Bundle Rate Mining: Why Large Power Stations Bundle 3x Better
Summary
Measuring bundle purchase rate per SKU split the catalog into two populations: large-capacity models bundle at 27.6% vs 9.8% for small units — and BP3000's 24 sales with zero reviews exposed a social-proof leak.
Quick answer
Mine bundle opportunities by computing a bundle purchase rate per SKU — the share of orders where the product ships with a companion. Large-capacity models bundle at 27.6% (small-capacity P800: 9.8%), so design companion kits and accessory slots for large units first, follow real purchase pairs, and prioritize reviews for high-sales, zero-review SKUs like BP3000 (24 sold).
Raising average order value is one of those goals everyone talks about and few measure properly. The standard reflex is to build bundles everywhere. But bundling only works when the buyer already needs the second item. Instead of guessing which products could carry a bundle, we measured the data.
The metric: bundle purchase rate
We define the bundle rate as the share of orders in which a product is bought alongside a companion — a second unit, an accessory, or an expansion. Computing it per SKU split our catalog into two very different populations. Large-capacity models showed a bundle rate of 27.6%; small-capacity models such as the P800 sat at just 9.8%.
The gap is not random. A large power station implies an ecosystem — battery expansion, solar panels, cables. Small units are bought as impulse or single-purpose devices; there is nothing left to bundle with them.
The bundle rate also served as a sanity check for merchandising: instead of asking 'where should we add bundles?', we asked 'where do buyers already self-bundle?' — and let the orders answer.
What the data told us
- Design bundles where the need already exists. For large-capacity products we built companion kits and accessory slots on the product page, because the buying intent is already there.
- Follow the purchase groups. When users who buy A routinely also buy B, that pair is a bundle — the data named the pairs, not our instincts.
- High sales with zero reviews is a leak. A premium model (BP3000) sold 24 units with zero reviews — a conversion leak, because social proof is missing exactly where the highest-value customers make their decision.
Why the bundle rate matters for AOV
AOV is not raised by bundling everything; it is raised by bundling the products whose buyers already arrive with accessory intent. The 27.6% versus 9.8% gap told us exactly where to put our effort — and where not to waste page space on bundle widgets. In our own catalog, the large-capacity tier now carries the bundle placements and the small tier stays clean — the numbers made that call, not opinion.
The same view tells us which accessories to stock and which to retire, because it reflects real co-purchase behavior, not a merchandiser's assumption.
Reusable checklist: bundle rate mining in 4 steps
- Compute the bundle purchase rate per SKU.
- Identify the high-bundle groups — which items travel together in orders.
- Design bundle sets and accessory placements for large-capacity products first.
- Prioritize review acquisition for high-sales, low-review SKUs.
Bundle-rate analysis is a standard view inside our Shopify operations analysis module, right next to AOV tracking.
Frequently asked questions
What is bundle purchase rate?
The share of orders in which a product is bought alongside a companion — a second unit, an accessory, or an expansion.
Why do large-capacity products bundle roughly 3x better?
Large power stations imply an ecosystem — battery expansion, solar panels, cables — so bundle rate hits 27.6%, while small units bought on impulse or for one purpose sit at 9.8%.
What did the BP3000 data reveal?
A premium model sold 24 units with zero reviews — a conversion leak, because social proof is missing exactly where the highest-value customers decide. Such SKUs should get review acquisition first.