Meet the SaaS Dataset

Before computing any KPI, spend ten minutes getting a feel for the data. This page is guided exploration: nothing is graded. Run each query, then tweak it until the shape of the dataset feels familiar.

Three tables, three grains:

• saas_accounts: one row per account (dimension)
• saas_subscriptions: one row per account per active month (the MRR fact)
• saas_invoices: one row per invoice (fact)

Start by counting rows. 12 accounts producing ~96 subscription rows over 12 months already tells you not every account was active every month.

The subscription fact runs from 2023-06 to 2024-05. Count active accounts per month: the count rises as new accounts land, and dips when accounts churn or pause. This per-month view is the spine every MRR metric hangs on.

Trace a single account through time. Account 7 (Quill & Parcel) is the most interesting one: it pays from 2023-09 through 2023-12, disappears for two months, then comes back in 2024-03: a churn followed by a reactivation. Remember: the gap is missing rows, not NULLs.

Now look at total MRR per month and its month-over-month change. The portfolio grows from 2,500 in 2023-06 to 8,748 in 2024-05, but not in a straight line. The dips and jumps are exactly the churn, expansion, contraction, and new-business events you will quantify in the drills.

Finally, peek at the invoices. Monthly invoice history starts in 2024-01 (earlier billing lived in a legacy system), two enterprise accounts prepaid annually instead of monthly, one invoice was voided and re-issued, and two May invoices are still open. These quirks are deliberate — reconciliation tasks later in the track lean on them.