Proposal: The ability to implement self-hosted aggregate analytics for a fixed time period


This proposal is to add the ability to implement self-hosted analytics in Hypha-core, for a fixed period of time.

Without configuration the analytics would not function.

NB: I want to stress - the analytics I am proposing are aggregate anonymised usage. I am not proposing, and would not propose, individual user analytics.

Time-limited implementation

The recommended implementation for these analytics would be for a fixed period of time, in order to minimise the data collected. This time period could be for example; a week before the CfP period opens until the closing date, or a fixed period of weeks/months.

Why propose this?

The objective of these analytics is to improve Hypha. They would be used to identify aggregate usage patterns by fund applicants, and fund staff users. By staff users I mean - anyone: fund management, finance users, external reviewer users.

The knowledge gained from these aggregate usage patterns would be combined with qualitative feedback (interviews, bug reports) from users - both fund applicants and management staff - to identify areas of Hypha that are underused, causing issues, need improvement, or possibly can be removed.

What data would be gathered?

The analytics would cover 2 areas.

1. Fund applicant analytics

What information would be useful to gather?

It would be useful to be able to answer the following questions:

  • what are the different browser languages applicants use? (to determine if there are localisations needed)
  • aggregate numbers to answer the following questions:
    • do applicants use the guidebook links in applications, where present?
    • what is the total time average time an applicant spends on an application page? (useful to identify if there are some pain-point questions that need improvement)
    • what are the most common paths applicants take through the site?
    • where do applicants drop off in the application process?

2. Fund management staff analytics

What information would be useful to gather?

It would be useful to gather aggregate numbers to answer:

  • what are the most common paths staff-users take through Hypha? (this is useful to understand what areas of Hypha are not used, under-used, cause issues)

Analytic software options

There are different options available. I expect self-hosted would be most acceptable. In this case the most common would be Matomo (formerly Pikwik).

Privacy and transparency

There would, of course, be a need to publicly state what analytics are implemented on the site.

In order to build trust between the organisation and applicants, staff, all/some of the aggregate analytics could be made public to make show what is gathered.

As mentioned above, the intent of this proposal is for improvement to Hypha. I’d like to hear what people’s thoughts are.