Analyzing Extension Usage

Data Collection

When an update is created via the extension, we set the via field to bookmarklet in the updates database. We’ll grab all of the updates created with the extension in the past 60 days with the SQL query below.

select
  up.id
  , date(up.created_at) as created_date
  , up.user_id
  , u.billing_plan_name as plan
  , up.profile_service
  , r.is_target_customer
from dbt.updates up
inner join dbt.users u
  on up.user_id = u.id
left join dbt.target_customer_survey_responses as r
  on up.user_id = r.user_id
where up.via = 'bookmarklet'
  and up.was_sent_with_buffer
  and up.created_at is not null
  and up.created_at >= current_date - 60

There are around 3.7 million updates in this dataset.

How many people use the extension?

Let’s start by plotting the number of extension updates created per day.

## `summarise()` ungrouping output (override with `.groups` argument)

It looks like around 40 thousand to 80 thousand extension updates are created each day. The number of posts is highly seasonal and dependent on the day of the week. Now let’s plot the number of users that create these posts.

## `summarise()` ungrouping output (override with `.groups` argument)

There are around 4 thousand to 9 thousand users that create updates from the extension on a given day. This data is highly seasonal, so let’s look at the number of users that use the extension in a given week instead.

## `summarise()` ungrouping output (override with `.groups` argument)

There are around 20 thousand active users using the extension each week. This indicates that there are some users that are creating huge numbers of posts through the extension. Let’s count the number of posts created by each user in the past 60 days. The plot below only shows the top 30 users in terms of the number of extension updates created.

## `summarise()` regrouping output by 'user_id' (override with `.groups` argument)

Wow, one user created over 50 thousand updates with the extension in the past 60 days. The color of the bars represents the plan that the users are on.

Who are these users?

Let’s plot the number of extension updates created by users on each plan.

## `summarise()` regrouping output by 'created_date' (override with `.groups` argument)

We can see that users on the Free plan create the majority of extension posts. Users on the old Awesome plan follow, as do users on the Pro plan. These are our largest customer segments, so this makes sense. Plotting the percentage of updates coming from users on each plan might help clear up some of the noise.

## `summarise()` regrouping output by 'created_date' (override with `.groups` argument)

We can see here that Free users create almost half of all extension updates. Legacy Awesome users create slighly less than 30% of extension updates, and Pro users create around 15% of extension updates.

## `summarise()` regrouping output by 'created_date' (override with `.groups` argument)

Of the 633 users that signed up in the past 60 days and fit our target customer description, only 13 have created updates with the extension.

Which networks do people share to?

We can create similar plots to the ones above to show the most popular networks. We don’t want the percentages to be heavily skewed by the small number of users sending thousand of updates, so we will calculate the proportion of users sharing to each network.

## `summarise()` regrouping output by 'created_date' (override with `.groups` argument)

Twitter is by far the most common network, followed by Facebook and Linkedin.

Summary

The data shows that a large number of users still use the extension. Some of these users rely on the extension heavily, but their usage numbers lie at the end of a long tail.

Most of the feature’s usage comes from Free users, but a significant number of paying customers make use of the extension as well. A small percentage of our target customers have found it useful so far.

Twitter is the most popular network for extension updates by a wide margin, followed by Facebook and Linkedin.