Startup Analytics

I’m going to be working with startups for two weeks advising them on how to use analytics to help fuel growth for their organization.

Taking notes here from materials I’m reading that I’ll eventually package up with my own thoughts and execution techniques through Google Analytics.

Metrics, you want to measure Outputs not Inputs. Draft a dashboard to tie this all together. You want to look for the anomalies, you don’t want to look for expected behavior.

At PayPal – none of the metrics said eBay. This was spotted through powersellers on ebay asking for PayPal – building a product for them to make it easy.

At LinkedIn – 30% of clicks from homepage went to a person’s own homepage. Made no sense. Max Lechin said it was vanity. People are looking at themselves in the mirror. Makes them feel good. You can test that, more content / more endorsement, you look at yourself more. It clarified what users of the product really wanted.

Keith Rabois – How to Operate – How to Start a Startup

From: Lean Analytics

Don’t sell what you can make; make what you can sell.

Data driven learning is the cornerstone of success in startups. It’s how you learn what’s working and iterate toward the right product and market before the money runs out.

Sometimes growth comes from what you don’t expect. If you’ve found an insight, decide how to test it quickly w/ minimal investment. Define what success looks like before hand and know what you’ll do if your hunch is right.

Analytics is about tracking metrics that are critical to your business.

You don’t always know what metrics are key b/c you don’t always know what business you’re in.

Jon Snow vs Daenerys Targaryen

California and Hawaii are catching on

Shortly after #deletefacebook became a thing, a survey that found 57% of Americans didn’t know Instagram was owned by Facebook. Looks like California and Hawaii might be catching on.

Speakeasy, Microbrewery, or Beer Garden

Using Google Search Trends to Correlate and Forecast Revenue

Google search can be highly correlated to revenue for many categories. In the example above, I pulled the yearly data from 2010 to 2018 for worldwide search queries for “Beer” from Google Trends and the worldwide revenue (USD) for Beer from Statista. I ran a linear regression against the data and it shows an R2 value of .987.

This is a really great fit and we can use this information to predict what Beer revenues for a particular period will be given the searches in that period or vice-versa.