Metadata
Metadata is the set of custom key/value fields you can attach to objects in Stripe and Paddle. Your billing provider treats them as inert notes; Chartsy reads them and turns them into dimensions you can filter and break down by. This is usually what separates generic SaaS metrics from metrics about your specific business.
Definition
Metadata - a key/value pair stored on a customer, subscription, or payment in your provider. The key is a label you choose (acquisition_source, company_size, coupon_code) and the value is what applies to that particular record (webinar, 51-200, LAUNCH20).
Metadata describes what a record is, and it is attached to your billing records rather than to site traffic - so it is unrelated to events, which count tracked visits and nothing else. Setting or changing a metadata field consumes no usage at all.
What it unlocks
Every standard metric can be sliced by any metadata key you use. That turns one number into a comparison:
- Breakdowns
- MRR by acquisition source, churn by company size, revenue by coupon code - the same metric split along a dimension only you have.
- Filters
- Scope a whole dashboard to one segment - just enterprise accounts, just customers who used a launch discount - and read the familiar charts about that group.
- Cohort questions
- Ask ChartsyAI things like "what is the churn rate for customers whose plan_type is annual?" without defining a segment in advance.
The practical walkthrough lives in How to use metadata analytics.
Naming keys so they stay useful
Metadata is only as good as its consistency, and the failure mode is always the same: three keys that mean one thing.
- Pick one key name and keep it.
acquisition_sourceandsourceandutm_sourceon different records become three separate dimensions, each with a partial view. - Keep values to a small, closed set.
Enterprise,enterpriseandENTwill not group together. - Prefer buckets over raw numbers.
51-200breaks down cleanly; an exact headcount of 137 produces a chart with one bar per company. - Set metadata at creation time. A field added later exists only on records created after that point, which quietly biases historical comparisons.
Do not put personal data in metadata
Metadata is descriptive, not a place for names, emails, or anything else identifying - particularly if you run GDPR mode, where the point is that such fields never reach Chartsy at all.
Metadata vs. Growth attribution
Both can tell you where customers came from, and they are not interchangeable. Metadata records what you knew at the moment the record was created, which means it depends on your own code writing the right value. Growth attribution observes the visit itself, so it captures channels you never thought to tag.
Use Growth for marketing channel questions, and metadata for facts about the customer that no amount of click tracking would reveal - their industry, their size, their contract type.
