A free gift is the cheapest way to turn a one-time buyer into a subscriber. It costs you once, on the first box, while the subscription renews at full price. The hard part was never the economics — it is getting anyone to choose the plan in the first place.
Why the gift beats a bigger discount
The instinct is to discount the subscription. It is the wrong lever, and the maths is not close.
A recurring 15% off costs you margin on every delivery, forever. A $10 gift costs you once. On a subscriber who stays twelve months at $40 a box, the recurring discount costs $72 in margin; the gift costs $10 — and the gift usually converts better, because a free product reads as a bigger deal than a percentage.
There is a second, subtler advantage. A discount trains the customer to value the subscription at the discounted price. When it eventually ends, the renewal reads as a price increase. A gift has no such tail: the first box was special, the rest are normal, and nothing about the arrangement feels like it got worse.
The part that actually blocks conversion
Most subscription programs do not fail on the offer. They fail on the interaction.
The customer has to find the subscribe toggle on the product page, understand the delivery interval, pick one, and only then discover whether the promised gift attached. Every one of those is a place to give up — and on mobile, where most of the traffic is, the toggle is often below the fold and visually indistinguishable from the one-time option.
A link removes the whole sequence. The plan is already selected, the gift is already in the cart, and the first screen the customer sees is the one that says what they will pay.
Choosing the gift
Three constraints, in order:
- It has to cost you materially less than it reads. A trial size of an adjacent SKU is close to ideal — low COGS, high perceived value.
- It must not cannibalise a future order. Gifting a full-size unit of the product they are subscribing to just delays the next delivery.
- It should introduce something. The best gifts are a second product the customer has not tried. A meaningful share of them become a second subscription later, which makes the gift an acquisition cost for two products rather than one.
Avoid branded merchandise. It is the classic mistake: it costs real money, it has near-zero perceived value to someone who is not yet a fan, and it teaches the customer nothing about your catalogue.
Scoping it to the first box only
This is the mechanic that keeps the offer bounded. The gift applies to the first delivery and nothing after it, so:
- The acquisition cost is a known, one-time number rather than a compounding one.
- Renewals bill at the normal subscriber price, so lifetime value is unaffected.
- The offer cannot leak into the recurring stream if someone shares the link.
Practically, the gift is a line-level discount on the first delivery, which means it drops out of the subtotal cleanly and never touches a renewal invoice.
The economics, in detail
Worth working through properly, because the intuition that a discount is cheaper than a gift is wrong in most catalogues and right in a few.
Take a $40 box at 45% gross margin — $18 contribution per delivery. Two acquisition offers:
A recurring 15% discount. Costs $6 per delivery, forever. Contribution drops to $12. Over a twelve-month subscriber: $72 of margin.
A $10-retail gift on the first box. Costs you COGS, not retail — call it $4. Contribution on box one drops to $14, and every subsequent box is a full $18. Over twelve months: $4.
The gift is roughly eighteen times cheaper on a year-long subscriber, and the gap widens the longer they stay. Even on a customer who cancels after two boxes, the gift costs $4 against the discount’s $12.
When the discount does win: very short expected lifetimes and very low gross margins. If your median subscriber lasts two deliveries and your margin is 25%, the arithmetic narrows enough that the simpler offer may be worth it. But a two-delivery median is itself the problem to fix, and it is usually an interval problem.
The number to compute for your own store: gift COGS versus (discount rate × contribution × expected deliveries). If you do not know expected deliveries, you cannot price either offer, and that is the first thing to measure.
Why the gift converts better despite costing less
Counter-intuitive, and consistent across the merchants running it:
- A product is concrete. “A free jar of X” is a thing; “15% off” is arithmetic the customer has to do to know what it means.
- It has a reference price. A $10 gift is worth $10 regardless of order value. A percentage is worth whatever the customer happens to spend, which they do not know yet at the moment of deciding.
- It introduces something. The gift is a product sample with a job beyond the promotion — a share of recipients buy it later at full price, which no discount does.
- It does not reprice the subscription. The customer’s mental anchor stays your real price, so the first full-price renewal is not experienced as an increase.
Where this fits against your other retention work
A subscription gift is an acquisition offer for a retention mechanic, which makes it easy to misfile.
Against a loyalty program. Loyalty rewards customers for behaviour they already have. A subscription gift changes behaviour — it converts a manual repeat buyer into a subscriber, which is a different and generally more valuable outcome. Run both; do not treat one as a substitute.
Against a win-back discount. Win-back targets lapsed customers with a price cut. The subscription gift targets active ones with a structural change. If a customer has already lapsed, the gift is the wrong tool — they have not chosen to keep buying, so a plan is not the ask.
Against post-purchase upsell. These compete for the same moment. A post-purchase upsell converts more often and is worth less; the subscription offer converts less often and is worth far more. On a second or third order, the subscription offer should generally win the slot.
The sequencing that works: let a customer buy once, let them reorder manually, and make the subscription offer at that second or third purchase — the point at which they have demonstrated the repeat behaviour a plan formalises.
Where to send it
Ranked by how well they convert:
- Existing one-time buyers, especially repeat ones. They already like the product and have simply never chosen a plan. This is the highest-yield audience by a wide margin.
- Post-purchase, on the second or third order. The moment someone reorders manually is the moment a subscription obviously makes sense to them.
- A Klaviyo flow on reorder cadence. If someone buys every five weeks, a subscription is what they are already doing by hand.
- Cold traffic and ads. Works, but converts far below the above, because you are asking for a recurring commitment from someone who has not yet tried the product.
Picking the plan the link lands on
The gift gets the attention, but the delivery interval decides whether the subscriber survives to the second box. Get this wrong and the gift bought you one discounted order.
Match the interval to genuine consumption. Work out how long a unit actually lasts your median customer — from reorder gaps in your own order data, not from the label. Most brands set intervals shorter than real usage, because a shorter interval looks better in projected revenue. It produces cancellations instead: the second box arrives while the first is half full, the customer does the mental arithmetic on how much product is accumulating, and they cancel.
Offer fewer options, not more. A picker with six intervals is the thing the link exists to remove. Choose the one that fits most customers and put that on the link. Customers who want something different will change it in their account, and far more of them will subscribe at all.
Consider a longer first interval. If your data is ambiguous, err long. A customer who runs out slightly early will shorten the interval themselves — an action that signals engagement. A customer drowning in product cancels, and cancellation is terminal in a way that adjustment is not.
For multi-product subscriptions, align the intervals. Two products on different cadences generates two shipments, two charges and two chances to reconsider.
Making the gift feel like a gift
The mechanics can be right and the offer still land flat, because how the gift is presented changes what it is worth.
Name it, don’t discount it. “Free Debloat Gummies with your first box” reads differently from “$10 off”. Identical economics, materially different response — one is a product you are giving, the other is a number off a price. The discount framing also invites comparison shopping in a way the gift framing does not.
Show it in the cart before checkout. The gift should be visible as a line at $0, not applied invisibly. A customer who cannot see the gift does not believe in it, and a benefit they do not believe in does not motivate the subscribe decision.
Do not make them claim it. Any step between “subscribe” and “have the gift” costs conversions. The gift is in the cart when they arrive; nothing to add, no code, no checkbox.
Say what it is worth, once. A single reference to the retail value anchors the offer. Repeating it starts to feel like selling.
The surprise variant
An alternative worth testing: do not advertise the gift at all, and include it in the first box unannounced with a note.
You lose the conversion lift on the subscribe decision — which is the main reason to run the offer — but you gain something on retention, because an unexpected gift lands harder than a promised one and arrives at the moment the customer is deciding whether the subscription was a good idea. For stores whose problem is second-box churn rather than first-box conversion, this is often the better use of the same $10.
You can run both: advertise a gift on the link to win the subscribe, and include a second small unannounced item in the box.
Handling the second box
Everything above is about acquisition. The gift’s return is decided after it, and the first renewal is where subscriptions die.
Warn before charging. A pre-renewal email a few days ahead costs a small number of cancellations and prevents chargebacks and angry support tickets, which are more expensive. Customers who cancel on a warning were going to cancel; customers surprised by a charge tell other people.
Make the second box feel considered. If the first box had a gift and the second is identical minus the gift, the experience is of something being taken away. A note, a sample, a small variation — anything that makes box two a continuation rather than a downgrade.
Give them a pause option, prominently. Pause is the single most effective churn intervention available, because most cancellations are about accumulation rather than dissatisfaction. A customer who pauses for a month usually comes back; a customer who cancels usually does not. If the cancel flow does not offer pause first, it is manufacturing churn.
Watch the cancel reasons. If “too much product” dominates, the interval is wrong and no gift will fix it. If “too expensive” dominates, the subscriber price is wrong. If it is “just trying it”, the gift attracted trial-seekers and the offer needs to be less generous, not more.
Common mistakes
Gifting a full-size unit of the subscribed product. It delays the next delivery, which is the opposite of what a subscription gift should do. Gift something adjacent.
Advertising the gift as a permanent benefit. “Gift with every box” is a recurring cost that compounds exactly like the discount you avoided, and it is very hard to withdraw once established.
Running it as a site-wide banner. The offer’s economics depend on it reaching people who would otherwise buy once. Putting it on the storefront means every existing subscriber sees a better deal than the one they are on, and some will cancel and re-subscribe to get it.
Measuring at the link. Conversion on the link will look excellent because you sent it to people who already buy from you. The number that matters is second-delivery retention, and it arrives a month later.
Setting the interval from the projection. Covered above, and worth repeating because it is the single most common cause of subscription programs that acquire well and retain badly.
Building the link
The mechanics, concretely, so the offer behaves the way the economics assume.
Pre-select the selling plan. The link encodes the plan and interval, so the customer lands with the subscription already chosen rather than a toggle to find. This is the whole conversion argument — every store with subscriptions has a subscribe control on the product page, and the drop-off happens there.
Add the gift as a first-delivery line at zero. Scoped to delivery one only. It should appear as a visible line rather than a silent price adjustment, and it must not attach to renewals — a gift that recurs is the compounding cost the whole approach exists to avoid.
Zero the shipping if you can afford it. Shipping is the most-cited abandonment reason, and on a subscription the customer is evaluating a recurring commitment. Removing one variable from that decision is usually worth more than the freight.
Set the offer to expire. Not for scarcity theatre — for control. An offer link with no expiry is a permanent alternate price for your subscription, indefinitely forwardable.
Prefill what you know. Sending to existing customers means you have their email and often their address. Every field pre-filled is friction removed at exactly the moment they are reconsidering.
Testing it before it goes out
Three checks that catch the common failures:
- Complete a real purchase. Confirm the plan attached, the gift is at zero, and the renewal amount shown excludes the gift. A gift that silently attaches to renewals is the expensive failure.
- Check the renewal invoice. In your subscription app, look at what box two will charge. This is where scoping errors surface.
- Open the link logged out, on a phone. Most recipients will. Confirm nothing requires an account and nothing overflows.
Segmenting the send
Not everyone should get the same offer, and the differences are worth the extra flows.
Repeat buyers who never subscribed. The core audience — they have demonstrated the behaviour a subscription formalises. Standard offer, and expect your best conversion here by a wide margin.
One-time buyers, 30-60 days in. Approaching a natural reorder point. Worth a slightly stronger offer, because you are asking them to commit on less evidence.
Lapsed subscribers. A separate case. They tried it and left, so the gift is not the issue — find out whether they cancelled on interval, price or product, and address that instead. Sending the acquisition offer to someone who already rejected the plan reads as tone-deaf.
Existing subscribers. Exclude them. This is the segmentation mistake that costs real money: an active subscriber who sees a better offer than the one they are on will cancel and re-subscribe to get it, and you have paid a gift to lose a month.
That last exclusion is worth building into the flow rather than remembering each time.
What to measure
Not conversion rate on the link. The numbers that matter are downstream:
- Second-delivery retention. If people cancel before the first renewal, the gift bought you a discounted one-time order, not a subscriber.
- Gift cost as a share of first-year value. This is the real acquisition cost, and it should be comfortably below what you pay for a subscriber through ads.
- Attach rate of the gifted product. If a good share of gifted customers later buy the gift SKU at full price, the gift is doing double duty.
If second-delivery retention is weak, the problem is rarely the gift — it is usually the interval. People cancel because the second box arrived before they finished the first.
When not to run this offer
The pattern is not universal, and three cases where it underperforms:
Consumables with unpredictable usage. If consumption varies wildly between customers — seasonal products, occasion-driven purchases — no single interval fits and churn will be high regardless of the gift. Fix the interval problem, or accept that subscriptions are the wrong model for that SKU.
Very high-consideration products. Where the purchase decision takes weeks of research, a gift does not shorten it. The customer is not hesitating over $10.
Catalogues with nothing adjacent to gift. A single-product store has only two options: gift the product itself, which delays the next delivery, or gift branded merchandise, which has near-zero perceived value. Neither works. Broaden the catalogue before running the play.
Further reading
- Shopify discount codes playbook — why the recurring-percentage instinct is usually the expensive one.
- VIP offers playbook — gifts and private pricing for customers who already buy repeatedly.
- Abandoned cart recovery playbook — the other place a pre-built cart does the heavy lifting.
- Checkout Links for Klaviyo — sending the offer to a repeat-buyer segment.
- Checkout Links for abandoned cart — the link layer itself.
