What we see in ecommerce analyses is that assortment performance is often reviewed SKU by SKU while shoppers experience a choice system. They compare the entry price, the apparent best value, the bundle, the premium option, delivery thresholds, subscription savings, and the risk of buying too much. A product can be healthy alone while the surrounding price-pack architecture makes the decision confusing or unprofitable.
GA4’s ecommerce reporting can measure item views, items added to cart, items purchased, and item revenue when the recommended event parameters are implemented correctly. Those statistics create a behavioral base. The commercial analysis must then add cost, discount, fulfillment, return, and inventory data.
Contents
- What price-pack architecture means
- The analysis dataset
- A practical price-ladder scorecard
- How to find choice problems
- Experiment without damaging margin
- Frequently asked questions
What price-pack architecture means
Price-pack architecture is the relationship between quantity, format, price, value, margin, and shopper mission. It answers questions such as:
- Is there a credible entry point for first-time buyers?
- Does the middle option communicate value without making the premium option irrelevant?
- Does a multipack improve contribution after pick, pack, and shipping costs?
- Does subscription pricing reward retention or simply subsidize an order that would have happened anyway?
- Are bundles helping discovery, or hiding slow inventory behind an attractive headline discount?
This is not a recommendation to manufacture artificial decoys. The objective is a legible assortment where each option has a real customer role and an economically defensible place.
The analysis dataset
Build the model at product, variant, bundle component, order line, and customer cohort levels.
| Field group | Examples | Analysis enabled |
|---|---|---|
| offer | unit count, size, list price, selling price | price and quantity ladder |
| economics | COGS, packaging, pick cost, payment fee | contribution by option |
| behavior | views, selections, carts, purchases | choice and conversion |
| order | basket value, shipping subsidy, discount | halo and threshold effects |
| post-purchase | refund, return reason, repeat timing | retained value |
| inventory | stock cover, age, availability | productivity and risk |
Use a consistent base unit where sensible: price per 100 ml, per item, per serving, or per month. Keep the actual selling unit too, because buyers do not experience unit economics alone. A twelve-pack can have excellent unit value and still create an unaffordable cash outlay.
Bundles need component-level cost and inventory consumption. If the analytics platform records only a parent bundle SKU while fulfillment decrements components, merchandising and operations will see different realities.
A practical price-ladder scorecard
| Metric | Calculation | Interpretation |
|---|---|---|
| entry conversion | purchasers of entry option / qualified viewers | accessibility |
| trade-up rate | buyers choosing larger or premium option / category buyers | value communication |
| contribution per order | retained sales − variable cost | commercial quality |
| contribution per unit | contribution / units shipped | pack efficiency |
| bundle attachment | orders with bundle / eligible orders | offer relevance |
| threshold subsidy rate | shipping or gift subsidy / retained sales | cost of basket growth |
| option concentration | share held by top option | choice balance or redundancy |
| repeat interval by pack | days to next purchase | consumption and retention fit |
The scorecard should be segmented by new versus returning customer, source, device, market, and promotion state. A large pack may work for branded search and repeat buyers while suppressing conversion on a paid prospecting landing page.
For the inventory side, pair the model with ecommerce assortment productivity and margin analysis. Price architecture is stronger when it releases cash as well as raising AOV.
How to find choice problems
Start with the category page and PDP as a connected decision surface. Look for:
- Invisible unit value. Shoppers cannot compare sizes or quantities without mental arithmetic.
- A broken step. The jump from entry to next pack is too large in cash terms or too small in value terms.
- Margin inversion. The promoted “best value” option contributes less after shipping and discount.
- Redundant options. Several variants split demand without serving distinct missions.
- Unavailable anchors. The option that makes the ladder understandable is frequently out of stock.
- Mobile ambiguity. Unit count, saving, or subscription condition is hidden below a sticky CTA.
An anonymous merchandising review found a category with healthy overall conversion but weak clarity between single, multipack, and subscription options. The team did not need another sitewide promotion. It needed consistent unit labels, a clearer default, and economics that included fulfillment. The analysis changed the question from “Which SKU sells most?” to “Which choice structure creates retained contribution?”
Experiment without damaging margin
Write a hypothesis that names the buyer problem and the protected metric. For example: making unit value visible will increase confident trade-up without reducing contribution per visitor. Then measure option views, selections, add-to-cart, conversion, contribution per visitor, refund rate, and repeat behavior.
Do not change price, copy, default selection, badge, and shipping threshold simultaneously unless the test is intentionally measuring a complete new proposition. Otherwise the team will know that something worked but not why.
Protect customers from misleading comparisons. Discounts should have a credible reference, subscription terms should be visible, and bundles should explain contents. Local consumer-pricing requirements vary; this article is operational guidance, not legal advice.
Evaluate inventory effects before scaling. A winning bundle can drain one constrained component and create stockouts across the rest of the catalogue. The test decision should include forecast capacity and substitution rules.
A monthly operating rhythm
Weekly trading should flag availability, margin inversion, promotion distortion, and sudden option shifts. Monthly analysis should revisit price ladders, customer cohorts, returns, and inventory turns. Quarterly reviews can consider structural changes such as a new pack, discontinuation, bundle strategy, or subscription proposition.
Assign ownership across merchandising, finance, ecommerce, and operations. If only growth owns the metric, AOV may rise while contribution or stock health deteriorates. If only finance owns it, useful customer choice may be removed because its role is not visible in a SKU margin report.
Frequently asked questions
Is higher AOV always evidence of better pack architecture?
No. AOV can rise because of discounting, shipping thresholds, price inflation, or fewer low-value customers. Evaluate contribution, conversion, returns, and repeat behavior beside it.
How many options should a PDP offer?
There is no universal number. Each option should serve a distinct mission and remain understandable on mobile. Remove options that create complexity without incremental customer or commercial value.
Should the cheapest unit price be the default?
Not automatically. The default should fit buyer intent, affordability, inventory, and transparent value. A larger pack can be efficient yet unsuitable for a first-time customer.
EcomToolkit’s view
The strongest assortment is not the one with the most SKUs or the biggest headline saving. It is the one where customers can understand the trade, operations can fulfill it, and the business keeps healthy contribution after discount, delivery, and returns. Contact EcomToolkit if your catalogue grows while choice clarity and margin confidence fall.
