Personalizing abandoned cart emails by cart value means splitting your flow into two or three branches at the trigger: low-value carts get a short reminder and never a discount, mid-value carts get your standard sequence, and high-value carts get more patience, more reassurance, and a human touch instead of a percentage off. The mechanic is a simple condition on cart total — for example, under €40, €40–150, over €150 — but the effect is that your most expensive mistake disappears: sending the same 10% code to someone abandoning a €15 phone case and someone abandoning a €400 espresso machine. One discount costs you €1.50 you didn’t need to spend; the other cheapens a purchase that needed confidence, not a coupon.

One flow, three very different shoppers

If you run a store past the €20k-a-month mark, your abandoned carts aren’t one audience. The €18 cart is often an impulse that faded — recover it cheaply or let it go. The €90 cart is your bread-and-butter buyer weighing shipping costs. The €350 cart is someone doing real research, probably comparing you against two other stores in open tabs. A single generic flow treats all three identically, which means it’s mistuned for at least two of them at any moment.

The usual response to soft recovery numbers is to turn up the incentive — 10% becomes 15%, free shipping becomes free shipping plus a gift. That raises recovery rate while quietly wrecking the economics underneath it, because the incentive applies equally to carts where it’s margin-fatal and carts where it’s irrelevant. Ads have the same flaw at this stage: retargeting bids don’t know your margin per cart either.

The margin math of a blanket discount

Illustrative numbers: a store sends 500 final cart emails a month with a 10% code, evenly split between €20 carts and €200 carts. Say 40 orders are recovered. On the €20 carts, each code costs €2 — trivial. On the €200 carts, each costs €20, and if your net margin on that product is 25% (€50), one blanket code just handed back 40% of the profit on every recovered high-value order — including the ones that would have completed after a simple reminder. Value-based branching exists to stop exactly this transfer. For catalogs where margins are thin across the board, cart emails for low-margin products takes this further.

How to roll it out, in order

1. Get a standard flow working first. Branching a flow that doesn’t exist helps nobody — the base structure is in how to build a three-message abandoned cart flow.
2. Find your thresholds from data, not gut. Pull three months of orders and abandoned carts. Your low threshold is roughly where a discount stops being pocket change relative to margin; your high threshold is around your top 10–20% of cart values. For many stores that lands near “under €40 / €40–150 / over €150”, but yours may differ by a lot.
3. Split the flow at entry with two conditions. Start with just two branches (below/above one threshold) if three feels heavy. You can add the third later.
4. Differentiate the endings before the openings. The first reminder can stay nearly identical across branches; it’s the incentive and timing at the end that must diverge.
5. Not needed yet: per-product personalization, predictive scoring, or AI-written copy. Cart total is one field you already have, and it captures most of the signal.

The spec for each branch

Shared: trigger on cart/checkout abandonment for identifiable contacts; exit on purchase; condition step on cart total immediately after the trigger.

Low-value branch (illustrative: under €40)

• Messages: two emails maximum. Reminder at 1 hour; short nudge at 24 hours. Channel: email only — SMS costs eat these margins.
• Content: product photo, checkout button, done. No discount ever; a code here often costs more than the profit on the item.
• Goal: cheap recovery. If it doesn’t convert in 48 hours, the flow ends quietly.

Mid-value branch (€40–150)

• Messages: your full three-message sequence — reminder at 1 hour, objections at 24, last call at 48–72.
• Content: standard playbook; incentive in the final email only, only for first-time buyers, and only if your margin allows — sanity-check it against when to add a discount to a cart flow.
• Goal: maximize recovered revenue per recipient.

High-value branch (over €150)

• Messages: three or four touches over 5–7 days — big carts have longer decision cycles. First email still at ~1 hour, then 24h, 72h, day 6.
• Channel: email, plus an optional SMS near the end; for the biggest carts, a plain-text “anything I can answer?” email that looks like it came from the owner.
• Content: reassurance over pressure — warranty, returns, delivery handling, a substantive review. Free shipping or a bundled extra instead of a percentage. No countdown theatrics.
• Goal: full-price recovery. This branch deserves its own article’s worth of detail — see how to treat high-value abandoned carts differently.

Illustrative example: same store, same hour, two carts

A kitchenware store sees two abandonments at 19:00. Cart one: a €22 silicone spatula set. Cart two: a €280 knife set. The spatula shopper gets a one-line reminder at 20:00 and nothing after a second nudge — total cost of recovery attempt: effectively zero. The knife-set shopper gets the reminder, then next evening an email about the 25-year blade guarantee and free sharpening, then on day three a personal-feeling note: “Happy to answer anything about the set — just reply.” If she buys, the store kept the full €280; a blanket 15% code would have cost €42 for the same outcome. Numbers invented, mechanism real.

What to measure per branch

• Revenue per recipient, per branch — the whole point of splitting is that this number can be optimized separately.
• Recovery rate per branch — expect the low branch to look worst; that’s fine, it’s also the cheapest.
• Discount cost per recovered order — should trend toward zero on the high branch.
• Average recovered order value — if it’s drifting down, your flow is recovering the easy carts and losing the ones that pay the rent. Make sure you’re attributing all of this cleanly — measuring abandoned cart revenue correctly shows how.

Setting the split up in Omnisend

I came to behavioral automation reluctantly — what pushed me was watching ad fatigue set in on my own stores while margins got squeezed from both sides. Value-based branching was one of the first changes that paid for itself, because it protects margin instead of spending it. In Omnisend, you add a conditional split on cart value right after the abandonment trigger and build each branch’s emails and delays separately; SMS steps and unique codes can live in one branch without touching the others. The requirement people underestimate: your cart totals must sync correctly (currency, taxes, discounts applied), or the split sorts shoppers into the wrong branches and you’re personalizing on noise.

Your next step

Export last quarter’s abandoned carts and look at the value distribution. Find the line above which your top fifth of carts sits — that single number is your first threshold, and adding one conditional split at that line is a one-evening change with margin you can measure by next month.