Gifting ROI is real. It just doesn't look like paid ROI, and the most common mistake operators make is pretending it does. You send 80 PR packages, three creators post, you pull a 7-day attribution report, see $1,400 in revenue against $4,200 in COGS plus shipping, and conclude that seeding doesn't work. Then you cancel the program, and two months later your branded search is down 22% and nobody can quite explain why. The same brands often quietly switch to paid sponsorships at this point, which is its own trap I picked at in the gifting vs paid breakdown.
I want to lay out a way to score a seeding program that a co-founder or boss will accept, without forcing the numbers to behave like a Meta ads dashboard. It involves admitting up front that you can't see most of the lift, then building a scoreboard that captures what you can see plus some leading indicators for what you can't.
For context, I work on Seed, a Shopify gifting app, so I see how brands account for these sends. The brands that stick with seeding over a year are the ones who stopped trying to attribute it like performance media. The ones who churned out of it almost always had a finance person asking for a 30-day ROAS number that gifting cannot produce.
The true cost per send
Before any ROI conversation, get the cost right. Three components, and brands routinely miss the third.
COGS is the easy one. If the product cost you $11 landed and you ship a $58 retail item, your cost is $11, not $58. Don't put the gift on your P&L at retail. I have watched founders panic about "$30,000 of gifting spend" that was actually $5,800 of COGS.
Shipping is the second. Domestic ground for a small item is $5 to $9 with a label, plus the mailer, plus tissue and a thank-you card if you do that. Call it $9 to $13 all-in for most beauty and apparel brands. International doubles or triples this and is where seeding gets expensive fast.
Coordinator time is the one nobody counts. Vetting a creator, getting them to fill out an address, building the draft order, sending the tracking, following up about a post. At a manual DM pace, that's 15 to 25 minutes per send. If your coordinator costs $28/hour loaded, that's $7 to $12 of labor per package. For a 50-package month you've quietly spent $500 on coordination alone.
Add it up. A typical send for a Shopify beauty or apparel brand costs $25 to $45 fully loaded. Not the $80 retail figure your accountant might use. Not the $11 COGS your gut wants to use. Somewhere in between.
What you can attribute directly
Three signals are clean enough to put in a deck.
Unique discount codes per creator. Give each creator a code like MAYA15 with their 15% off. When it gets redeemed, you know the order came from their audience. This is the cleanest signal you'll get. The catch: redemption rates on creator codes are usually 0.5% to 3% of their follower count, and a lot of buying audiences just go straight to the site without using a code at all.
UTM links in bio or swipe-ups. Same idea, slightly noisier because people share links and forget where they came from. GA4 will give you sessions and a checkout funnel by source if you tag properly.
Branded search lift. This is the one most operators ignore and it's the most honest. Pull weekly branded query impressions from Google Search Console for the 12 weeks before the campaign and the 12 weeks after. If your brand keeps showing up in feeds, people start typing your name into Google. A 15% to 40% rise in branded impressions after a seeding push is a real signal, even if you can't tie any single creator to it.
What you can't attribute
Most of the lift. I mean that literally. The view-through effect of someone watching a 14-second Reel, scrolling past, then buying a week later when they see a retargeting ad: invisible to your reports, but it's how the system actually works. Comments where people tag a friend ("@jess this is the brand I told you about") almost never convert in a window your attribution model can see. The "I keep seeing this brand everywhere" effect compounds over months and shows up as cheaper paid ads, not as gifted-attributed revenue.
If you only count code redemptions, you will conclude that gifting loses money on roughly every campaign. That conclusion is wrong, but you cannot disprove it with the data you have. You have to accept the gap or you'll cancel a program that's working.
A simple per-campaign scorecard
Pick five metrics, score every campaign on all five, and stop arguing about a single ROAS number.
- Post rate. Of creators who received product, what percent posted within 30 days. Healthy is 35% to 55% for cold seeding, 60%+ for warm. Targeting is usually the lever here, and how you build the creator list in the first place matters more than the pitch.
- Content quality. Subjective but consistent: rate each piece 1 to 5 on whether you'd repost it. Track the average.
- Engagement on posts. Total likes plus comments across all creator posts in the campaign. Compare to last campaign, not to a benchmark.
- Code redemptions and UTM revenue. Direct attributed revenue. Expect this to be small. That's fine.
- Branded search lift 4 weeks after the send window closes. Pull from GSC.
Five numbers. Put them in a sheet. Compare campaign to campaign. Don't reduce to one number, because the reduction always lies in one direction or the other.
Time horizon
Gifting compounds over 4 to 12 weeks. Not 48 hours. A creator posts on Tuesday. Their audience sees it, screenshots it, mentions you in a group chat, sees you in two more places, finally clicks an ad three weeks later. If you close the books on a campaign after 7 days, you are measuring the wrong window. Wait at least 6 weeks before you call a campaign a result, and 12 if you can stand to.
When to call a creator a hit vs a miss
A hit posted within 30 days, the content was a 3 out of 5 or better, and engagement was at least at their account average. A miss didn't post, posted late with low-effort content, or posted content you'd be embarrassed to share. Track both. The brands I see succeeding have a hit rate near 40% and a clear rule for not re-seeding misses for at least six months.
When to graduate a creator from gifting to paid
A creator earns a paid slot after two seeded posts that hit your bar, when their code does at least 8 redemptions, or when they DM asking about a deeper partnership. Paying a creator who has never posted for you organically is gambling. Paying a creator who has already posted twice and moved units is just a normal media buy.
FAQ
Should I include retail value of the product as the "cost" of gifting? No. Use COGS. Retail value is a marketing number, not an accounting one.
What if a creator wants a flat fee on top of free product? That's a paid partnership now, not gifting. Move it to a different line item and judge it as media.
How long until I should see branded search lift? 4 to 8 weeks after the send window closes, assuming you sent to at least 30 creators with audiences over 5,000. Smaller programs are hard to detect against weekly noise.
What about UGC rights, does that count toward ROI? Yes, and you should price it. If you would have paid $200 to a UGC creator for a usable clip, every gifted post you can license is worth that. Tools like Seed let you bundle a UGC license into the gifting form so you don't have to chase it later.
How do I justify gifting to a CFO who only respects ROAS? Show the scorecard, show the branded search trend over 12 weeks, and show the blended CAC change in the months you ran heavy seeding versus the months you didn't. If those three move, the CFO will get it. If they don't, you have an honest answer too.