Website User Behavior Modeling - current visitor data gathering

[ExperimentsInHonesty]

Dependencies

• Google Analytics #1730
• Create Organizational Dashboard page #542: Organizational Dashboard
• Setup Google Analytics on website #503: Google Analytics
• After dependency is cleared, read through whole issue and either rewrite or make a new issue that has the relevant information.

Overview

We need to model the behavior of people on the current hfla website.

Action Items

• Review Modeling Questions.

• Add any others behaviors to track and their purpose in the comments.

• Review the documentation on what is possible and then outline the value of each platform to our need and the coding burden for each.

• Google Analytics

• FullStory free version

• Set up Hotjar analytics tool on website #432.

Modeling Questions

Purpose: To determine information value ranking.
What % of visitors click a top menu item?
Of that %, what is the breakdown of which menu item they click?

Purpose: to determine if the project's links are of interest to visitors.
For visitors who go to the project section (via click or scrolling), what percent click on the project's links?
Of that %, how many people click on more than one?

Purpose: to determine if visitors are seeing that the projects are being worked on one of two locations and pre selecting project exploration based on that.
Does there seem to be a pattern of people clicking projects that are in one of the two locations?

Purpose: to determine how people are using the maps, if at all.
What is the interaction with the link to maps? (Specifically, do people click on the link?)
If yes, do they click on something else, how soon do they click on something else?
If yes, is it the RSVP via meetup?

[thekaveman]

IMHO any click/etc. tracking should have an explicit purpose or hypothesis. E.g. we could theoretically have click-tracking on every link on the site, but why? Are there specific questions we're hoping to answer by having more granular use data? If not, I suggest we hold off on this until those questions are formulated.

[ExperimentsInHonesty]

@thekaveman I agree. I have updated the issue.

[stu562]

I was able to take a look at fullstory, simple sign up and got it to work on my local machine, it only needs a script added to the head of the main html file.
It tracks all users session into the signed up account.

For privacy concerns, adding css selectors to exclusion list from FullStory user's account dashboard is possible.

[switchzer0]

I think a way to focus what we're optimizing around would be on generating more of the two critical inputs to H4LA output (skills, project ideas):

• converting more new people/skills to show up at events (they'll get manually matched at events)
• converting more new ideas or partner-orgs/departments into potential projects (again, they'll get manually matched after H4LA reviews them)

Only after the site is set up for the above do I think we get tactical about e.g., button rearrangement. The volume of responses from the 43 survey responses to date supports this approach (given that skills can only be matched to projects if projects of interest exist).

Response	Frequency	%
Volunteer opportunities	32	74%
Current project information	25	58%
When and where do we meet?	20	47%
Hack for LA leadership / governance structure	7	16%
How to start a new project	5	12%
Sponsors and sponsorship information	1	2%

(data edited/updated 7/2)

[gitklu]

Current progress and low-fidelity wireframes for website re-design:
Hack for LA Website Redesign

(Doc includes link to survey results of website feedback.)

[ExperimentsInHonesty]

See me for fullstory user/pass