Being a data driven organization

From DevSummit
Revision as of 17:47, 5 May 2015 by Vivian (talk | contribs) (1 revision imported)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to navigation Jump to search

Being a data driven organization

David Taylor, notes by Josh Levinger (Citizen Engagement Lab)

what can we learn from obama campaign?

  • they learned it all from silicon valley

what do we mean by data?

  • raw chunks, metrics, informed decision making
  • dashboards

how do we act on data we collect?

  • only produce analytics that are actionable (vs reportable?)
  • vanity metrics are killing startups (clicks, hits, users)
  • actionable ones key (cost per acquisition, revenue per user, dropoff rate)
  • very similar to member-funded orgs (dollars raised per user, value provided to them)
  • * lead generation -> conversion -> lifetime value
  • pirate metrics talk

what is most central metric? just actions, dollars?

  • twitter used avg engagement per user per day in their early days)
  • campaigns could use virality, or growth velocity (recruitment)
  • ability of people to spread your message (shares)

knowing what your goal is as organization

  • what are we trying to do in the world?
  • at the end of the day, what do we care about?

most of the things that are measurable are best for lead conversion

  • 90% of what we can measure are vanity metrics
  • define qualitative metrics / proxies for real world engagement (clicks vs higher bar asks)
  • move people up the engagement ladder

fundraising culture is far behind software development in following pace of change

  • rigidity (obliviousness) vs plasticity
  • how do we create pressure on foundations to change?
  • some orgs stuck in loop of email list growth & rapid response
    • key performance indicators should be balanced against each other
  • data collection and decision making systems don't talk to each other
    • by data we mean constituents (and their interactions with the org)

How do we make this easy?

  • you need a lot of data before being metrics driven makes sense
  • collect everything you can, go back and design what's relevant later
  • use automated tools that are built for this (kissmetrics, mixpanel)
    • google analytics gives you traffic numbers (vanity metric)
    • cohort tracking, not just spikes
  • decide on one kpi per channel (+- on baseline, something that indicates health)
    • know your baseline, and the sector baseline
  • start with goals, work backwards
  • * but check that it fits your organization model

visualization helps connect between different staff cultures

  • intelligence dashboards
  • chart.io (visually build charts from many data sources)
  • make everything transparent, everyone in the organization has access to every graph

design tests, knowing most of them will fail

  • but with limited resources "burning" them can be scary
  • phrase a hypothesis vs "fine tuning" or optimization
  • when is there an end any more?