9 Types of Bundles to Boost Sales and Margin in E-commerce
11 min read · AstraLoop Studio
The bundle is probably the most underrated lever in Italian e-commerce. Everyone knows that "selling more products together pays off," few do it with any method. The result is a catalog full of improvised "kits" that don't convert, or aggressive discounts that raise the order value by a few euros while burning 12 points of margin.
Yet the numbers are clear: a well-built bundle lifts average order value (AOV) by 20-35%, and industry estimates suggest up to 30% of a mature e-commerce store's revenue can come from products sold as a package. The point isn't "do bundles": it's choosing the right type of bundle for the right objective (clear stock, protect margin, push a launch, raise perceived value) and building it with data instead of gut feeling.
In this guide we cover 9 bundle types, when to use each, and how to set them up. At the end you'll find the AstraLoop method for optimizing composition and price with AI, so you stop guessing which pairing works.
Before picking a type: the two questions that matter
A bundle isn't an offer: it's a pricing and product-mix decision. Before choosing which of the 9 types to use, answer two questions.
1. What's the economic goal? There is no "best bundle" in the abstract. There's the right bundle to clear unsold stock, the right one to raise margin on an already-large cart, the right one to get a new product tried. If you don't know what you want to achieve, you end up discounting products that would have sold anyway, destroying margin with zero incrementality.
2. Are the products complementary or substitutes? A bundle works when the products complete each other (the cream plus the serum, the shoe plus the technical sock) and the customer's valuation of each individual piece is of comparable order. That's the conclusion of the academic research by Bakos and Brynjolfsson on the economics of bundling: a bundle performs better when demand for the various products is weakly correlated and individual valuations are similar. In practice, avoid pairing your best-seller with a dead product hoping it will get dragged along: customers spot the forced pairing and don't buy.
Keep both questions in mind, and here are the types.
The 9 types of bundle (and when to use them)
1. Pure bundle
The products are sold only together, never separately. Think of a gift box that exists exclusively as a set.
It sounds like the simplest formula, but it's also the riskiest. A classic Harvard Business School study on Nintendo showed that pure bundling, by forcing a "take it or leave it" choice, made sales drop 20% compared to offering the same items both individually and as a package. Removing the freedom to buy just one piece pushes away anyone who only wanted that one.
When to use it: products designed natively as an indivisible set (a limited-edition collection, a gift box, a kit whose separate pieces wouldn't make sense on their own). Outside these cases, you almost always lose sales.
2. Mixed bundle
The customer can buy the products individually or together at a discounted price. It's the most common form and, in the vast majority of cases, the most profitable: you preserve freedom of choice while still pushing toward the larger package.
In that same Nintendo study, mixed bundling increased console sales by 100,000 units and game sales by over a million. The takeaway is simple: always keep the single-item option, then make the bundle irresistible with the right discount.
When to use it: it's the default. Unless you have a strong reason to do otherwise, your bundle should be mixed.
3. BOGO (Buy One Get One)
"Buy one, get the second free" (or at 50% off). Technically it's a two-item bundle, but the psychology is different: to the customer it feels like a gift, not a discount. And perception matters: according to a survey by AMG Strategic Advisors, 93% of customers prefer BOGO bundles over an equivalent percentage discount.
When to use it: to clear excess stock, to get a new product tried ("buy the best-seller, get the new arrival free"), or on high-margin products where you can absorb the cost of the second item. Careful: on thin margins BOGO eats you alive. Always run the numbers on the second item, not the first.
4. "Build your own" bundle
The customer builds the package by choosing N items from a selection (3 t-shirts of their choice, 4 flavors of a supplement, a custom skincare set). Perceived value rises because the choice is theirs, and you still increase the number of items per order.
When to use it: catalogs with many variants of equal price (sizes, colors, flavors, scents). Works great for apparel, cosmetics and supplements. It does require a bit of store configuration to handle variants and dynamic pricing.
5. Volume bundle (multipack)
The same product in multiple quantities at a decreasing unit price: "3 units, save 15%; 5 units, save 25%." The classic multipack.
When to use it: recurring-consumption products (cosmetics, supplements, household goods, socks, underwear). The volume bundle has an often-overlooked advantage: it grows the customer's stock at home and lengthens the time before they buy from a competitor. It's a retention lever disguised as a promotion, and it pairs perfectly with a subscription model.
6. Themed / occasion bundle
Products grouped around a theme or a moment: a "first apartment kit," a "travel set," a "gift for him under 50 euros." You're not selling products, you're selling a solution to a situation, and that raises perceived value well beyond the sum of the parts.
When to use it: during seasonal peaks (Christmas, Valentine's Day, Mother's Day, back to school) and to capture gifting intent. A ready-made "gift idea" bundle solves the choice anxiety of someone buying a present and converts better than a single product. If you plan your sale and holiday creatives around these bundles, your ads work much harder.
7. Cross-sell bundle (complementary products)
The main product plus its natural accessories: the smartphone plus case and screen protector, the shoe plus technical sock and waterproofing spray, the camera plus memory card and bag.
It's the bundle with the highest real incrementality, because it adds items the customer would buy elsewhere (or forget to buy at all). The secret is relevance: the pairing has to be obvious. Data helps a lot here, because order history tells you exactly what gets bought together. It's the same territory as upsell and cross-sell strategies, applied in pre-packaged form instead of suggested at checkout.
8. "Gateway" bundle (entry plus premium)
Pair an accessible, high-turnover product with a premium, high-margin one. The cheap item attracts, the premium one makes the margin. It's an elegant way to introduce an expensive product on the shelf that would get little attention on its own.
When to use it: when you have a little-known premium product and a best-seller that drives traffic. The bundle uses the first product's popularity to give visibility to the second. It's the same aggregation logic described by Bakos and Brynjolfsson: a bundle performs when the individual products' valuations are comparable, so take care with the price ratio between the two items.
9. Threshold bundle (gift with purchase / spend-and-get)
It's not a fixed package, it's a rule: "spend €79 and get a free sample," "above €100 shipping is free plus a free pouch." The bundle activates once a spending threshold is crossed.
When to use it: to push the order value toward a target threshold defined with data. If your average AOV is €62, setting the gift threshold at €79 nudges a chunk of customers to fill out the cart to get there. It's one of the cleanest levers for raising AOV without discounting the whole catalog, because you only reward those who spend more.
The real problem isn't the bundle type: it's the pricing
Once you've picked the type, the question that makes or breaks the P&L remains: how much discount? Here most e-commerce stores go by gut feel, and gut feel loses margin.
The rule of thumb is simple: a bundle discount only makes sense if it generates incrementality, i.e. sales that wouldn't have happened buying the items separately. If you discount a package of products the customer would have bought separately at full price anyway, you're simply giving away margin. The bundle has to either add items to the cart, convert a hesitant buyer, or clear dead stock. If it does none of these three, that discount is a straight cost.
The second mistake is anchoring the discount to a round percentage ("let's do -20% and call it a day") the same for every bundle. Each package has different margins on its components: an identical discount produces completely different profitability. A bundle of purely high-margin products can withstand -25%, one that includes a thin-margin item goes into loss already at -12%. The discount has to be calculated on the package's weighted margin, not on the sticker price.
The AstraLoop method: composing and pricing bundles with data and AI
This is where our work at AstraLoop comes in. Instead of picking pairings by gut feel, we start from the data the e-commerce store already has and use it to answer three operational questions.
What to put together. Order history contains the map of what customers already buy together (market basket analysis). A model run on transactional data surfaces the product pairs and triplets with real purchase affinity, including non-obvious combinations no one would have paired by hand. It's the same principle behind many AI use cases in e-commerce: turning behavioral data into catalog decisions.
At what price. For any given bundle you can estimate elasticity: what's the minimum discount that unlocks the purchase without underselling. Combining component margin with the segment's price sensitivity finds the discount threshold that maximizes absolute margin (not gross revenue). It's often lower than what would have been used by gut feel, and it performs better.
Who to show it to. Not every bundle should go to everyone. Whoever already bought product A is the ideal candidate for the refill or accessories bundle; whoever has a high cart is the target for the threshold bundle. Segmenting the base with an RFM analysis or with lead scoring gets the right bundle to the right person instead of running one generic offer for everyone.
These three pieces (composition, price, targeting) live in your own systems: catalog, order history, e-commerce CRM. Getting them to talk to each other is what turns bundles from "improvised kits" into a measurable margin lever.
Want to know which bundles would actually raise your margin, and at what price? Request an analysis of your store's order history: we'll identify the highest-affinity pairings and the optimal discount, product by product.
Mistakes to avoid (we see them in almost every catalog)
- Bundles that are too big. The "8-piece kit" impresses you, not the customer: it pushes the absolute price past the psychological threshold and stalls the decision. Better to have 2-3 well-targeted packages of 2-4 items than one mega-bundle.
- A dead product hidden inside. Slipping unsold stock into a bundle hoping it goes unnoticed doesn't clear the stock, it kills the bundle: the customer senses the forced pairing and grows suspicious of the whole offer.
- The same discount for every bundle. As noted, each package has a different margin. A single flat percentage is easy to manage but leaves money on the table for some and runs at a loss for others.
- No incrementality measurement. If you don't compare margin with and without the bundle on the same products, you don't know whether you're growing or just shifting sales to a discount. Define the right e-commerce KPIs first: AOV, margin per order, bundle attach rate.
- An invisible bundle. The best package is useless if it's buried at the bottom of the product page. It needs to be pushed on the product page, at checkout and in ads. Refining the product page and showing the bundle at the right moment matters as much as its composition.
Bundles and margin: summary table
A quick reference for which type to use depending on the economic goal.
| Goal | Recommended bundle type | Margin risk |
|---|---|---|
| Raise average order value | Threshold bundle, mixed bundle | Low (you only reward bigger spenders) |
| Clear dead stock | BOGO, volume bundle | Medium-high (check the margin on the second item) |
| Protect / increase margin | Cross-sell, gateway premium | Low (you're adding high-margin items) |
| Push a launch | BOGO with a free new item | Medium (trial cost on the new product) |
| Raise perceived value | Themed bundle, build your own | Low (you pay in packaging, not discount) |
| Retention and repeat purchase | Volume bundle (multipack) | Low (you grow stock and frequency) |
In summary
A bundle isn't "selling more for less." It's choosing the right type for the right goal and building it (products, price, recipient) on data rather than gut feel. The 9 types cover every scenario: from BOGO to clear stock, to cross-sell to build margin, to the threshold bundle to raise AOV without discounting the whole catalog.
The mixed bundle remains the safe default. The pure bundle is almost always a mistake outside native sets. And above every type, the variable that decides the P&L is the discount: calculated on the package's weighted margin, not on a flat percentage applied to everyone. That's where applying AI to order history makes the difference between a bundle that grows margin and one that burns it.
Frequently asked questions
What's the difference between a pure bundle and a mixed bundle?
In a pure bundle the products are sold only together, never separately. In a mixed bundle the customer can buy them separately or as a discounted package. Mixed is almost always more profitable because it doesn't remove freedom of choice: the Harvard study on Nintendo showed pure bundling dropped sales 20% compared to offering both options.
How much discount should you offer on a bundle?
There's no fixed percentage. The discount should be calculated on the weighted margin of the package's components, not on the sticker price. A bundle of purely high-margin products can withstand -25%, while one with a thin-margin item goes into loss already at -12%. The rule: a discount only makes sense if it generates incremental sales that wouldn't have happened buying the items separately.
Does BOGO always pay off?
No. BOGO works great for clearing stock, getting a new product tried, or on high-margin products where you can absorb the cost of the second item. On thin margins, though, it erodes profitability: the free second item weighs on the whole transaction. Always run the numbers on the second item's margin, not the first, before launching it.
How much can bundles raise average order value?
Industry estimates point to a 20-35% AOV increase with well-built bundles, and up to 30% of a mature e-commerce store's revenue can come from packaged products. But the result depends on composition: a bundle with sensible pairings and a calibrated discount grows margin, an improvised one burns it.
How do I know which products to pair in a bundle?
The most reliable way is to start from order history with a market basket analysis: the data shows which products customers already buy together, including non-obvious combinations. Avoid pairing a best-seller with a dead product hoping it gets dragged along: it works when products are complementary and have comparable price valuations, not by forcing unsold stock into the package.
Is it better to have few bundles or many?
Few and targeted. Catalogs with dozens of improvised kits confuse customers and don't convert. Better to have 2-3 well-thought-out packages of 2-4 items, each with a clear economic goal (clear stock, build margin, raise AOV), pushed on the product page and at checkout, than an 8-item mega-bundle that stalls the purchase decision.
If you want to turn bundles from a guessing game into a measurable margin lever, let's talk: we build composition, pricing and targeting on the data your e-commerce store already has.