Ethical Analytics

Alex Green

 

The challenge

To monitor and evaluate services, charities collect data through various commercial web analytics services like Google Analytics. However these services pose ethical and privacy risks: even just accessing a service may disclose sensitive information about an individual which can then become part of a commercially traded dataset. Furthermore, commercial web analytics aren’t designed to monitor the kinds of services charities run. They can be too complex for smaller organisations to effectively administer and implementing them often requires the creative misuse of features originally intended for tracking sales funnels. 

The opportunity

Charities have an ethical imperative to move to a more privacy-focused analytics tool. They also have a need to demonstrate that their digital services deliver value and collect useful data to enable them to improve these services. There is a clear opportunity to develop a charity-oriented web analytics service, which better respects user privacy and provides meaningful insight ‘out of the box’. 

In phase 1 this might include:

    • Volume metrics: e.g. page views, total page views
    • Engagement metrics: e.g. bounce rate, dwell time, scroll tracking
    • Acquisition data: e.g. top referrers, traffic source (organic search, social, paid)
    • Social metrics?
    • Built in benchmarking against similar services (using a tag categorisation system) – this is a USP
    • Simple display via a 1 page dashboard
    • Ability to define period for data display
    • Ability to define groups of pages for data display
    • Download of data to csv
    • Download of dashboard for print

In the future this might include:

    • Pop up surveys which allow users to actively opt in to more detailed service research for a limited time period (for example by providing demographic data or allowing full user journeys to be tracked). This shifts individual tracking to ‘by exception’ not ‘by default’.
    • Data API (in and out) allows connection to other services (e.g. social media post scheduling) or datastores. This could address some tracking issues as you can correlate effect rather than tracking individuals (e.g. correlate traffic rises after social post)
    • Journey / user flow mapping (would need opt in?)
    • Aggregate dashboard for tag categories (what impact have all the digital mental health services using the platform had? – needs enough users to be valuable)
    • Automated actions (e.g. email out dashboard)
    • CMS plugins