Event analytics
Answers before you have finished asking
Cadence reads your events as they happen and answers in under a second — on your own data, in your own region, with the query it ran printed underneath.
Running in production at
- Northbank
- Alto Freight
- Verity Health
- Ostara
- Fieldmark
Three things it does properly
-
Answers while you are still asking Events land in the store in under two seconds and queries return in under one. There is no nightly build, so the dashboard you open at 09:00 includes what happened at 08:59.
-
Shows the query it ran Every chart has a "how was this made" link that prints the exact SQL behind it. An analyst can copy it, change it, and save the result as a new chart without leaving the page.
-
Stays in one region Your account is pinned to Frankfurt, Dublin or Virginia and nothing leaves it — not backups, not logs, not the copy our support team reads when you open a ticket.
Four steps to the first chart
Measured on the last two hundred accounts: half were reading their own data inside a working day.
-
Point it at your data
Postgres, BigQuery, Snowflake, S3 or an HTTP endpoint. Read-only credentials, and we tell you exactly which tables we touched.
-
Name your events
Map the columns you already have onto the four fields Cadence needs. No new tracking code and no schema migration.
-
Open a board
Start from one of eleven boards built for common questions, or from an empty one. Both are editable by anyone who can read a chart.
-
Put it somewhere useful
Send a board to Slack on a schedule, alert on a threshold, or expose it as an endpoint your own service reads.
Against a warehouse-and-BI stack
The arrangement most teams already have, and where it costs them.
| What you are comparing | Cadence | Warehouse plus a BI tool |
|---|---|---|
| Time from event to chart | Under two seconds | One to twenty-four hours |
| Time to the first useful board | Hours | Weeks, once the pipeline is built |
| Query behind a chart is visible | ||
| Needs a pipeline to maintain | ||
| Cost of one more dashboard | Nothing | A seat, or an analyst's week |
| Data leaves your region | Depends on the vendor |
We had a warehouse and four dashboards nobody trusted, because nobody could see how any figure was produced. The "show me the query" link ended that argument in a week.
Priced on events, not on seats
Everybody in your company can read a board on every tier. Charging per seat makes people share logins, which is a security problem sold as a pricing model.
Starter
One product, one region, one team.
€290 /month
Ask about a trial- Up to 20 million events a month
- Unlimited people
- Unlimited boards
- Thirteen months of history
- Email support, answered in four hours
Most accounts
Company
Several products, and somebody whose job is the data.
€890 /month
Ask about a trial- Up to 200 million events a month
- Everything in Starter
- Region pinning
- Audit log export
- Single sign-on
- Shared Slack channel with our engineers
Estate
Volumes where the invoice deserves a conversation.
Let us talk
Discuss a deployment- Everything in Company
- Private deployment in your own account
- A named engineer
- Contractual availability with automatic credits
- Security review with your team, not a questionnaire
The questions we are asked before every signature
If yours is not here, write to us and it will be by next week.
How long does moving off our current tool actually take?
Reading your data takes an afternoon. Rebuilding the boards people rely on takes as long as there are boards — most teams find they rebuild a third of them and quietly retire the rest. We will do the first five with you on a call, and we do not charge for that.
Where does our data live, and who can read it?
In the region you pick, and nowhere else. Our staff have no standing access to it; access is granted per incident, expires after four hours, and is written to the same audit log you can export. The security page has the detail your compliance team will ask for.
Can we run it inside our own cloud account?
On the Estate tier, yes — in your account, in your VPC, with us holding no keys. It is the same build, and it updates on the schedule you choose rather than ours.
What happens if we go over the event allowance?
Nothing stops. Overage is billed at the tier rate and appears on the next invoice, and we email you when you cross ninety per cent so it is never a surprise. We have never switched an account off for going over.
What do we get back if we leave?
Everything, in Parquet, for thirty days after the subscription ends — the raw events and the definition of every board and alert. Export is a button, not a support ticket.
Fourteen days, no card
Point it at one week of your own events
A trial runs on your data, not on a sample set, because a demonstration on somebody else's numbers proves nothing about yours. Nothing is deleted when a trial ends: if you do not continue, the account is exported and closed and we write to say it is done.