Google Analytics App now offers Google Now-like automated insights

The latest update to the Google Analytics apps (iOS and Android) offers a sleek new machine learning-powered set of insights a la Google Now.

You’ll find the insights on the Assistant screen, accessed from the menu icon in the upper left corner of the app.

Below are a couple of examples from Google of the types of insights you might see for your account. One highlights fast-rising products, showing week-over-week revenue growth at the product level. The other spotlights an increase in users week-over-week.

The insights don’t just point out the good. Cards can call out decreases in ecommerce conversion rate, average session duration, and even point out below average performance on a particular device (see below) or screen resolution.

You may also see Quick Tips and other recommendations in some insight cards.

Often there is the option to view more details for an insight. The insight detail provides data such as which segments are influencing performance. For example, the screenshot below shows a partial view of the detail that a site “performs below average on tablet”.

You’ll notice there is also a feedback request (thumbs up or down) at the bottom of this detail screen. Google takes that feedback along with how one uses the insights — engaging with, pinning, sharing, etc. — to continue iterating and improving on the insights provided.

To customize a sort of insights dashboard, you can “pin” (click the pin icon in the top right) insights and then from the Assistant home screen click the pin icon slider to show only those selected.

You can easily share insight cards — to text, email, Slack, social media etc. — by clicking the upload/share icon in the top right menu.

Right now, automated insights are only available in English, but Google plans to add support for more languages and bring this functionality to the desktop version of Analytics in the coming months.

I can see the Assistant tab becoming a top destination for Google Analytics app users.