Adjusting Annualized RPV Projections For Q1 Returns And Refund Lag
Q4-heavy revenue drags a return-and-refund tail into Q1 that can erase 4–8% of a gross RPV win. Here's how to net returns out of your seasonality index before you annualize.
Quick answer
Before you annualize a Q3-tested RPV win, subtract a category-typical return rate from each seasonality index cell — weighted heaviest on Q4 orders that refund in Q1. For most apparel and beauty stores that means shaving 4–8% off the gross annualized number so finance sees a net figure, not a GA4 gross-revenue mirage.
Adjusting annualized RPV projections for Q1 returns and refund lag
Netting expected returns and refund timing out of a seasonality-weighted RPV projection so the annualized number matches booked revenue.
When you annualize a Q3-tested RPV win using a seasonality index, GA4 hands you gross revenue by month. But Q4 orders — the ones that dominate the index — return at a much higher rate than the annual average, and most of those refunds hit the books in January and February. If you don't net them out, your projection double-counts revenue that finance will later reverse.
The adjustment is small in math and large in credibility: subtract a category-typical return rate from the Q4 and Q1 index cells before you multiply through, so the annualized RPV lift you present is a net-of-returns figure.
This page assumes you already have a Q3 test result and a seasonality index — the parent concept covers how to build one. The problem here is narrower: the index itself is biased upward by gross Q4 revenue, and the fix is a single netting step you apply before annualizing.
Why the Q4-into-Q1 refund tail breaks the projection
GA4 records purchase revenue at checkout. Refunds — when they're tracked at all — land weeks later as negative events, or more often, don't flow back into the analytics warehouse at all. Your seasonality index is built on the gross side of that ledger.
For an apparel store, Q4 might represent 38% of annual gross revenue. But apparel return rates spike in Q4 gifting orders — often 25–35% versus a 15–20% annual baseline. Most of those returns process in the first six weeks of Q1, meaning your Q4 index cell is inflated and your Q1 cell is masking a refund crater.
The double-hit
Q4 is overweighted (gross revenue includes items that will be returned) AND Q1 is overweighted (gross revenue hides the refund outflow). If your RPV win came from a Q4-heavy AOV lift — bundling, gift-with-purchase, size upsell — the return rate on the incremental revenue is likely higher than baseline, not lower.
How to detect the size of the problem in your data
Pull net revenue from Shopify (or your OMS) by month for the last full year, alongside GA4 gross purchase revenue. The gap between them, by month, is your refund lag signature. In apparel and beauty, expect January net to be 10–20% below January gross.
If you can't get monthly net revenue, use the category benchmarks below as a starting subtraction. Then reconcile against your finance team's actual return rate once — the delta tells you whether your category is higher or lower than the industry mean.
Typical return rates by category and quarter (online retail)
| Category | Annual avg return rate | Q4 return rate | Q1 refund concentration |
|---|---|---|---|
| Apparel (fashion) | 20–30% | 28–38% | 55–65% of annual refunds |
| Footwear | 18–25% | 22–30% | 50–60% of annual refunds |
| Beauty & skincare | 5–10% | 8–14% | 45–55% of annual refunds |
| Home & decor | 8–15% | 12–20% | 50–60% of annual refunds |
| Electronics | 10–18% | 15–22% | 45–55% of annual refunds |
| Health supplements | 3–7% | 5–9% | 40–50% of annual refunds |
How to net returns out of the seasonality index
Take your monthly index cells and multiply each by (1 − quarterly return rate). For an apparel store, that means Q4 cells get multiplied by roughly 0.70, Q1 cells by 0.82 (baseline return rate plus the inflow from Q4 refunds), and Q2/Q3 by 0.80. Then re-normalize the index so it still sums to 12.
Apply the Q3 test's RPV lift against the re-normalized index. The resulting annualized number is net-of-returns and will reconcile against a full-year P&L within a few percent. That's the version you take to a forecasting meeting.
Concrete example
A women's apparel Shopify store tests a size-guide redesign in Q3 and measures a +€0.42 RPV lift (from €12.10 to €12.52, +3.5%). Gross annualization on 4.2M annual sessions = €176k incremental. After netting Q4 index by 30% return rate and Q1 by 18%, the re-normalized figure is €162k — an 8% haircut. Finance signs off on €162k, not €176k.
Experiment ideas to tighten the adjustment
The size-guide, fit-quiz, and PDP-imagery tests that lift Q4 AOV also tend to reduce Q4 return rate — because the incremental revenue comes from better-fitted purchases, not just bigger baskets. Instrument return rate as a secondary metric on any test that ships before September so you have real, not assumed, netting factors by January.
For beauty and skincare, shade-finder and sample-first tests are the equivalent. Measure return rate by variant for at least 60 days post-purchase — the refund window closes slowly, and cutting off measurement too early flatters the lift.
Frequently asked questions
You can, and it's the cleaner path when you have clean OMS data. Most teams end up with GA4 because it's session-linked and the RPV test itself was measured there. If you build the index from net revenue, skip the netting step — you're already there.
The parent concept — applying a seasonality index to annualize a Q3-tested RPV win — assumes gross revenue is a fair basis for weighting months. This page adds the netting step for stores where Q4 gross revenue includes a material return tail that lands in Q1.
Start with the category benchmark in the table above, then reconcile against a single finance close. For apparel assume 25% annual and 32% Q4; for beauty assume 8% annual and 12% Q4. Update once you have your own numbers.
Yes, and it matters more. A Q4-tested lift measured on gross revenue will be biased upward by the very orders most likely to refund. Net Q4 return rate off the observed lift before you annualize, then apply the seasonality adjustment on top.
For most online retail, 85–90% of returns close within 60 days. The long tail matters for high-AOV categories like furniture and consumer electronics where the return window can run 90–120 days. Use a 90-day window there and accept a small residual.
Only if they're material — usually under 1% for established stores. Bundle them into a single 'refund and adjustment' factor rather than modelling separately. The signal-to-noise doesn't justify a dedicated cell.
That's common for stores with heavy gifting revenue in December — brands where 40%+ of Q4 revenue is a gift purchase. Increase the Q1 netting factor by 3–5 percentage points and validate against last January's actuals before you sign off on the projection.
Yes. Conversion-rate lifts refund at roughly the baseline rate. AOV lifts — especially from bundling or upsell — often refund at higher rates because return rate scales with basket size in most categories. Apply a modest extra haircut (2–3 points) on AOV-driven lifts.
Lead with the net number and show the gross as a footnote with the netting factor explicit. Finance will trust the net figure; marketing will want to see the gross for channel-level ROAS reasoning. Both are legitimate, just labelled clearly.
Once a year, after the January close. Return behaviour is sticky enough that quarterly updates add noise, not signal. Reconcile your assumed netting factor against actual returns from the prior year and adjust before you plan the next test roadmap.
Test ideas before you ship them
Run unlimited A/B tests, attach hypotheses to outcomes, and build a searchable archive of what works — and what doesn't.