Avinash Kaushik, whose blog Occam’s Razor is a must-read, had a lot to say at the Emetrics Summit this week. I tried to write it down for those not lucky enough to attend.
Avinash made seven recommendations:
- Go for the bottom line. If you find out how to increase revenue, cut costs, raise customer satisfaction, or whatever the execs at your company want, you’ll be listened to. “People love their paychecks,” Avinash advised, so appeal to what they care about to increase or safeguard their personal paychecks.
- Reporting is not analysis. Both reporting and analysis can consume all of your time, but analysis is far more valuable. Avinash cleverly called reporting the art of finding three errors in 1000 numbers and analysis the art of knowing that the three errors don’t matter.
- Depersonalize decision making. Avinash described the phenomena of the “Highest Paid Person’s Opinion” (HiPPO) making every decision—”HiPPOS rule the world,” as he succinctly put it. But he said that you can turn this culture around into a data-driven culture by being a slave to what customers are telling you with the data. He says it is the rare HiPPO who will look at your data about customers and say “Screw them.”
- Proactive insights rather than reactive. If you think about what a CEO thinks about and get in front of his thinking, you’ll be listened to. “20% of your time should be spent on analysis no one asked for” Avinash told us.
- Empower your analysts. “Senior analysts should spend 80% of their time on analysis and just 20% on reporting,” Avinash recommended. “Hire an intern for reporting,” he went on to say. He encourages risk-taking and critical thinking among his team.
- Solve for the Trinity. Avinash reminded us that click streams are only one of the critical measurements, what he called “the what”—customer behavior. We also need to research “the why” and track outcomes (“the how much”) of orders, leads or other conversions.
- Got Process? “Report scheduling and publishing is not a process,” warned Avinash. He called decision making a journey, not a destination. You must decide what roles are involved in your process (business user, analyst, designer, etc.) and hold each role to its part of the process. You must assign ownership for actions and track them so that the metrics drive the projects which then drive the collection of new metrics. That’s process.
Avinash implored this audience of metrics analysts to understand the limitations of what metrics can tell you. Pointing to fellow blogger Marshall Sponder in the front row, he exclaimed, “If Marshall comes back to my site again and again, is he engaged or frustrated?” The metrics don’t tell you. If no other example told you how important analysis is over reporting, that should do it.
