Distributions
Analyze the shape of order values.
Use the page to
- Set a CAC threshold against your typical order, not your average.
- Check whether an AOV initiative — bundles, upsells, free-shipping thresholds — actually made things better.
- Pick where to set free-shipping or bundle thresholds.
What is a distribution?
A distribution is a visualisation of how your orders are spread out by value. Instead of collapsing every order into one number like average order value, it shows you which order values have the most orders. The higher the chart — the more mass — the more orders around that value.
The intuitive way to draw a distribution is a histogram: define different intervals and count how many orders fall in each. The problem is that the picture you get depends entirely on how wide you make the intervals. Pick intervals so narrow that each one holds only a single order and you get a flat histogram. Pick intervals so wide that every order lands in the same one and you get a single tall column. The right interval width sits somewhere in between — but there’s no single width that works across every store, currency, and product type.
So instead we use a more correct method called kernel density estimation (KDE). Rather than slotting orders into fixed intervals, KDE moves along the chart and, at every point, asks how many orders sit nearby. Kind of like if you imagine an abacus where every order is a bead sitting at its own value — the more beads piled up at a point, the higher the chart climbs there.
Metrics
Switch between Sales and CM2 at the top of the page to change what is measured. Sales is the order value before any returns. CM2 is contribution margin 2 — what’s left after returns, COGS, and all variable costs.
Mean is the arithmetic average. If three orders are worth 300, 400, and 500, the Mean is (300 + 400 + 500) / 3 = 400.
Mode is normally the value that repeats most often. For example, in 200, 300, 300, 400 the Mode is 300. But order values rarely repeat exactly down to the cent, so we find Mode by looking at where orders cluster on the chart instead — the highest point of the curve.
Why this may differ from Shopify AOV
Shopify’s AOV calculation doesn’t include shipping revenue, taxes, or order edits. Kleio’s Sales includes shipping, can include tax (via the toggle in the filter panel), and reflects items added or removed via order edit.
Valuable analyses
Set a CAC threshold against your typical order, not your average
Mean doesn’t tell the whole story. If you have nine $10 orders and one $110 order, the average is $20. But if you pay $10 per order in CAC, nine of those orders are deeply unprofitable after ads, shipping, and everything else — they rely on the tenth order to make up the gap.
There’s technically nothing wrong with that. If the big order keeps showing up, you might be fine. But starting the fulfillment machine for orders that lose money on their own is inefficient, and it leaves you exposed to cash-flow risk and variance.
Verifying an AOV initiative
After running a bundle, upsell, or free-shipping threshold change, turn on the comparison toggle and see whether the curve shifted to the right and whether Mean or Mode increased. If the whole curve moved, the initiative shifted the distribution. If only the long tail moved, you only persuaded the customers who were already buying a lot.
Picking free-shipping or bundle thresholds
Setting a threshold just above where orders cluster pushes more customers to add another item rather than coast on a threshold they already hit.