search

matomo-org / matomo

Empowering People Ethically with the leading open source alternative to Google Analytics that gives you full control over your data. Matomo lets you easily collect data from websites & apps and visualise this data and extract insights. Privacy is built-in. Liberating Web Analytics. Star us on Github? +1. And we love Pull Requests!
https://matomo.org/
GNU General Public License v3.0
19.95k stars 2.66k forks source link

Make automated Recommendations to users #6179

Open mattab opened 10 years ago

mattab commented 10 years ago

The goal of this issue is to implement a mechanism to let plugins create recommendations. A recommendation is an advice text given by Piwik about something actionable to improve, something to watch out for, or anything else that may be useful to tell the user.

Here is an example of this feature in a SEO product: automate-seo recommendations

These recommendations would really improve the value of Piwik by making it the smart analytics companion that any website owner needs. This could recommend ecommerce modifications (bringing more insights in #6177 ) and/or SEO recommendations (related to #1459) and/or better Insights in general (maybe Insights data can be used to issue smart recommendations? #5246)

wikiloops commented 10 years ago

@mattab, this issue here is closely related to a lot of other issues I found by going thru the list, including #6576, #5246 and #6507. Also, there are a group of issues concerned with "How to guide a first time user thru his first experience with PIWIK" & "How to make first installation process easy", including the major #3185, which are also aiming in a similar direction.

Maybe it makes sense to create a label concerned with "new user UX" to keep those things in sight - I belive the topic will not be prominent in the GitHub issues, simply because few new users will head over here right away to give feedback, while at the same time it is crucial to not forget about those first minutes in which a user will reach a first conclusion about "is PIWIK worth further investigation?".

concerning the recommendations feature: Maybe its a good start think about the different roles the people looking at a PIWIK report may have. The sales department will look for other stats than the UX developer, and a hobby blogger will have interest in other metrics than someone managing a high performance platform. One could try to either filter the automated recommendations according to these roles (no need to tell the CEO about increased average generation times or bounce rates, really), or implement a display that offers a hint on which "department" a recommendation is aimed at - some examples: "server monitoring advisor recommends: Check server workload, increase in generation time monitored!" "ecommerce advisor recommends: Check item XY, orders have dropped by X percent." "development advisor recommends: mobile iOS visitor bounce rate has increased by 20%, better investigate display & functionality." "SEO advisor recommends: Page XY gets 80% more hits from search engines than average, try to adopt pages SEO concept to other content if possible." "CEO advisor recommends: Raise the salary of your web team, your visitors have increased by 10% and you are making 15% more revenue than last month, good job!"

just brainstorming a little, with the non-expertised user in mind...