How Long Does it Take Trialists to Convert?

Data Collection

We collect the data from BigQuery and Redshift. Much of the code has been hidden.

How Long it Takes to Convert

Below we’ll plot the distribution of the number of days it takes to convert. We can see that Pro trialists are much more likely to convert on the first day of the trial and Business and Premium trialists are much more likely to convert on days 14-30.

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

How Acquisition Source Affects Trial Selection

Let’s plot the types of trials from users that signed up with the different attribution sources specified in the onboarding survey. We can see that users from each source are most likely to start a Pro trial. A Business trial is the second most common, followed by the Premium plan trial.

People that heard of Buffer through an acquaintance or through the blog are slightly more likely to have started a Pro trial. Users that heard of Buffer through ads, search, or social media are slightly more likely to have started a Business trial.

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

How Acquisition Source Affects Time to Upgrade

Now let’s see if users that heard of Buffer through different avenues take a different amount of time to convert. We’ll want to control for the type of trial here, so let’s start with the Pro trial.

The chart below shows the cumulative distribution function (CDF) for each segment. It’s a useful way to compare distributions. For each point on the x-axis that represents the number of days to convert, the y value represents the proportion of users that converted in x days or less.

For example, on day 0, we see that around 12% of conversions occur for the “advertisement” group, while around 26% of conversions for the “acquaintance” group occur on day 0.

In general users that heard of Buffer through ads take the longest time to convert.

Let’s look at the distribution in a more straightforward way.

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

Now let’s look at Business trials. Again we see that users that heard of Buffer through ads are the slowest to convert.

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