Discount-Led AOV Erosion: How Promo Cadence Quietly Breaks LTV:CAC
When AOV declines month over month and repeat orders cluster around promos, the discount calendar — not acquisition — is usually what broke LTV:CAC. Here's how to prove it and fix it.
Quick answer
If your AOV has fallen for 3+ consecutive months while repeat-purchase rate stayed flat, the discount calendar is training customers to wait. Split your last 12 cohorts into full-price vs promo-acquired and compare 90-day contribution margin per customer — a gap of 25%+ confirms discount-led AOV erosion as the cause of LTV:CAC slipping below 1.
Discount-led AOV erosion
A pattern where frequent promotions train customers to wait for deals, pushing AOV and contribution margin per order down until LTV:CAC breaks.
Discount-led AOV erosion is a slow-moving failure mode specific to stores that run a predictable promo cadence — monthly sales, recurring 15% codes, sitewide events every 3-4 weeks. Customers learn the rhythm and time their orders to it, which compresses three things at once: average order value, contribution margin per order, and the share of revenue that arrives outside promo windows. The acquisition side of LTV:CAC often looks fine; the lifetime value side hollows out from within. Because each individual discount feels small, the erosion only becomes visible at the cohort level, usually 6-9 months after the cadence started.
This page is a focused branch of the broader LTV:CAC Ratio Below 1 diagnostic playbook. Use it when acquisition cost looks stable but the LTV side is the one drifting down.
Why it happens: the mechanism
Customers form reference prices from what they've seen recently. Once they've paid £42 for an item twice in a row during a promo, £56 at full price feels like a premium they shouldn't accept.
Predictable cadence is the accelerant. If a 15% code appears every third Thursday, repeat buyers don't churn — they just delay. Their second order still happens, but at lower margin and on your worst day of the month.
The hidden cost is not the discount
A 15% discount feels like a 15% margin hit. It's almost never just that. On a typical apparel order with 55% gross margin and £6 fulfilment, a 15% sitewide promo on a £60 AOV order erases roughly 40% of contribution margin — and that's before you account for the full-price orders the promo cannibalised.
How to detect it: the cohort split that proves it
Tag every customer by their first-order context: full price, soft promo (under 15%), or hard promo (15%+, BFCM, sitewide events). Then track 90-day contribution margin per customer for each cohort across the last 12 months.
If hard-promo cohorts return at a similar rate to full-price cohorts but generate 25-45% less contribution margin, you have discount-led erosion. If they return less often AND spend less, you have a quality-of-acquisition problem on top — the promo is also pulling in the wrong customer.
Typical 90-day cohort behaviour on Shopify apparel and beauty stores, by first-order acquisition context
| First-order context | AOV (order 1) | Repeat rate 90d | Contribution margin per customer 90d |
|---|---|---|---|
| Full price | £62 | 28% | £24 |
| Soft promo (≤15%) | £54 | 31% | £17 |
| Hard promo (>15% or BFCM) | £48 | 26% | £11 |
| Email-code repeat buyers | £51 | 44% | £14 |
AOV trend by cohort type — 12 months of monthly cohorts
Full-price acquired
Promo-acquired
How to fix it without killing volume
The instinct is to cut promos cold turkey. Don't — you'll punch a hole in this quarter's revenue and the board will reverse you in six weeks. Erosion took months to build; the unwind should take months too.
Break the cadence first. If you promo every third Thursday, your customers know it. Move to irregular intervals — 19 days, then 34, then 23 — so reference-price formation breaks down. This alone typically lifts AOV 4-8% within two cycles.
Replace sitewide codes with bundle thresholds. A 'spend £75, save £10' offer protects margin floor and shifts AOV upward, where a sitewide 15% code does the opposite. On Shopify, the native discount engine handles this without an app.
Shift discount budget toward post-purchase and reactivation rather than acquisition. A code in the order-2 email targeting day 35 lifts repeat rate without conditioning the first-time buyer. Klaviyo flows make this routine.
Finally, segment your promo audience. Customers who have only ever bought on promo should see different offers (loyalty points, free shipping) than full-price buyers. Sending the same 15% code to both is the cheapest way to keep eroding.
Experiment ideas to run this quarter
Test 1 — cadence break. Hold one monthly promo. Measure full-price revenue in the 4 weeks following vs the 4 weeks prior. Primary metric: full-price AOV. Guardrail: total revenue within 5% of forecast.
Test 2 — threshold swap. Replace your next sitewide 15% promo with a 'spend £X, save £Y' bundle where the implied discount is similar. Primary metric: contribution margin per order. Guardrail: order count.
What good recovery looks like
Six months into a disciplined unwind, expect: full-price share of revenue up 8-15 points, blended AOV up 5-10%, and LTV:CAC back above 1.5 — without acquisition cost changing. The fix lives almost entirely on the LTV side of the ratio.
Edge cases worth knowing
Beauty subscription SKUs behave differently — promo-acquired customers there often convert to subscription at similar rates, so the LTV gap closes by month 4. Run the cohort split with a 180-day window before declaring erosion.
Stores under £1M annual revenue can rarely afford a full cadence break. For them, the threshold swap is the higher-leverage move; cadence discipline becomes feasible once paid traffic stabilises.
Frequently asked questions
It's usually invisible for the first 3-4 months of a new cadence and obvious by month 7-9. The fastest signal is full-price share of weekly revenue — track it weekly and you'll see the trend before AOV moves materially.
Discount addiction describes the customer behaviour; discount-led AOV erosion describes the financial outcome. The first causes the second. Treating them as one diagnosis is fine operationally, but the metrics you watch are different — behavioural metrics for the cause, margin metrics for the effect.
Yes — product-mix shift toward lower-priced SKUs, a new paid channel pulling lower-intent traffic, or removal of a high-AOV bundle can all do it. The cohort split by acquisition context separates these. If full-price cohorts also show declining AOV, the cause is upstream of the discount calendar.
Almost never. Discounts have a legitimate role for new-customer acquisition, inventory clearance, and reactivation. The problem is predictable sitewide cadence aimed at existing buyers. Keep targeted discounts, kill predictable ones.
Discount-led erosion is one of the three most common causes of LTV:CAC dropping under 1 — alongside rising paid CAC and falling repeat rate. It's the hardest to see because acquisition metrics often look healthy. The parent LTV:CAC diagnostic playbook walks through all three causes in sequence.
There's no universal number, but the rule is: customers should not be able to predict the next promo within a 2-week window. For most apparel and beauty stores that means 4-7 promotional events per year, irregularly spaced, with reasons attached (launch, season, clearance) rather than calendar slots.
No — a one-time welcome code on first email signup is acquisition, not cadence. It only becomes a problem if you re-send it monthly to non-converters, which trains them to wait for the next one. Cap welcome-code exposure at two sends.
Revenue minus COGS minus fulfilment minus payment fees minus any discount applied, aggregated across all orders in the cohort window. Pull it from your Shopify order export joined to COGS by SKU. Don't use blended gross margin — promo orders skew it.
Done gradually, no — the threshold-swap and post-purchase-shift moves are revenue-neutral in quarter one and margin-positive from quarter two. The cadence break is the riskier one; pilot it on a single month before committing.
Importing your historical GA4 and order data lets the cohort split run on day one, without waiting for new data to accumulate. The AI hypothesis layer flags the AOV-by-cohort divergence automatically when it appears in your store, rather than waiting for someone to query it.
Get an AI expert review of your site
Paste your URL — Metricuno's AI runs the same heuristic checks a senior CRO consultant would, scoring your page and prioritising the fixes that'll move conversion fastest.