How to set up Stripe reports
Jun 21, 2024
Creating and analyzing reports for your Stripe data helps you understand how you are making money and how you can improve.
This tutorial shows you how to sync your Stripe data to PostHog and then create a report of insights like customer count, gross revenue, recurring revenue, revenue churn, and more.
Linking Stripe data to PostHog
To start, you need both a Stripe and PostHog account. Once have those, head to PostHog's data warehouse tab and:
- Click Link source
- Choose the Stripe option by clicking Link
- Enter your account ID and a restricted API key that can read the resources you want to query
- Press Next, keep all tables selected and click Import
Once done, PostHog will automatically pull and format your Stripe data for querying. You can adjust the sync frequency, see the last successful run, and more in data warehouse settings.
Note: If you are missing a table, check your data warehouse settings to make sure it synced correctly.
Creating insights for your Stripe report
Now that your Stripe data is synced into PostHog, you can use it to create insights for your report. Each of these requires you to create a new insight in the product analytics tab.
Customer count
To start, we create a trend of customer count over time.
On the trends tab, change the aggregation to Unique users, and then click the data series, go to the Data Warehouse tab, hover over the stripe_customer
table, and click Select. You might want to change the Distinct ID field from id
to email
as Stripe can give multiple id
values to the same user email.
This can also be done for any of the other Stripe data like charges, subscriptions, and invoices. You can also add filters based on property values like created_at
, email
, status
, and more.
Gross revenue
Next, we can get our gross revenue by doing a similar process and selecting stripe_charge
. For aggregation, we want Property value sum and then choose amount
. We also want to filter out failed charges by clicking the filter button, selecting the status
property, and making sure it doesn't equal failed
.
Finally, to clean up the visualization, click enable formula mode to divide by 100 (the amount
value is in cents) and click Options on the chart to add $
as a prefix.
Monthly recurring revenue (average revenue per customer)
There are many ways to calculate monthly recurring revenue, but the easiest and most common is multiplying the number of customers by the average revenue per customer per month.
To do this, we'll rely on the stripe_invoice
series. Make sure to set customer_id
as the distinct ID field before you select it, and then change the aggregation type to unique users. We also want to filter out invoices with nothing paid, so add a filter where amount_paid
is greater than 0.
Next, we can copy that series to create another but modify it to aggregate by property value average of amount_paid
.
Finally, we use formula mode to divide the amount by 100 and then multiply by the number of users with (B/100)*A
. You can add the prefix and likely want to change the graph to the last 180 days grouped by month.
Monthly recurring revenue (the Stripe way)
Stripe calculates MRR by "summing the monthly-normalized amounts of all active subscriptions at that time."
To mimic this calculation in PostHog, we need to write an SQL query that gets all the subscription items, normalizes the subscription amount, and then sums them up for each month. Because a lot of this data is in JSON
, we need to extract the values.
WITH subscription_items AS (SELECTid,current_period_start,JSONExtractArrayRaw(items, 'data') AS data_itemsFROM stripe_subscription),flattened_items AS (SELECTid,current_period_start,arrayJoin(data_items) AS itemFROM subscription_items)SELECTtoStartOfMonth(current_period_start) AS month,sum(casewhen JSONExtractString(JSONExtractRaw(item, 'plan'), 'interval') = 'month'then JSONExtractFloat(JSONExtractRaw(item, 'plan'), 'amount')when JSONExtractString(JSONExtractRaw(item, 'plan'), 'interval') = 'year'then JSONExtractFloat(JSONExtractRaw(item, 'plan'), 'amount') / 12else 0end) / 100 AS MRRFROM flattened_itemsWHEREJSONExtractBool(JSONExtractRaw(item, 'plan'), 'active') = trueGROUP BY monthORDER BY month DESC
Revenue churn
For many companies, the amount of money they lose is just as important as the amount they retain. To measure this, we can track revenue churn.
To do this, we write SQL to query both the stripe_invoice
and stripe_subscription
for users with an invoice but without a subscription. This requires joining the tables together on the customer_id
value, and looks like this for the last 30 days:
withrecent_invoices as (select customer_id, amount_paid, created_atfrom stripe_invoicewhere created_at >= now() - INTERVAL 30 day),subscribed_customers as (select DISTINCT customer_idfrom stripe_subscription)select sum(recent_invoices.amount_paid)/100from recent_invoicesLEFT JOIN subscribed_customers on subscribed_customers.customer_id = recent_invoices.customer_idwhere empty(subscribed_customers.customer_id)
Revenue growth rate
To get revenue growth rate, query stripe_invoice
to get the monthly amount paid sum, but then we use window functions to calculate growth. These smooth out the month-over-month changes to give us a 3-month average to use in our growth rate calculation.
WITH monthly_mrr AS (SELECTtoStartOfMonth(created_at) AS month,sum(amount_paid) / 100 AS mrrFROM prod_stripe_invoiceWHERE status = 'paid'GROUP BY monthORDER BY month),mrr_with_growth AS (SELECTmonth,mrr,avg(mrr) OVER (ORDER BY monthROWS BETWEEN 2 PRECEDING AND CURRENT ROW) AS mrr_avg,avg(mrr) OVER (ORDER BY monthROWS BETWEEN 3 PRECEDING AND 1 PRECEDING) AS previous_mrr_avgFROM monthly_mrr)SELECTmonth,mrr_avg AS mrr,previous_mrr_avg AS previous_mrr,(mrr_avg - previous_mrr_avg) * 100.0 / previous_mrr_avg AS mrr_growth_rateFROM mrr_with_growthWHERE previous_mrr_avg IS NOT NULLORDER BY month
The nice part about this is that we can still visualize the query data as a graph by choosing the line chart option below the query and then choosing month as the X-axis and MRR growth as the Y-axis.
Usage by top customers
The great part of syncing your Stripe data in PostHog is using it alongside your product data. An example of doing this is querying the usage of top customers.
To do this, we get the top customers from the stripe_invoice
and join their emails with their PostHog distinct_id
to get a count of their events.
withtop_customers as (select customer_email, sum(amount_paid) / 100 as total_paidfrom stripe_invoicewhere created_at >= now() - INTERVAL 30 daygroup by customer_email),big_events as (select count(*) as event_count, distinct_idfrom eventsgroup by distinct_id)select customer_email, total_paid, event_countfrom top_customersleft join big_events on big_events.distinct_id = top_customers.customer_emailorder by total_paid desc
You can further break this down by filtering for specific events like home_api_called
.
Further reading
- The basics of SQL for analytics
- Using HogQL for advanced breakdowns
- Adventures in null handling: Null money, null problems