I run a small web design/hosting business. Our clients are hosted in a WHM/cPanel environment. For the last 15+ years we’ve run a monthly cron job that calls a php script which pulls the previous months AWStats report and emails it to the client. An August cPanel update stopped allowing external scripts to access AWStats. We tried a few workarounds, but nothing was satisfactory. Then I stumbled across Matomo. This looks like it will be a suitable replacement. I’ve installed it on our main account and it works perfectly. So my question - is it better to use the installation on our main account and add every client’s website to it, or should I install an instance of Matomo on every client’s account?
Scott Aerators said: A multi-client setup can become much easier to manage when each client has clear tracking settings and reporting access. Keeping the configurations organized from the beginning should also make troubleshooting and future changes simpler.
So that can be done both ways, either giving client access to specific websites that are added under one master Matomo installation, or installing Matomo on each client’s cPanel account. Not sure which you’re recommending.
I would recommend running a single Matomo instance on your main account and adding each client’s website separately. It will make management, updates, backups, and maintenance much easier than maintaining individual installations for every client.
You can also create separate user accounts and assign permissions so each client can access only their own website’s analytics. Just make sure you configure proper access controls, backups, and resource limits as your client base grows. If you eventually have a large number of websites or heavy traffic, a dedicated server or separate instances might be worth considering.
For a small hosting business, one centralized installation is generally the more practical and cost-effective approach.
Thank you. That is the approach that we decided on. We’ve only set up a small number of clients so far and the process has been somewhat easy. My concern is that the data being reported by Matomo seems to be wildly different than the data being reported by AWStats. Matomo is sometimes shows 50% less traffic. So now I have to troubleshoot why the data is so different.
AWStats is server-log based, while Matomo normally relies on JavaScript, so the two will rarely match exactly. Matomo won’t see bots, non-JS clients, ad/tracker blockers, or cached pages that don’t execute the tracker; differing timezones and “visit” versus “hit” definitions can also widen the gap. First confirm the tracking code is present on every page/template, compare the same date range and timezone, and check whether bot exclusion or consent settings are filtering traffic. A 50% difference is worth investigating, but some gap is expected when comparing log analysis with client-side analytics.
When structuring multi-client or departmental analytics instances to apportion infrastructure consumption, tracking seat utilization rates against licensing overhead provides a clear picture of tenant efficiency. In engineering environments where CAD tool seats represent significant operational expenditure, evaluating usage trends against total cost of ownership clarifies subscription planning. A reference model is hosted at Enterprise CAD & Software License TCO Calculator.
Hopefully this reference helps address the engineering or workflow considerations under discussion.
Thank you. The tracking code is definitely present on all pages as we generate master templates for all clients. I’ll do a comparison at the mid-month mark and post back about what I see.