◆ PRODUCT ANALYTICS WITHOUT THE DATA TEAM
Ship it. Measure it.
Know by morning.
Kilowatt reads your product events and tells you, in plain sentences, what changed and why. No SQL, no warehouse, no analyst in the loop.
kilowatt.app / northwind / overview
⌕ Search or ask…⌘K
3
WORKSPACE
Overview
Alerts3
Metrics12
Segments7
Ask
SOURCES
Segment
Postgres
Stripe
SAVED VIEWS
Onboarding funnel
Enterprise cohort
Churn watch2
TEAM
+5
18 of 25 metrics
Overview
LIVE
ACTIVE TEAMS
1,284+6.2%
TRIAL → PAID
18.4%−18.0%
ALERTS SENT
37+3
MRR
$84.2k+2.4%
Trial → paid conversion
variant A
variant B
Mar 1Mar 6Mar 11Mar 16Mar 21
RECENT ALERTS
−18%Trial → paid2m
+6%Activation rate1h
−4%Weekly retention3h
+11%Signup → trial6h
−2%Invite accept9h
View all 37 alerts→
TOP SEGMENTSSHARE OF DROP
/pricing variant B412 users62%
Mobile Safari188 users24%
Self-serve, EU96 users9%
Invited teammates41 users5%
SOURCE HEALTH
Segmentlive · 84ms
Postgreslive · 210ms
Stripelive · 96ms
Amplitudepaused
SHIPPING WITH KILOWATT
NORTHWIND
ORBITAL
PAPERCUT
HELIOSTAT
MERIDIAN
FLATIRON
◆ WHAT IT DOES
Three jobs, done without you asking.
-
THRESHOLD
Watches every metric
Set a threshold once. Kilowatt checks it every five minutes and only speaks up when the number actually moves.
-
−18%
Trial → paid, 7d
Segment/pricing variant B
Releasev4.18 · 3 days ago
Share of drop62%
Explains the why
Every alert arrives with the segment, the release and the cohort that caused it — written as a sentence, not a dashboard.
-
› why did trial conversion drop last week
Variant B of /pricing converted 18% worse. It shipped 3 days before the drop.
SELECT variant, conv_rate FROM trials …
Answers in plain English
Ask “why did trial conversion drop last week” and get an answer with the query it ran, so you can check its work.
kilowatt / alerts / trial-conversion
DROP −18%
Trial → paid conversion, last 7 days
Caused by: signups from the /pricing A/B variant B — 62% of the drop.
◆ ALERTS THAT EXPLAIN THEMSELVES
The alert already
did the digging.
Most tools tell you a number moved. Kilowatt tells you which cohort moved it, which release it followed, and how much of the change it accounts for — before you open the app.
Segment + release attribution on every alert- Slack, email or webhook — your call
- Shows the query it ran, always
- CHECK INTERVAL
- 5 min
- LINES OF SQL
- 0
- ALERTS RATED USEFUL
- 93%
- MEDIAN TIME TO CAUSE
- 11 min
Live in about nine minutes.
-
01Drop in the snippet
One script tag, or the SDK if you prefer. We read the events you already send — no new instrumentation, no schema to agree on.
<script src="kilowatt.js"
data-key="kw_live_8f2a…" />
</script>
-
02Name what matters
Pick three to five metrics in plain language. “Trial to paid”, not a SQL fragment nobody else on the team can edit.
trial → paid activation weekly retention invite accept + add metric -
03Close the tab
Kilowatt watches from here. You hear from it only when a number actually moves — with the cause already attached.
Kilowatt
now
Trial → paid fell 18%. 62% of it came from /pricing variant B.
Teams that stopped checking.
NORTHWIND
★★★★★
“We cancelled the weekly metrics meeting. The alert that used to start it now just arrives, already explained.”
Rosa LindqvistVP Product · NORTHWIND
ORBITAL
“Our analyst got her week back. Kilowatt answers the questions she used to queue up on Mondays.”
Tomas Reiner · Head of Growth
PAPERCUT
“It caught a checkout regression four hours before support did. That alone paid for the year.”
Amara Okonjo · CTO
◆ PRICING
Move the slider. That's the price.
No seats, no tiers to outgrow. You pay for the metrics we watch — nothing else.
Unlimited seats — invite the whole company- 12-month history and full alert attribution
- Ask-in-English queries with the SQL shown
◆ NINE MINUTES TO SET UP