01
Analysis on demand
Dashboards, reports, and bespoke investigations for teams that need an answer, not a slide deck.
- Exploratory analysis
- Executive reporting
- Metric design
- Causal inference
CLARYON 0 · DATA ANALYSIS & LITERACY STUDIO
Decode. Discover. Decide.
We do the analysis your team can't, and teach your team to read numbers the way a journalist reads a quote. Half studio, half school. All receipts.
- 4.8 / 5
- cohort rating, n=312
- 120+
- audits delivered
- 2025
- est. · startup13
FIG.01 · REVENUE INDEX · 2Y
Same data. Two stories.
§ 01
What we do
Three products, one discipline. Each delivered by analysts who'd rather be precise than impressive.
02
Teaching & workshops
Live and async courses that train your team to read numbers with the same rigor a journalist reads a quote.
- Data literacy 101
- Statistics without tears
- Charts that don't lie
- Cohort programs
03
Data audits
Second opinions on the numbers your business runs on. We find the decimal point that moved.
- Metric reconciliation
- Dashboard review
- Forecast critique
- Bias & sampling check
§ 02
Read the numbers — a mini lesson
Three 90-second lessons from the first week of the curriculum. Drag, toggle, and watch the story change.
LESSON 01 / 03
Does revenue look like it exploded, or barely moved?
A chart where the y-axis starts near the data instead of at zero can turn a rounding error into a revolution. Toggle the axis below.
BE SUSPICIOUS
The line screams upward — but the y-axis is only showing a ~2-point window. The drift is noise, not growth.
READ CORRECTLY
Both groups average around 50, but Group B's standard deviation is nearly 7× Group A's. One group is uniform. The other contains two camps. An average without a distribution is a rumor.
ICE-CREAM SALES
119
cones / hour
DROWNINGS
3
incidents / day
CORRELATION (r)
0.92
very strong, positive
READ CORRECTLY
Both series rise together because both are caused by temperature. Selling fewer cones won't save anyone. The lurking variable is the real story.
§ 03
Selected work
Three engagements we're allowed to talk about. The rest are under NDA, as they should be.
RETAIL · 12 STORES
The revenue climb that wasn't
Same-store sales looked up 14% YoY. We found a store-count change and a pricing round-up. Real growth: 1.8%.
−12.2 pts
ADJUSTMENT AFTER AUDIT
SAAS · SERIES B
A churn metric that hid the truth
Monthly logo churn was flat. Net revenue retention by cohort told a different story — the largest accounts were shrinking quietly.
88 → 103
NRR AFTER RE-CUT
NON-PROFIT · HEALTH
Teaching a board to read its own data
Six two-hour sessions. The board now asks for denominators before dollar signs. Board satisfaction: up.
6 weeks
PROGRAM DURATION
§ 04
The school
An 8-week cohort for operators, founders, and anyone who's tired of pretending to understand the dashboard.
- WK 01What a number actually meansUnits, denominators, and the quiet assumptions baked into every metric.↗
- WK 02Averages lie (usually)Mean, median, mode — and when each one is the honest choice.↗
- WK 03How a chart can deceive youTruncated axes, dual axes, area tricks, and the 3D pie chart of shame.↗
- WK 04Correlation ≠ causationConfounders, Simpson's paradox, and the spurious correlations you'll love to spot.↗
- WK 05Sampling & survivorshipWhy your customer survey is probably wrong, and how to fix it.↗
- WK 06Forecasts & their confidenceReading error bars, P-values, and the grown-up version of "my gut says…".↗
- WK 07Telling the storyPicking the right chart, the right number, and the right headline.↗
- WK 08Capstone — read a real reportDissect a published analysis end-to-end. Find the flaws. Rewrite the headline.↗
§ 05
Questions
-
Yes — roughly half our engagements start there. We set up the basics, then teach your operators to run it.
-
We can. A typical audit starts by reproducing three or four numbers from first principles, then comparing.
-
Cohorts run quarterly. Private cohorts for teams of 8+ run on your schedule.
-
Whatever yours are. We've shipped in SQL, Python, R, spreadsheets, and the occasional napkin.