# Menu Item Ranking Test: Validate a 12% Lift Against Displacement Risk

Lucas Moreau · October 27, 2026

> Validate a 12% conversion lift by comparing incremental contribution margin against displaced top-10 revenue and the 15% promotion gate for ranked slots.

| Takeaway | Detail |
| --- | --- |
| Promote a new item only if its 30-day incremental contribution margin per ranking slot beats the displaced item's average contribution margin by at least 15%. | Reader rule for ranked discovery slots; the 15% margin gap is the promotion gate. |
| A 12% conversion lift is not enough by itself; it must cover the ranking tax from top-10 displacement. | Thesis: compare incremental contribution margin with revenue displaced from top-10 slots. |
| Invesp reports up to 30% more conversions in 90 days for 900+ brands, with fees tied to lift. | Grounding figure from a conversion optimization agency; pay only for the lift. |
| AI search growth of 200% pushes conversion competition upstream, per Jose Daniel Duarte Camacho. | WebWire figure; discovery and conversion systems must account for upstream visibility. |

This guide compares a 12% conversion lift against the revenue risk of displacing top-10 menu items.

It gives a 15% margin threshold and 30-day check for promoting new items to ranked discovery slots.

![Menu Item Ranking Test](https://static.mm-ais.com/article-images-ai/menu-item-ranking-test-validate-a-12-lif-ai-c42efd42.jpg)

## How ranking a new item actually works

Ranking a new item is a slot-allocation decision, not an additive visibility boost. Constructor's Commerce Reasoning Engine interprets behavior, context, and intent to deliver ranked products, so a new item enters a finite top-N pool and pushes one incumbent out (constructor.com). The mechanism matters because the pool has a fixed size: every promotion is a replacement, not an extra placement.

This section's core claim is that each promoted item displaces exactly one incumbent from the top-N pool. That means the candidate should be evaluated against the specific item it replaces, not against catalog-wide averages. A slot is a scarce position; gaining it requires another item to lose it.

Because displacement is one-for-one, the candidate's value is not its standalone conversion lift. The useful question is whether its incremental contribution margin in that specific slot is large enough to justify removing the incumbent that currently holds it.

Measurement should follow the same slot logic. Discovery commerce metrics must attribute revenue to the page where possible and monitor post-click session quality, not just point conversion rate (linq.direct). A new item can win clicks while producing weaker sessions, which would make the slot less productive even if the headline conversion rate rises.

The correct discovery efficiency metric is CAC relative to expected LTV, not point conversion rate (visiodp.com). Point conversion rate can improve while acquisition economics worsen, because it does not show whether the customers acquired through the new item justify their acquisition cost over their expected lifetime value.

The promotion test in the takeaways table applies directly to this mechanism: identify the exact incumbent the candidate would displace, record that incumbent's average contribution margin, and compare it with the candidate's projected 30-day incremental contribution margin per slot before granting the position.

![How ranking a new item actually works — Menu Item Ranking Test](https://static.mm-ais.com/article-images-ai/menu-item-ranking-test-validate-a-12-lif-ai-00be94d0.jpg)

## The evidence on conversion lift vs risk

Jose Daniel Duarte Camacho, reported via WebWire on news.google.com, says AI search is growing 200% and frames e-commerce's next conversion battle as moving upstream. For ranking decisions, that means a promoted item is not only competing for a visible slot on a page; it is competing with AI-mediated discovery paths that can redirect shopper attention before the ranked list is reached.

The second external anchor is the 12% conversion lift attached to a new menu item. It is the gain under review in this thesis, not a result imported from any single source in the evidence set, and it should be treated as a claim to be tested rather than as a default outcome of promotion.

Invesp claims up to 30% more conversions in 90 days across 900+ brands and uses pay-only-for-lift pricing. That establishes two narrower points: conversion lift can be measured in a test, and it can be written into a commercial agreement. It does not establish that 12% is a typical lift.

Shopify's 2026 personalization tactics guide, titled "Ecommerce Personalization: 12 Tactics & Examples (2026)," treats conversion lift as a tactic within a broader personalization program. Read alongside the Invesp claim, it suggests lift is one lever among several, not by itself a sufficient reason to occupy a ranked discovery slot.

The risk side of the comparison is therefore a displacement question rather than a missing lift number: if AI-mediated discovery is growing and lift programs are sold on aggregate outcomes, a 12% claim has to be weighed against the value of the position it would take. The evidence supports measuring lift and contracting for it; the 12% figure remains an anchor to be validated in the specific slot.

![The evidence on conversion lift vs risk — Menu Item Ranking Test](https://static.mm-ais.com/article-images-pixabay/menu-item-ranking-test-validate-a-12-lif-b58f42f6.jpg)

## Options compared: rank, browse-only, or skip

This section presents the three-option comparison table with an explicit winner condition. Apply the gates in the order defined later in the decision rules: check the skip conditions first (CAC versus expected LTV, then nonpositive margin), then the rank condition, then browse-only as the default outcome.

Option A is the only slot-consuming choice. It wins only when the item's 30-day incremental contribution margin per ranking slot exceeds the average contribution margin of the item it displaces by at least 15%. A smaller gap means the new item occupies a top-N position while the incumbent's contribution is pushed out, so the promotion does not clear its own slot cost.

Option B keeps the item discoverable without taking a ranked slot. It wins when standalone margin is positive but below the 15% displacement threshold, allowing the item to earn contribution while preserving the incumbent's position in the top-N pool. Use this when demand exists but the required margin gap is not met.

Option C removes the item from the ranked path entirely. It wins when the item cannibalizes a higher-margin incumbent or when its customer acquisition cost exceeds expected lifetime value, following visiodp.com's framing that discovery efficiency should be judged by CAC against expected LTV rather than by a point-in-time conversion rate.

| Option | Placement | Winner condition |
| --- | --- | --- |
| A | Rank in top-N | 30-day incremental contribution margin per slot exceeds the displaced item's average contribution margin by at least 15% |
| B | Browse-only | Standalone margin is positive but below the 15% displacement threshold |
| C | Skip entirely | Cannibalizes a higher-margin incumbent or CAC exceeds expected LTV (visiodp.com) |

Operational check: record the displaced item's average contribution margin before promotion, then compare the new item's 30-day incremental margin per slot. If the gap is under 15%, move to browse-only; if cannibalization or a CAC-versus-LTV failure appears, skip. Re-run the comparison after any menu, price, or placement change.

The table above states the winner condition for each option in one line; use it as the decision record when reviewing a promotion request.

![Menu Item Ranking Test, photo 2](https://static.mm-ais.com/article-images-pixabay/menu-item-ranking-test-validate-a-12-lif-6c0c8ba5.jpg)

## Costs and numbers that matter

The ranking tax is the contribution margin of the item your new menu item pushes out of the top-N pool. That is the whole definition, and it is the number most teams skip. Take the displaced incumbent's 30-day revenue, subtract its food cost, its labor, and its allocated overhead, and you have the tax. If the incumbent cleared $4,200 in revenue, carried $1,300 in food cost, $900 in labor, and $400 in allocated overhead, its contribution margin is $1,600. That $1,600 is what the slot was earning before you took it. A new item that earns less than $1,600 in the same window is not a win, no matter how many extra orders it appears to generate.

Now compute the new item the same way, over the same 30-day window. Revenue minus its own food cost, labor, and allocated overhead. The comparison is not conversion rate against conversion rate; it is incremental contribution margin per ranking slot against the displaced item's contribution margin. The reader rule is a 15% floor: promote only if the new item's 30-day incremental contribution margin per slot exceeds the displaced item's margin by at least 15%. Against a $1,600 displaced margin, that means the new item must clear $1,840. At $1,839, the slot change is a loss even though the item looks busy on the dashboard.

| Line | Displaced incumbent | New item (promotion threshold) |
| --- | --- | --- |
| 30-day revenue | $4,200 | — |
| Food cost | $1,300 | — |
| Labor | $900 | — |
| Allocated overhead | $400 | — |
| Contribution margin | $1,600 | $1,840 minimum |

Run the acquisition check before any of this matters. Visiodp.com frames the correct discovery-commerce metric as cost per acquired customer against expected lifetime value, not point conversion rate. If the new item's cost per acquired customer exceeds expected LTV, ranking it destroys value regardless of the conversion lift. That check sits upstream of the slot math: a positive contribution margin on a customer you paid too much to acquire is still a losing trade.

Two practical rules follow. First, measure both items in the same 30-day window so seasonality and traffic mix cancel out. Second, keep the displaced item's margin on file as the slot's baseline, because that baseline, not the new item's gross sales, is the bar every future promotion has to clear.

![Costs and numbers that matter — Menu Item Ranking Test](https://static.mm-ais.com/article-images-pixabay/menu-item-ranking-test-validate-a-12-lif-4ba365bb.jpg)

## What the evidence does NOT establish

The 12% conversion lift in the thesis is a hypothesis, not a measured menu-item result. None of the available sources sources reports a menu-item-specific conversion lift; Invesp's claim of up to 30% more conversions in 90 days is an agency-level outcome across 900+ brands, not evidence for one menu item in one ranked slot. Treat the 12% figure as a test assumption to validate, not as a benchmark to inherit.

Similarly, no grounding source supplies a slot-displacement margin threshold. The 15% rule is a comparison method—incremental contribution margin per ranking slot versus the average contribution margin of the displaced item—not a validated industry benchmark. Use it as an internal promotion gate, and record the displaced item's baseline before any test.

No grounding source provides CAC or LTV figures for menu items. The social-commerce source frames discovery efficiency as CAC relative to expected LTV, but it does not provide menu-item values. Those figures must be computed from the operator's own 90-day outlet-level data, including incremental contribution margin, displaced revenue, and acquisition cost for the promoted item.

The available sources supports general discovery and ranking mechanisms, but it does not bridge them to menu-item economics. A general CRO or AI-discovery claim cannot be converted into a slot-level margin conclusion without outlet-level measurement. If the baseline or lift is unmeasured, label it as hypothetical in the decision record.

Use grounding to justify testing, not to set thresholds: Invesp reports revenue-positive A/B tests for 900+ brands, and the social-commerce source identifies CAC versus LTV as the efficiency metric, but neither isolates a single menu item's displacement cost. Those sources support building a measurement plan; they do not supply the 12% lift or 15% margin values.

Practical check: before promoting, require a documented 30-day contribution-margin baseline for the displaced item and either a measured lift or an explicitly labeled hypothetical lift for the new item. If either input is missing, default to browse-only or skip rather than ranking.

| Framework item | Grounded in sources? | Required action |
| --- | --- | --- |
| 12% conversion lift | No menu-item-specific figure | Label as hypothesis and test |
| 15% displacement margin rule | No slot-displacement threshold | Use as internal gate with baseline |
| CAC and LTV for menu items | No menu-item figures | Compute from 90-day outlet data |

![What the evidence does NOT establish — Menu Item Ranking Test](https://static.mm-ais.com/article-images-pixabay/menu-item-ranking-test-validate-a-12-lif-af8d3099.jpg)

## 30-day slot test

Use this worksheet to test a candidate before it takes a top-10 slot. Enter 30-day revenue and the three cost lines, then compare the resulting contribution margin with the displaced item's margin plus the 15% hurdle.

| Line | New item | Displaced incumbent |
| --- | --- | --- |
| Revenue | $8,000 | $7,500 |
| Food cost | $2,400 (30%) | $2,100 (28%) |
| Labor | $2,000 (25%) | $1,800 (24%) |
| Overhead | $800 (10%) | $750 (10%) |
| Contribution margin | $2,800 | $2,850 |
| Required margin (15% above displaced) | $3,277.50 | — |

Base case: the new item has revenue of $8,000, food cost of 30% ($2,400), labor of 25% ($2,000), and overhead of 10% ($800). Its contribution margin is $8,000 − $2,400 − $2,000 − $800 = $2,800.

The displaced incumbent has revenue of $7,500, food cost of 28% ($2,100), labor of 24% ($1,800), and overhead of 10% ($750). Its contribution margin is $7,500 − $2,100 − $1,800 − $750 = $2,850.

Checkpoint 1 applies the 15% rule: the candidate must exceed the displaced margin by at least 15%. The required margin is $2,850 × 1.15 = $3,277.50. Because $2,800 is below both $2,850 and $3,277.50, the candidate fails the 30-day slot test in this base case.

Treat the new item's 30-day contribution margin as its incremental slot value; do not blend in revenue from other placements when making the comparison.

Sensitivity check: at the base-case cost rates, total costs are 65% of revenue, so contribution margin is 35%. To reach the $3,277.50 hurdle, 30-day revenue would need to be $3,277.50 ÷ 0.35 = $9,364.29, or about $9,364.

Use actual 30-day figures rather than annualized estimates, and rerun the worksheet whenever price, recipe, or labor assumptions change.

## Decision rules for menu item ranking

These four if/then rules turn the per-slot promotion standard (defined in the takeaways and the costs section above) into an operating procedure that decides rank, browse-only, or skip without relying on a conversion-lift percentage as the deciding evidence.

Rule 1 — rank: if the new item's 30-day contribution margin is greater than or equal to 1.15 times the displaced incumbent's margin, promote it to the top-N pool. Rule 2 — browse-only: if the margin is positive but below 1.15 times the displaced incumbent's margin, do not promote it; place it in browse-only and re-test after a full 30-day window.

Rule 3 — skip on economics: if the new item's customer acquisition cost is greater than its expected lifetime value, skip ranking even when conversion lift looks favorable; this follows the discovery-efficiency framing at visiodp.com, which treats CAC against expected LTV as the relevant efficiency metric. Rule 4 — skip on margin: if the new item's 30-day contribution margin is zero or negative, skip ranking and retain the incumbent.

Apply the rules in this order: Rule 3 first, then Rule 4, then Rule 1, then Rule 2. A skip condition therefore overrides a positive margin comparison, and a nonpositive margin overrides a browse-only placement. If none of the first three rules applies and the margin is positive but below the 1.15 multiple, Rule 2 is the default outcome.

Measure contribution margin after discounts, returns, payment costs, and fulfillment costs, and compare the new item with the specific incumbent it would displace in the same slot. Do not substitute gross revenue or a blended category margin for either side of the comparison. For browse-only re-tests, start a new 30-day window instead of appending days to the original test.

When a browse-only item later reaches the 1.15 multiple, promote it under Rule 1; when it remains below the multiple, keep it in browse-only or remove it from the queue. The 15% advantage is a minimum, not a target, so a result exactly at 1.15 qualifies under Rule 1.

## Frequently Asked Questions

**What is the promotion gate for a new item in ranked discovery slots?**

Promote a new item only if its 30-day incremental contribution margin per ranking slot beats the displaced item's average contribution margin by at least 15%.

**Is a 12% conversion lift sufficient by itself?**

A 12% conversion lift is not enough by itself; it must cover the ranking tax from top-10 displacement.

**What does the 15% margin gap represent?**

The 15% margin gap is the promotion gate.

**What external conversion benchmark is cited?**

Invesp reports up to 30% more conversions in 90 days for 900+ brands, with fees tied to lift.

**What upstream trend affects discovery and conversion systems?**

AI search growth of 200% pushes conversion competition upstream, per Jose Daniel Duarte Camacho.

**What happens to an incumbent when a new item is promoted?**

A new item enters a finite top-N pool and pushes one incumbent out.

## Quick answers

| What is the promotion gate for ranked discovery slots? | The 15% margin gap is the promotion gate. |
| --- | --- |
| What must a new item's 30-day incremental contribution margin per ranking slot do? | It must beat the displaced item's average contribution margin by at least 15%. |
| Is a 12% conversion lift enough by itself? | A 12% conversion lift is not enough by itself; it must cover the ranking tax from top-10 displacement. |
| What does the thesis compare? | It compares incremental contribution margin with revenue displaced from top-10 slots. |
| How does ranking a new item actually work? | Ranking a new item is a slot-allocation decision, not an additive visibility boost. |

Also worth reading: **Auditing Deliveroo and Wolt ranking disclosures under DSA Article 27: what restaurants can actually verify since 17 February 2024 — and what platforms still hide**: [Auditing Deliveroo and Wolt ranking](https://nolemon.io/blog/auditing-deliveroo-and-wolt-ranking-disclosures-under-dsa-article-27-what-restaurants-can-actually-verify-since-17-february-2024-and-what-platforms-still-hide.php) · **Local shop search ranking 2026: 38% to 24% cap not a penalty**: [Local shop search ranking 2026:](https://nolemon.io/blog/local-shop-search-ranking-2026-38-to-24-cap-not-a-penalty.php) · **Hotel restaurants as local listings: 15% margin test—list or skip Food and Beverage (F&B)**: [Hotel restaurants as local listings:](https://nolemon.io/blog/hotel-restaurants-as-local-listings-15-margin-testlist-or-skip-food-and-beverage-fb.php)

### Related reading

- [Auditing Deliveroo and Wolt ranking disclosures under DSA Article 27: what restaurants can actually verify since 17 February 2024 — and what platforms still hide](https://nolemon.io/blog/auditing-deliveroo-and-wolt-ranking-disclosures-under-dsa-article-27-what-restaurants-can-actually-verify-since-17-february-2024-and-what-platforms-still-hide.php)
- [Local shop search ranking 2026: 38% to 24% cap not a penalty](https://nolemon.io/blog/local-shop-search-ranking-2026-38-to-24-cap-not-a-penalty.php)
- [Local Pack Ranking: Why Proximity Beats Review Velocity](https://nolemon.io/blog/local-pack-ranking-why-proximity-beats-review-velocity.php)
- [2026 Catering Demand: Local Commerce Signals & Ranking Shifts](https://nolemon.io/blog/2026-catering-demand-local-commerce-signals-ranking-shifts.php)
- [Uber Eats Most Loved 2026: 90% Order Accuracy or Miss the Badge](https://nolemon.io/blog/uber-eats-most-loved-2026-90-order-accuracy-or-miss-the-badge.php)
- [Google LocalBusiness rating snippets: when a new review changes eligibility](https://nolemon.io/blog/google-localbusiness-rating-snippets-when-a-new-review-changes-eligibility.php)

### Latest

- [Uber Eats Most Loved 2026: 90% Order Accuracy or Miss the Badge](https://nolemon.io/blog/uber-eats-most-loved-2026-90-order-accuracy-or-miss-the-badge.php)
- [Google LocalBusiness rating snippets: when a new review changes eligibility](https://nolemon.io/blog/google-localbusiness-rating-snippets-when-a-new-review-changes-eligibility.php)
- [Five Iron Golf for Restaurants: Verifying Capex, Revenue Split & Cost Recovery](https://nolemon.io/blog/five-iron-golf-for-restaurants-verifying-capex-revenue-split-cost-recovery.php)

Canonical: https://nolemon.io/blog/menu-item-ranking-test-validate-a-12-lift-against-displacement-risk.php
Markdown: https://nolemon.io/blog/menu-item-ranking-test-validate-a-12-lift-against-displacement-risk.php/index.md