Most brands start thinking about BFCM in September. By then it's too late to learn anything. You're locked into whatever offer feels right in the moment, running it live in front of your biggest audience of the year, and hoping the math works out.
That's not a strategy. That's a guess, with your Q4 margin riding on it.
Mid-summer is different. Traffic is normal. Stakes are low. Nobody's watching your AOV dip for a week while you figure out what actually converts. This is the window to test, not October.
The Offer Test Nobody Talks About
We ran this exact experiment with a client, Optimize Minerals. Same product, same audience, same channel. We tested six offer formats against each other:
$10 off
$15 off
$20 off
15% off
20% off
Free gift with purchase
One of the six beat the other five by 300%. Not 30%. Three hundred percent.
I'm not going to tell you which one won, and that's not a tease. That's the whole point of this article. If I told you, you'd go copy it, and you'd probably get it wrong.
Here's why. The winning offer format isn't universal. It depends on your price point, your margin, and how your specific audience does math in their head. A flat $10 off reads as generous at a $40 AOV and reads like an insult at a $150 AOV. A percentage off feels bigger at low price points, and can actually underperform a flat dollar amount once you cross a certain threshold, because the dollar savings shrinks even as the percentage stays fixed. A free gift with purchase can beat both discounts entirely if your margin lets you attach something with high perceived value and low COGS.
Optimize Minerals' winning offer works because of their price point, their margin structure, and their audience. Yours won't be identical. That's exactly why you need your own version of this test before BFCM, not theirs.
Offers Aren't the Only Lever
Discount format is the one everyone obsesses over, but it's not the only thing worth testing this summer.
Merchandising. How you present the product matters almost as much as the price attached to it. Test a hero single-SKU page against a multi-product collection and see which one actually moves units.
Bundles. Pairing complementary SKUs at a fixed bundle price can lift AOV without touching your discount depth at all. Worth testing against your best single-item offer, not just assuming it wins.
Multi-packs. For consumable or repeat-purchase categories, a buy-more-save-more structure can outperform a flat percentage off, because you're selling volume instead of margin.
All three are testable right now, on a normal week, before BFCM traffic and stakes are both maxed out.
How to Actually Run This
Pick one variable at a time. Offer format, or bundle, or merchandising layout. Don't test five things in one send.
Split your list into even segments, ideally through Klaviyo or your ESP's native A/B tooling, and keep the sample size big enough to trust the result.
Run it on a normal week. Don't wait for a promo period. You want clean signal, not noise from a sale.
Track conversion rate and AOV together, not just one. An offer that lifts conversion but tanks AOV isn't automatically the winner.
Give yourself 4 to 6 weeks of runway before BFCM to run two or three rounds if the first test is inconclusive.
One offer format beat the other five by 300%. Same product. Same audience. Same channel. The only variable was the offer itself.
Final Thought
You don't need to guess your way into BFCM. The brands that walk into November already knowing their answer aren't smarter than you. They just ran the test in July instead of assuming in October.
Whatever you test, test it now, while a bad week costs you nothing. In November, a bad week costs you the quarter.
See you next week.
