# Ladders Nobody Had to Build: Smart Upsell Ladders, One Product Family at a Time What the Smart ladders rule engine picks, why each campaign gets one shared upgrade list, and how to split a large catalog into campaigns that work. URL: https://www.discountprime.app/case-studies/smart-upsell-ladders Updated: 2026-10-09 ## Summary Illustrative scenario: a cookware brand with 640 products uses Upsell Smart ladders, an Ultimate feature, to choose upgrade products by rule: same product type, priced 1.15x to 3x the trigger, margin no more than 5 points lower, in stock. Each campaign stores one top-4 list at activation and offers its first in-stock product to every trigger, so the brand runs one campaign per product family and price band. ## Short Answer Smart ladders lets a rule engine choose the upgrade products for an Upsell (Upgrade offer) instead of picking them by hand. It is useful when you have many entry products and do not want to research every premium sibling. It is not a per-product upgrade map for a whole catalog: - **One campaign gets one ranked list.** The engine scores candidates for each trigger product, merges the results, and stores the **top 4 products** as the campaign's upgrade list. The storefront then offers the first of them that is in stock, to every trigger in that campaign. - **The list is chosen at activation.** It does not follow the catalog on its own. Reactivating the campaign picks a fresh top 4. So the practical design is one campaign per product family and price band, not one campaign for the whole catalog. Smart ladders requires the Ultimate plan. Atlas Kitchenware is a fictional store used to illustrate the decision. All prices, costs and counts are illustrative assumptions, not results. ## Fits / Does Not Fit | Fits when | Does not fit when | |---|---| | A group of entry products shares one product type and a similar price | One campaign would cover many types or a wide price range | | Product types in Shopify are clean and consistent | Product type is blank or used loosely | | Unit costs are filled in Shopify, so the margin rule can work | Many products have no cost (the margin rule is then skipped) | | You can reactivate campaigns on a schedule after launches | You expect the upgrade list to update itself | | The upgrade products are single-variant, or any variant suits everyone | Customers must pick a size of the upgrade in the cart (see Limits) | ## The Scenario and the Data You Need **Atlas Kitchenware**, a fictional Shopify brand, sells 640 products: pans, pots, knives and small appliances. About 180 are entry-level items, and most have a premium sibling of the same type. A hand-built upgrade list was never finished. Before choosing a setup, Atlas needs: - **Product type** on every product, because the engine only compares products of the same type. - **Unit cost** in Shopify for triggers and candidates, because the margin rule needs both costs. - **Price bands** within each type, because the downsell check compares upgrades with the cheapest trigger in the campaign. - The number of **live set-layer campaigns** already running. Bundles, cross-sell, free gift and upsell campaigns share a limit of 25 live campaigns per shop. ## How the Rule Engine Chooses For each trigger product, a candidate must: | Rule | Detail | |---|---| | Same product type | Compared case-insensitively with the trigger's type | | In stock | At least one variant in stock | | Priced 1.15x to 3x the trigger | Using each product's cheapest in-stock variant | | Margin no more than 5 points below the trigger's | Only when both costs are known; with a missing cost, this rule is skipped | Candidates priced 1.2x to 1.8x the trigger rank higher, and so do candidates with a better margin. The per-trigger lists are merged into one, each product keeping its best score, and the top 4 become the campaign's upgrade list. Offers priced at or below the cheapest trigger are then dropped as downsells. Note what the margin rule is: "not much worse than the trigger", not "better than the trigger". A candidate 3 points below the trigger's margin but 50% more expensive usually still earns more gross profit per unit. And a candidate with an unknown cost passes without any margin check. ## The Options Compared | Upgrade source | How the list is built | Stays current | Per-trigger control | Plan | |---|---|---|---|---| | Manual picks | You choose up to 6 products, in order; the builder shows rule-engine suggestions with evidence (same type, price difference, margin points or "margin unknown", stock) and an Add button | Only when you edit | Full, through what you add | Prime and above | | Pool | A collection, tag, vendor or product type; offered cheapest first | Re-resolved automatically every few minutes | Through the collection or tag you curate | Prime and above | | Smart ladders | Top 4 rule-engine candidates, stored at activation | Only when reactivated | None inside the campaign; scope the triggers instead | Ultimate | Manual picks with the suggestion panel is the "rules propose, people decide" option, and it does not need Ultimate. Smart ladders saves that review step. Pool is the option that follows the catalog automatically, if you maintain a collection or tag such as `upgrade-skillet`. ## Why Atlas Splits the Catalog If Atlas put all 180 entry items into one Smart ladders campaign, the merged top 4 would come from whichever types scored best, and a knife buyer could be offered a premium sauté pan. Splitting fixes that: | Campaign | Triggers | Typical entry price | What Smart ladders can pick from | |---|---|---|---| | Skillet upgrades | 14 entry skillets | $45 to $60 | Skillets priced 1.15x to 3x a trigger (about $52 to $180 across this band) | | Saucepan upgrades | 11 entry saucepans | $35 to $50 | Saucepans in the same multiple range | | Chef knife upgrades | 9 entry knives | $40 to $65 | Knives in the same multiple range | Keep each campaign's triggers in a narrow price band. If a $45 and a $90 skillet share a campaign, an upgrade priced at $70 passes the downsell check (it is above $45) but would be a downsell for the $90 buyer, and the card would say "at no extra cost". With a limit of 25 live set-layer campaigns per shop, Atlas covers its highest-volume families first rather than all 180 items. ## The Incentive: 5% Off the Upgrade, With the Math Atlas adds a 5% "% off upgrade" incentive. It applies only to lines swapped with the cart swap card, per unit of the upgraded line. Illustrative skillet pair: entry $60 (unit cost $22, gross profit $38), premium $95 (unit cost $33, gross profit $62). | Customer | Pays | Gross profit | vs no campaign | |---|---|---|---| | Would have bought the entry skillet, upgrades | $90.25 | $57.25 | +$19.25 | | Would have bought the premium at full price, swaps instead | $90.25 | $57.25 | -$4.75 | At 5%, each cannibalized full-price sale costs about a quarter of what one genuine upgrade earns, so the incentive is low risk in this example. With no incentive the risk is zero, which is why a no-incentive run is the natural baseline. Nothing here shows that 5% converts better than 0%; that is what the measurement plan is for. If you raise the incentive, redo this table: a 20% incentive would make each cannibalized sale cost $19. ## Settings Used in This Scenario (one campaign) | Setting | Value | |---|---| | Campaign type | Upsell (Upgrade offer); Smart ladders needs Ultimate | | Upgrade mode | Better product | | Products the customer is buying | Specific products: 14 entry skillets, $45 to $60 | | Upgrade to | Smart ladders | | Incentive | % off upgrade: 5% | | Phrasing template | `Upgrade to {title} - only {diff} more` | | Placement | Cart swap card | | Eligibility | All customers | Why cart only: in Better product mode, the product page banner is a link to the upgrade's page, and a product added from there carries no swap tag, so it would not get the 5%. ## What the Customer Sees 1. A shopper adds an entry skillet. The cart card reads `Upgrade to Pro Skillet 28 cm - only $30.25 more` (illustrative: $95 minus 5% minus $60). 2. **Upgrade** removes one entry skillet and adds one Pro Skillet with the 5% applied at checkout. A toast offers **Undo** for 10 seconds by default. 3. If the Pro Skillet is out of stock, the card offers the next product in the stored list. Every skillet trigger in this campaign is offered the same first in-stock upgrade. ## Limits and Pre-Launch Tests - **Swaps add one fixed variant**: the upgrade product's cheapest in-stock variant at processing. If the premium skillet comes in 24 and 28 cm variants, the swap adds one of them. - **The incentive applies to every unit of the upgraded line**, including units added after the swap. - **Triggers are specific products.** New entry items must be added by hand. - **Refresh by reactivation.** New premium products do not enter the list until the campaign is reactivated. A retired product stays in the stored list until then; the card skips it only while it shows as out of stock, so reactivate after retiring or relaunching products. | Test input | Expected result | |---|---| | Builder preview before activation | Four upgrades, all of the trigger's type, none priced at or below the cheapest trigger | | Cart with the cheapest and the most expensive trigger | The offer is an upgrade for both; if not, narrow the price band | | Tap Upgrade, check checkout | 5% off the upgraded line only | | Raise the upgraded line to 2 | 5% off both units | | Premium pick out of stock | Card offers the next stored pick | | Product with no cost in Shopify | Shows "margin unknown" in suggestions; decide before activating | ## Measurement and When to Change Course - **Primary metric:** gross profit per order containing a product from the family, against a baseline period. - **Guardrails:** full-price sales of the premium products (cannibalization), incentive paid, Undo rate, and units per upgraded line. - **What Discount Prime records:** swapped lines per campaign and their revenue before discount, from paid orders. Attributed, not incremental; refunds are not subtracted. - **Confounders:** launches, seasonal gifting, and promotions on either product. A clean comparison is one family with the campaign and a similar family without it, or alternating periods with 0% and 5%. Without that, treat changes as directional. - **Continue** if gross profit per family order rises and premium full-price sales hold. - **Narrow the triggers** if the Undo rate is high for some entry products. - **Switch to Manual picks** if the stored top 4 keeps including products your team would not offer. Related: when the price gap itself is the obstacle, see [Pay at Most $15 More](/case-studies/upsell-cap-the-difference). ## Frequently Asked Questions **Q: How does Smart ladders choose upsell products?** A: For each trigger product, Discount Prime looks for in-stock products of the same product type, priced 1.15x to 3x the trigger, whose margin is no more than 5 points below the trigger's when both costs are known. It ranks them, merges the lists for all triggers in the campaign, and stores the top 4 as the campaign's upgrade list. If a cost is missing, the margin rule is skipped for that product. **Q: Does Smart ladders give each product its own upgrade?** A: No. One Upsell campaign has one stored list, and the cart offers its first in-stock product to every trigger in that campaign. To get sensible upgrades across a large catalog, create one campaign per product family and narrow price band. Note that bundles, cross-sell, free gift and upsell campaigns share a limit of 25 live campaigns per shop. **Q: Do Smart ladders update when I add or retire products?** A: Not automatically. The top 4 are chosen when the campaign is activated, and they stay until it is reactivated. A product that goes out of stock is skipped by the cart card, but new premium products only enter after reactivation. If you need a list that follows the catalog, use a Pool source based on a collection or tag, which is re-resolved automatically. **Q: Which plan includes Smart ladders, and what can I use without it?** A: Smart ladders requires the Ultimate plan. On Prime, Upsell works with Manual picks or a Pool, and the builder still shows rule-engine suggestions with their evidence (same type, price difference, margin points or margin unknown, stock) and an Add button, so you can use the rules to propose and choose the picks yourself. **Q: Can I pin or exclude specific upgrades in Smart ladders?** A: No. Smart ladders stores the engine's top 4 for the campaign; there is no pin or exclude control inside it. To control the result, narrow the trigger list to one product family and price band, fix product types and costs, or switch to Manual picks and add the suggested products you agree with.