being tracked:

Wait 3-4 days and compare again Piwik VS Google Analytics stats for these days → is the difference getting less?

floripaoliver · July 15, 2015, 6:02pm

Hi Matt,

of course I can test this in analytics, I check it out and we see again in some days…
BUT if Piwik exclude traffic from bots, should this not include impressions as well? I mean should this exclude traffic than it would be for visits AND for page impressions because it would not make sense to exclude the traffic from visits but count the pis?

either way, I will test this… for us

floripaoliver · July 16, 2015, 3:23pm

I switched now all traffic from bots and spiders out. lets see what happen over the next 2-3 days.

ZipperZapper · August 16, 2015, 7:54pm

I think I’m allowed to join in this conversation? I’ve been using Google Analytics for the past seven years now and just this week started to roll out Piwik. Partly because of the idiotic cookie-rules in the Netherlands, it’s easier with Piwik to draw up privacy statements and keep better control over privacy. But, to the point:

After a week tracking with both GA and Piwik, I’m not really happy with the result. I expected little difference between the both, but not what I’ve seen this week. Some points:

Real time visit counter: Piwik always shows around 10 people less (most off the time, exactly 10. Weird). When it’s really busy, this difference can run up to 20-25 visitors less. Sometimes even > 30.
Visitors: Between august 10 and august 15, GA tracked 23.032 visits (or, as they call it: sessions). Piwik counted 18.210.
Pageviews are also off: GA counted 40.297, Piwik only 37.061.

As said I expected difference between the two, but I personally find these differences to big. Or am I to precise?

Some further information: tracking bots is switched off in Google Analytics. I disabled DNF-support in Piwik today, I will post new results in a few days to see if this changes things. But I can’t image DNF is used by several thousand people, to be honest. Both Javascript-snippets are placed in the footer. Switching them around has no effect.

quba · August 16, 2015, 8:14pm

There’s also this issue that will be fixed in 2.15: Chrome’s “Data Saver” mode prevents tracking of many visitors using Chrome browser · Issue #7733 · matomo-org/piwik · GitHub