All your products' numbers on one screen you can trust
For each feature: what was painful, what you see on screen, the order you use it in, and what you end up with.
Same-definition comparison
The problem
Every product uses a different analytics tool with different definitions of visitor, signup, and activation, so you can't line up the numbers and see which product is rising or slipping. Comparing today, which isn't over yet, with yesterday always looks like a drop.
Product comparison
One definition, one table · changes compare two completed 30-day windows · counts instead of % when the earlier window is under 30
| Product | Visitors | Users | Signups | Activated | First paid |
|---|---|---|---|---|---|
Budget Note budget-note | 4,812 ▲ 12% | 412 ▲ 5% | 158 ▲ 8% | 96 ▲ 3% | 9 ▲ +3 |
Meeting Brief meeting-brief | 640 ▼ 4% | 133 ▼ 2% | 117 ▲ 3% | 13 ▼ −2 | 2 ±0 |
Wordbook wordbook Delayed · last event 7 h ago | 86 ▲ 9% | 21 ▲ +4 | 8 ▲ +2 | 4 ▲ +1 | 0 ±0 |
Swipe the table sideways to see more
How you use it
- 1Sign up, register a product, and add the one-line snippet or server events with its key.
- 2Home shows every product as one row. Pick 1, 7, 30, or 90 days.
- 3Click a row to open that product's overview: metric cards, daily trend, channels, and path, country, and device tables.
What you get
One screen shows which number moved for which product.
Server-only key events
The problem
Signup and payment events sent from the browser get blocked by ad blockers, arrive twice on reload, or can be faked by anyone who calls the endpoint. Signup and payment numbers end up inflated or missing.
Ingest health
Per product: events in the last 24 hours, the last event received, and why events were refused
| Product | Events 24h | Last received | Rejected 24h |
|---|---|---|---|
| Budget Note | 1,940 | 2 min ago | unknownEvent 2, rateLimited 5 |
| Meeting Brief | 310 | 11 min ago | None |
| Wordbook | 42 | 7 h ago | serverOnly 1 |
Swipe the table sideways to see more
serverOnly means an event like signup, activation, or payment that is never accepted from a browser. It is refused and only the reason is kept.
How you use it
- 1The browser (the snippet or your product's proxy) sends visits and the product events you registered.
- 2Signups, activations, and payments are sent by your server with a server key. The user identifier is computed on the server; any value from the browser is dropped.
- 3If the browser sends a server-only name, it is refused and shows up only as a rejection reason in the product's ingest health.
What you get
You can take signup and payment numbers at face value and compare products.
Retried server delivery
The problem
If the network drops just as your server sends a signup or payment, that event is gone and the day's numbers stay incomplete. You end up reading conversions without knowing something is missing.
baselogOutbox · Product database
Rows are added in the same write as the signup, activation, or payment
| kind | name | status | attempts | next try |
|---|---|---|---|---|
| event | signup | delivered | 1 | — |
| event | activate | pending | 2 | +4m |
| event | payment | delivered | 3 | — |
| delete | — | delivered | 1 | — |
How you use it
- 1When a signup, activation, or payment happens, the same database write adds a row to your product's outbox.
- 2Only events Baselog confirms as stored are closed as delivered; failures are retried with growing gaps for 7 days. If the same event arrives twice, it counts once.
- 3Events that were sent again show up on the overview and comparison from the next calculation.
What you get
Days with failed sends still have complete signup and payment numbers.
Try before signing up
Free calculators that run only in your browser. No account, nothing stored.
- Funnel Conversion CalculatorEnter the people at each step to get step conversion, overall conversion, and the weakest step.
- A/B Test Sample Size CalculatorFind the visitors you need per variant from baseline rate, minimum detectable effect, significance, and power.
- Cohort Retention CalculatorTurn weekly signup cohorts and returning users into a retention table with averages from finished weeks only.
Questions by feature
Same-definition comparison
What does the change compare?
Two completed windows. For 7 days, that's the 7 days ending yesterday against the 7 days before. Today, still in progress, is left out. If the earlier window has fewer than 30 people, the change is shown as a count instead of a percentage.
Is the period's visitor count the sum of daily visitors?
No. The same person visiting on several days would be counted more than once, so period counts come from raw events. Days are cut on the Korea time (KST) calendar.
Does a zero really mean nobody came?
The screen draws three cases differently: "Not collecting" when no event has ever arrived, "Delayed" when the last one is more than 6 hours old, and "Updating" while numbers are recalculated. Any other zero is a real zero.
Server-only key events
Are visitors counted on the server too?
No. Visits come from the browser or the app, so visit numbers are treated as unauthenticated estimates, and what ad blockers drop is shown separately as estimated collection loss.
Do you receive payment amounts?
Yes. Amounts ride only on payment and refund events and are summed per currency. They're there to see which channel led to payment; your payment provider and books remain the source of truth for revenue.
Where can I see refused events?
The product's ingest health shows counts by reason for the last 24 hours. Raw payloads are never kept.
Retried server delivery
Are browser visit events retried too?
No. Web visits are sent one at a time and not retried on failure; that share is shown as estimated collection loss. Retries are for server events.
What happens if sending keeps failing for 7 days?
The event is marked as no longer retried. Deletion requests from account deletion are the exception: they keep retrying until they succeed.
Do late events make it into the numbers?
Yes. Metrics are recomputed from raw events, so a late event shows up in the trend from the next calculation.
Register your product and receive your first event
Sign up with email and add one snippet line; your product gets a row on the comparison. Free for now.