Published by Dan Cunning on May 29, 2014

Finding Your Most Active Users

Google Analytics is normally used to track a website’s visitors and pageviews, but it can also track specific users, which can be invaluable information as you develop your customer base. Who is the most active? What are they using? Whose usage is fading?

Filed under Web Services

Introduction

You're using Google Analytics to track your web app's visitors, but you don't have visitors: you have users. Don't you want to know what they are doing? Who are your most active users? What do they use the most? Who should you contact directly about a feature improvement? Whose usage has dropped off?

Tracking User IDs

Google Analytics accepts custom, site-specific data. It's really simple, just add one line to the snippet Google Analytics provides. The new Google Analytics snippet uses custom dimensions, which you need to set up explicitly. The older snippet uses custom variables, which doesn't require any set up.

// For New Snippets
ga('set', 'dimension1', <%= current_user.id %>);

// For Older Snippets
_gaq.push(['_setCustomVar', 1, 'User', <%= current_user.id %>, 1]);

Querying Google Analytics

Now that you're associating a specific user with each pageview, you can use the Google Analytics API to group pageviews by user. I use garb.

# For Newer Snippets
class UserActivity
  extend Garb::Model
  extend Garb::Queryable

  metrics :pageviews, :uniquePageviews
  dimensions :dimension1

  def initialize(struct)
    @pageviews = struct.pageviews.to_i

    @user_id = struct.dimension1.to_i
    @user_id = nil if @user_id == 0
  end

  def user
    @user ||= User.where(id: @user_id).first if @user_id
  end

  def email
    user.email if user
  end
end
# For Older Snippets
class UserActivity
  extend Garb::Model
  extend Garb::Queryable

  metrics :pageviews, :uniquePageviews
  dimensions :customVarName1, :customVarValue1

  def initialize(struct)
    @pageviews = struct.pageviews.to_i

    @user_id = struct.custom_var_value1.to_i if struct.custom_var_name1 == 'User'
    @user_id = nil if @user_id == 0
  end

  def user
    @user ||= User.where(id: @user_id).first if @user_id
  end

  def email
    user.email if user
  end
end
module Garb::Queryable
  # Instead of returning a Struct, return instances of the given class
  def query(profile, options = {})
    results(profile, options).inject([]) do |arr, result|
      arr << new(result)
    end
  end
end

Generating an Email

Emails are a passive way of tracking information. You don't need to go anywhere: it comes to you and fits nicely into most people's daily routines. Let's run a Rake task that authenticates with GA and delivers an email via ActionMailer.

class AnalyticsMailer < ActionMailer::Base
  def user_activity_report(to_email, profile)
    # Back one week from midnight last night
    start_at = Date.today - 1.week
    end_at = Date.today

    @activities = UserActivity.query(profile, start_date: start_at, end_date: end_at)
    @activities.sort! {|a, b| b.pageviews <=> a.pageviews }
    return if @activities.empty?

    mail_to({
      to: to_email,
      subject: "[#{profile.name}] User Activity"
    })
  end
end
<!-- views/analytics_mailer/user_activity_report.html.erb -->
<h2>Active Users</h2>

<table>
  <thead>
    <tr>
      <th>Email</th>
      <th>Pageviews</th>
      <th>Unique Pageviews</th>
    </tr>
  </thead>
  <tbody>
    <% @activities.each do |activity| %>
    <tr>
      <td><%= activity.email %></td>
      <td><%= activity.pageviews %></td>
      <td><%= activity.unique_pageviews %></td>
    </tr>
    <% end %>
  </tbody>
</table>
# lib/tasks/analytics.rake
namespace :analytics do
  namespace :user_activity_report do

    desc 'Delivers the user activity report'
    task deliver: :environment do
      email = 'dan@cunning.cc'
      password = 'use-two-factor-auth-instead'
      site_name = 'dan.cunning.cc' # name of your Google Analytics "Property"

      Garb::Session.login(email, password)
      # Garb::Session.api_key = api_key # required for 2-step authentication

      profiles = Garb::Management::Profile.all
      profile = profiles.detect {|p| p.name == site_name }

      AnalyticsMailer.user_activity_report(email, profile).deliver
    end
  end
end

Scheduling the Email

The whenever gem provides a slick interface for managing your application's cronjobs, with an easy integration into Capistrano.

# config/schedule.rb
every :monday, at: "8:00am" do
  rake "analytics:user_activity_report:deliver"
end
# config/deploy/production.rb
require "whenever/capistrano"
# config/schedule.rb
every :monday, at: "8:00am" do
  rake "analytics:user_activity_report:deliver"
end

Conclusion

Now you'll get an email every Monday morning telling you who your most active users were last week. The email takes just a couple seconds to browse over and can really improve and target your customer service and product development.

Finding your most active users is only the beginning. Google Analytics can answer a lot more questions:

  • How much time did a user spend on each page? What was the average?
  • What individual pages did a specific someone use? In what order?
  • Who are the major users of a specific page?
  • Who had 50% more/less pageviews this week than last week?

Check out all the Dimensions & Metrics