Google Analytics Notes


Reference: Google Analytics for beginners

https://analytics.google.com/analytics/academy/course/6/unit/1/lesson/3

A user session starts from a user lands on a page contains the google analytics js code. It ends if there is 30 minutes of user inactivity.

In marketing, there are 3 stages to measure the digital business.

Acquisition – where does your traffic come from

Behavior – how users interact with your website

Conversion – the transaction happens

其实自媒体也适用这个model:你的粉丝从哪里来,你的粉丝如何与你互动(评论,转发,赞),你的粉丝是否购买你的产品或服务。

The google analytics

Organization (optional) -> account -> property(s)-> view(s)

properties: you can set one property for your US market, and set the other one for the APAC; or set one property for each of your brand; or set one property for website, and set another property for your app. However, you can NOT aggregate reports across multiple properties.

views:

filters: multiple filters can be set in one view. Such as exclude all internal traffic.

Best practice: create a test view, apply a filter, to see if it works as expected. If so, apply that filter to the master view or other real views.

The left navigation panel in Analytics

realtime – realtime users

audience – the geographic graph of your users. How old are they? Gender? Interests?

acquisition – where does the traffic come from? From organic search? paid click?

behavior – how people interact with your site? which page they view? which page they land and exit.

Bounce rate: the percentage of visitors who enter the site and then leave rather than continuing to view other pages within the same site. The lower the rate, the better the website performance. A high bounce rate indicates the site entrance page is not relevant to your visitors.

A 26 to 40 percent bounce rate is excellent. 41 to 55 percent is roughly average. 56 to 70 percent is higher than average, but may not be cause for alarm depending on the website. Anything over 70 percent is disappointing for everything outside of blogs, news, events, etc.

bounce rate在短视频app中的应用可以理解为:用户刷到了你的视频,只停留了几秒钟,便刷走了。

对于短视频而言,engage从低到高:刷到了视频,没完播,就刷走了;完播视频;点赞视频(应该有page rank类似的每日每用户有限定like总权重的算法);点开了作者主页,开始刷其他的视频。

Audience reports

gender, age, geographic, new vs returning

Acquisition reports

mediums: organic (unpaid search), CPC (paid search campaign), Referrals (traffic comes from another website (not search engine)), Email, none (user type url directly into the browser)

question: what category does the traffic goes into that is from IM apps like wechat?

source: provides information for the medium.

if the medium is Organic, source could describe the name of the search engine. such as Google, yandex, etc.

if the medium is referral, source describes the website url of the referral.

in the comparison view, you can check which medium sends you the traffic with the highest bounce rate. There might be some problems on the medium, probably 1) they are sending users to an unrelated landing page 2) they are sending unrelated users to your website.

Behavior report

the most frequently viewed page.

events: like track a button click event.

Advanced Google Analytics Course

https://analytics.google.com/analytics/academy/course/7/unit/1/lesson/1

How GA collects data – by sending the user data in url parameters to google-analytics server.

There are 3 kinds of hit events in GA.

Pageview hit – when user load a webpage with the tracking code.

Event hit – tracks user interaction with a particular element in your website. There are 4 parameters in the url of an event: action, category, label, value.

transaction hit – products purchased, transaction ID’s Stock Keeping Units(SKU).

Categorizing into users and sessions

GA uses browser cookies to identify users. Risks: a user is considered as a new user if 1) their browser cookie is cleared. 2) they forbidden the browser cookie 3) they use different devices to browse your website.

When a filter is added to a view, the data excluded from the view are lost, you can not add it back anymore.

Creating a measurement plan

Macro conversions: user purchase the product

Micro conversions: user sign up for the news-settler, new product notification.

Organize your analytics account

It is recommended to set up 3 views: raw data, test and master.

Test view is used to test the filter or other configurations, use the readtime traffic panel to check if your filter works or not.

Recommends to track website, mobile app in different properties. but analytics 360 users could also set an aggregated property to track the aggregated information for all properties.

Set up advanced filters on views

filter based on source ip addresses, by campaigns.

filter the url to ignore parameters and focus only on the uri.

For example, aggregates xxxx?id=2 with xxxx?id=3 into a same page view report.

Custom events

4 parameters in an event: category, action, label, value.

category: RSE

action: Open, add node, delete node, open protoshop Editor

label: “supported”, “unsupported”

total events: total triggered events

unique event: total users triggered the event

Reference: Google Analytics 360

https://analytics.google.com/analytics/academy/course/8/unit/1/lesson/3

Google Analytics 360 is the enterprise version of Google Analytics.

Custom Funnel

You can see the drop off rate on each stage when user interact with your site. And analyse what blocks them from further interacting with your site.

After you make the hypotheses, use the Google Optimize or Optimize 360 to run AB test to test your hypotheses.

Custom segment / remarketing

You can segment out those users who drop out at some stage on your website and trace back which Ads camp-ion acquires these users.

You can also segment out the users who finishes the whole journey and track back the ads campaign that acquires those.

Remarketing means bring the abandoned (the user who dropped off) back to complete a conversion; you can also use remarketing to offer additional products or services to users who have completed a conversion path. 简言之,就是没买货的,追上去让他们再买点;买了货的,让他们买更多的货。

Reporting with BigQuery

Analytics 360 connects to bigquery to run SQL on tegabytes data.

Leave a comment

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.