Different numbers using self-hosted and cloudhosted

I’m working on www.euralarm.org. This org uses the self hosted version for many years. Now they want to change to the cloud hosted version. We tried the 30 days free trial version and that seems to work fine, except for one thing. The numers in the PiWik hosted/cloud reports differs from eachother in a huge way.

Obviously we checked the forum and took Troubleshooting - Analytics Platform - Matomo as a checklist. I didn’t find anything to adjust or change. We also contacted the hosting company about the SSL settings, without any results.

What might cause this issue?

I’m assuming the numbers are different for current visits, or did you upload historical data and find the differences are also there for converted data?

No, we launched a new website on January 27 (new CMS, new hosting provider, new IP’s and we moved 11 GB existing content). Obviously we implemented the PiWik software and the code. In parallel we tried the 30-days cloud hosted free trial version. During the same 30 days, the reports in the cloud based version are much better (more visitors, more clicks, less bounce etc.).

I had a similar issue, and I chocked it up to the archive reports weren’t necessarily being run accurately (I ran counts on some archive tables by site ID to compare the number of rows). So for the new install, I deleted all of the archive tables and rebuilt them all from the raw tracking data.

Since you are running in parallel, this may not be your issue. On your self-hosted instance, you could always rebuild for a single site ID for those 30 days and see if that gets you higher visits, etc.

If nothing else, you might want to count some of the rows in the various data tables and analyze what might be missing. Best obviously to look at a small subset (single site, single day, ideally down to a few minutes, and if possible identify your test visits for comparison).