Shopify & direct to consumer
The channel you own, run on its own economics
Direct isn't Amazon with better margins. It's a different business, where the money is made in repeat purchase and lost in shipping, discounting and fulfilment.
Most Shopify stores are measured on conversion rate and abandoned carts. Those are real, and they are rarely where the money is. The larger numbers sit in the customer base you already have, and in operational settings nobody has revisited since launch.
Direct is also the only channel where the shipping bill is yours. On a marketplace the fee is a fee; here it is a decision you made once, at a round number, and have been paying for ever since.
Every order band on the store contributes except one: the band immediately above the free-shipping threshold. A customer who spends $52 is worth less than one who spends $46, because at $50 the store starts paying the carrier and the basket has not grown enough to cover it.
Contribution per order, by order value
Net of product cost, payment fees, pick-and-pack and the shipping actually paid — not the shipping charged. Bars above the line make money, bars below it lose money, and the depth of colour follows the size of the number in either direction.
View as table
Illustrative composite account · trailing 90 days, 18,400 orders.
What we would do about it. The threshold is not wrong, it is in the wrong place. Moving it to $65 turns the losing band into the second-best one; so does a lighter carrier rate on that weight, or one accessory merchandised at the cart to push the basket past its own cost. Which of the three is right is an arithmetic question, and this is the chart that answers it.
The same store, three acquisition cohorts, one year apart in behaviour. Direct is not won at the first order — it is won at the second, and a cohort curve is the only place that shows up early enough to act on.
Cumulative contribution per customer, by acquisition cohort
Each line is one quarter's new customers, followed from their first order. The dashed rule is what it cost to acquire them. Cohorts are ordered in time, so the ramp runs light to dark from oldest to newest — the colour carries the sequence, not a category.
Illustrative composite account. The newest cohort has seven months of history, so its line stops there.
What we run on a direct account
Six things, all of them measured in contribution rather than in sessions.
- 01
Retention and the customer file
RFM segmentation, second-order programmes and win-back — campaigns that follow from the segments rather than from the calendar.
- 02
Shipping and fulfilment economics
Rate, threshold and zone modelled against real orders by weight, so the free-shipping line sits where it stops paying rather than where it looked round.
- 03
Subscriptions and loyalty
Cadence, churn reasons and the offers that hold a subscriber past the third fill.
- 04
Lifecycle email and SMS
On whatever you already run — Mailchimp, Klaviyo, Braze. We do not ask anyone to migrate to ours.
- 05
Merchandising and offer hygiene
Promotions nobody turned off, codes still honoured after expiry, fraud flags, and the fulfilment delays that read as churn.
- 06
Site and checkout
The transactional pages, judged on contribution per order rather than on conversion rate alone.
Seen in the work
Direct is the only channel where you own the customer file. It is also the only one where the shipping bill is yours — and both facts show up in the same model.
How we model a channel →