Posted on
by
Less Is the New More
Adam Ware
@wheresitworking
hashtag: #somuchdata
  • SwellPath: Love child of a web analytics and search marketing company
  • Everyone’s dilemma:
    • An overarching problem collecting, organizing, reporting on, and analyzing data
    • Universal across all organizations, scaling large and small, even non-profit preschools, small business, and large corporations
  • How did we get here — What caused the growth of internet data?
    • Speed : real-time data
    • Scale: unprecedented processing power
    • Sensors: many new kinds of data
  • Growth
    • In 2002: 5 exobytes of data
    • In 2009: 281 Exabytes of data
      (1 exabyte = 1024 petabytes = 1M terabytes
  • Digital Marketing Data is a Mess. The problem is exacerbated by:
    • Most metrics not being aligned with business objectives
    • Disparate data sources: website, social, mobile, CRM, etc.
      • And it’s really hard/ impossible to track people across those experiences.
      • So the data looks like seperate people. A person on twitter isn’t tracked as the same person who was on your website yesterday looking for tech support.
  • How do we overcome it?
    • Ask: Do the Number Even Matter?
      • Reasons why they might not matter:
      • Aesthetics: Are you studying stuff that you wouldn’t do anyway, because it would look ugly. e.g. Do you need to do A/B testing on a flourescent orange buy button?
      • Brand values: Are you testing things that, if you took action on, would be outside your brand value.
      • Overarching business strategy
    • Look at the Few (whitepaper from a company called Juicy Analytics, section called choosing the right metric)
      • Metrics that are actionable
        • Can you do anything about them?
      • They can be commonly interpreted
        • Does everyone in the organization understand the numbers
      • The calculations are transparent and simple
        • Example: on a website, they needed to measure which visitors were engaged. They had a way of figuring that out, based on pages visited, videos watched, etc. They could explain up front how the numbers were calculated (transparent), and while not simplistic, they were simple enough to be understood.
      • The data is easily accessible and credible
  • Aggregate Metrics
    • Would it be less valuable in aggregate?
      • Example: If you are studying growth in community, it’s useful to look at twitter, facebook, and forums in aggregate, not individually.
    • Be broad, but be comprehensive
    • Helps avoid uncertainty, data-doubt.
  • Visualize It
  • Q and A
    • Q: Making sure the data you have is actionable.. Is that a chicken and egg problem?
      • We spend time with the data, isolate the things that are really actionable for the people who don’t have a lot of time to dive into the data.
      • But we are spending a lot of time looking at all that data figuring out what is actionable.
      • And, of course, you need to deep dive to figure out what to do.
  • Resources:
    • Marrisa Mayer at PSRC: innovation at google the physic of data
    • blog.kissmetrics.com/you-vs-the-data
    • juicenanalytics.com