52% Repurchase Rate Among Gift Recipients — Validating Your Free-Gift Strategy with Data

"Do free gifts actually help drive repurchases, or are they just a cost?"

Plenty of brands run a free-gift strategy: "Free mini pouch with orders over KRW 30K," "Sample kit included with your first order." Yet almost no brand checks with data whether those gifts actually contribute to repurchases. For most, the reality is: "We're giving them out anyway, so who knows if it works — let's just keep doing it."

The problem is that a gift is a fixed cost you pay on every single order. Put a gift costing KRW 2,000 into 10,000 orders a month and that's KRW 20 million a month — not a small amount to burn without knowing whether it works. This article covers how to ① choose the right metrics for gift effectiveness, ② measure it accurately while avoiding the common traps, and ③ optimize it for ROI.

The 3 effects free gifts create

1. Higher AOV on orders that include a gift

Orders with a gift:      AOV KRW 58,000  |  3.2 items per cart 
Orders without a gift:   AOV KRW 38,000  |  1.8 items per cart

The AOV of orders with a gift is 1.5x higher. A "free gift with orders over KRW N" condition works as an upselling trigger that makes shoppers fill their carts up to the threshold. 👉 How to see it in Laplace: split orders by gift included (is_gift / gift_item_set) and compare AOV and items per cart.

2. Repurchase rate of customers who received a gift

Gift recipients:       Repurchase rate 52% 
Non-recipients:        Repurchase rate 31%

The reciprocity effect of "I got something for free" kicks in. 👉 How to see it in Laplace: use lifetime_num_of_purchases (cumulative number of purchases) and customer_order_type (first purchase / repurchase) to compare repurchase conversion between the recipient and non-recipient cohorts.

⚠️ Why you shouldn't take this number at face value (the most important insight) Gifts are usually given for "purchases over a certain amount." That means gift recipients are likely heavy users who already spend more. So you can't tell whether the higher repurchase rate comes from the gift or from the fact that they were loyal customers to begin with — there may be correlation, but not necessarily causation. To see the real effect, you need to control for these conditions: • Compare recipients vs. non-recipients within the same spend bracket (removes spend bias) • Isolate and compare first-time buyers only (removes heavy-user bias) • Where possible, validate with an A/B test split by period (gift on/off). Only if a gap remains after these controls can you call it a true "gift effect."

3. Different gifts, different results — "what you give" is what matters

Gift A (mini pouch):   Repurchase rate 55%  |  COGS KRW 2,000 
Gift B (sticker set):  Repurchase rate 33%  |  COGS KRW 500 
Gift C (sample kit):   Repurchase rate 61%  |  COGS KRW 3,500

The sample kit performs best — it gets customers to experience other products, and that experience leads to the next purchase (cross-selling). Conversely, gifts unrelated to the brand experience, like stickers, are cheap but contribute little to repurchases. Measured by "repurchase contribution relative to cost," B is essentially a throwaway expense. 👉 How to see it in Laplace: pull the top 5 most frequently given gifts × repurchase rate of each and separate the "cheap and effective" from the "expensive and ineffective."

How to calculate gift ROI (example)

The naive calculation (easy to get wrong): "Recipients 52% vs. non-recipients 31% → 21pp higher, it's a clear win!" → This carries the heavy-user bias described above straight into the result, so the effect is overestimated.

The controlled calculation (accurate): Suppose that after re-running it for the same spend bracket and first-time buyers only, the net lift is 8pp —

Gifts per month:            5,000 
Gift COGS:                  KRW 2,000  → Monthly cost KRW 10 million
Net incremental repurchases: 5,000 × 8pp = 400
Margin per repurchase:      KRW 30,000  → Additional margin KRW 12 million
ROI = KRW 12 million ÷ KRW 10 million = 1.2x

An effect that the naive calculation would have inflated to 3–4x comes out at 1.2x (a narrow profit) after controls. The prescription here is clear → lower COGS (switch to the cheaper A or the sample kit C) or raise the threshold to amplify the AOV effect. You can only make decisions like this when you look at the numbers.

4 steps to designing a gift strategy with data

1. Get the comparison group right first: compare AOV and repurchase rate for orders with and without gifts, controlling for the same spend bracket and the same customer group

2. ROI per gift: (net incremental repurchases × repurchase margin) ÷ total gift COGS — judged on margin, not on repurchase rate

3. Optimal threshold: set the gift threshold at 1.2–1.5x your current average AOV to drive upsells (too high and shoppers drop off; too low and you're giving gifts away for free)

4. Item replacement loop: review the repurchase rate of your top 5 most-given gifts every quarter and replace "gifts that don't work → experiential gifts (samples/miniatures)"

Summary

A free gift isn't "a cost you just give away" — it's a measurable investment. It comes down to three things.

• AOV moves with how you design the threshold; repurchases move with which items you choose

• A repurchase-rate gap only reflects the true effect once you strip out bias (same spend bracket, first-time buyers, A/B)

• Make decisions based on margin-based ROI, not repurchase rate

Run these three with data, and you can generate more repurchases from the same gift budget.

Gift effectiveness — from AOV on gift orders to recipient repurchase rate and ROI per gift — check it all yourself with Laplace data.