Meta Advantage+ Shopping Campaigns: How Andromeda Powers ASC and How to Maximize ROAS

What Meta Advantage+ Shopping Campaigns and Andromeda Actually Do

Meta Advantage+ Shopping Campaigns Andromeda is not a single feature — it is the convergence of a campaign type (ASC) and the large-scale ranking model (Andromeda) that powers its delivery. ASC was designed specifically for e-commerce advertisers who want to consolidate prospecting and retargeting into one automated campaign. Andromeda is Meta's retrieval and ranking engine that replaced older candidate generation systems; it evaluates billions of user-creative pairs per second to identify which ad has the highest probability of driving a conversion for a given person at a given moment.

The implication is significant: you are no longer targeting audiences manually. You are training a model. Every purchase, add-to-cart, view content event, and product catalog interaction becomes a data point Andromeda uses to refine its delivery. To understand the full architecture behind this, read our deep dive on how Meta Advantage+ targeting works — it covers how Andromeda's candidate retrieval, ranking layers, and contextual signals interact.

"ASC campaigns that run with 8 or more distinct creative variants consistently outperform single-creative campaigns — not because of A/B testing logic, but because creative variety expands Andromeda's ability to match the right message to the right buyer."

The practical takeaway: your levers are creative diversity, pixel signal quality, catalog health, and budget. Everything else is handled by the model. Optimizing those four variables is the entire job.

Prerequisites: What You Need Before Launching ASC

Launching ASC without the right foundations wastes budget and produces misleading data. Before you create a single campaign, confirm all of the following are in place.

• Meta Pixel or Conversions API (CAPI) firing reliably. ASC optimization requires a minimum of 50 purchase events per week at the ad set level to exit the learning phase. If your pixel is dropping events or deduplication between CAPI and browser pixel is misconfigured, Andromeda receives corrupted signal and bids incorrectly.
• A clean, complete product catalog. Every product must have accurate titles, descriptions, prices, availability status, and high-quality images. Catalog errors suppress delivery and inflate CPMs.
• A minimum daily budget of at least 3–5× your average order value. Andromeda needs room to explore. Budgets below this threshold restrict the model to only the most obvious segments and prevent it from finding incremental purchasers.
• At least 6 months of purchase history in your ad account. ASC can work with newer accounts, but performance is substantially stronger when the account has historical conversion data that Andromeda can use as a prior.
• Creative assets ready across multiple formats. You need static images, short-form video (under 15 seconds), and catalog-linked dynamic formats at minimum before launch.

Step 1 — Structure Your Campaign to Feed Andromeda Maximum Signal

The structural decisions you make at campaign creation directly control how efficiently Andromeda can consolidate signal. Fragmentation is the enemy — too many campaigns splitting the same budget and audience pool creates competing auctions that drive up your own CPMs.

• Use one ASC campaign per business objective. If you sell across multiple product lines, test whether a single catalog and campaign outperforms separate ones. For most accounts under $100K monthly spend, consolidation wins.
• Enable the existing customer budget cap. ASC allows you to set a percentage of budget allocated to existing customers. Start at 20–30% and adjust based on your customer acquisition cost targets. This prevents the model from spending disproportionately on easy retargeting wins and masks true prospecting performance.
• Upload your existing customer list as a Custom Audience. This is how ASC distinguishes existing customers from new prospects for the budget cap feature. Without it, the cap cannot function correctly.
• Avoid overlapping ASC with manual prospecting or retargeting campaigns on the same catalog. Overlap creates auction competition against yourself. Consolidate first, then evaluate whether any manual campaigns are genuinely additive.
• Set campaign-level budget (CBO), not ad set level. Andromeda allocates budget more efficiently when it controls the full pool rather than working within fixed ad set constraints.

Step 2 — Build a Creative Portfolio That Drives Delivery

Creative diversity is the primary lever you control inside ASC. Andromeda's retrieval layer scores your creative assets against user profiles continuously — if all your ads look the same, the model has no variety to match against different buyer psychology segments. For a thorough breakdown of creative strategy within Andromeda's architecture, see our guide on Meta Andromeda ad targeting strategy.

• Vary the creative hook, not just the visual. Different value propositions (price, social proof, urgency, problem-solution) activate different buyer segments within Andromeda's delivery model.
• Refresh creative before fatigue, not after. Monitor frequency at the ad level. When frequency exceeds 3.5 among purchasers, introduce new variants rather than waiting for CTR to collapse.
• Keep winning ads running while introducing new ones. Pausing high-performing ads to test replacements disrupts delivery patterns. Add new creative alongside existing performers.
• Use Advantage+ creative enhancements selectively. Standard enhancements like background generation and text optimization can improve delivery in some accounts. Test them on and off at the ad level — do not apply blanket rules.

Step 3 — Set Budgets, Bids, and Attribution Correctly

Bid strategy and attribution window choices directly constrain what Andromeda optimizes for. Misaligned attribution windows are one of the most common and costly configuration errors in ASC accounts.

• Start with Highest Volume bid strategy. Unless you have a hard cost-per-purchase ceiling, avoid cost caps during the learning phase. Cost caps restrict Andromeda's bid range and frequently cause campaigns to underspend or stall in learning.
• Introduce a cost cap or ROAS target only after 50+ weekly purchases. Once the model has sufficient conversion data, cost-based bidding becomes stable. Introducing it before that point creates erratic delivery.
• Use the 7-day click, 1-day view attribution window as your primary measurement. This reflects the majority of purchase journeys accurately for most e-commerce categories. Switching to 1-day click only underreports conversions and may cause Andromeda to under-bid on valuable users.
• Align your ROAS target to blended business reality, not channel-level vanity. An aggressive ROAS target that reflects last-click attribution alone will restrict delivery to bottom-funnel audiences only, limiting the model's ability to find net-new customers.
• Scale budget gradually — increase by no more than 20–25% every 5–7 days. Sudden large budget increases force ASC back into learning and destabilize delivery for 7–14 days.

Step 4 — Analyze, Iterate, and Scale Without Breaking Learning

Reporting inside ASC requires a different mental model than manual campaigns. You cannot evaluate individual ad sets by audience performance because there are none. Your analysis must focus on creative-level data, conversion volume trends, and blended efficiency metrics.

• Review performance at the ad level, not the campaign level alone. The creative breakdown view in Ads Manager shows which ad formats and hooks are driving conversions. Use this to inform your next creative iteration cycle.
• Track the existing customer vs. new customer purchase split weekly. If existing customers account for more than 40% of conversions and you have prospecting goals, tighten your existing customer budget cap.
• Do not edit campaigns in the first 7 days after launch or a major change. Every significant edit — new creative, budget change over 25%, bid strategy change — resets the learning phase. Batch changes rather than making incremental daily adjustments.
• Use a 14-day rolling window for ROAS assessment. Single-day ROAS in ASC is volatile due to Andromeda's exploration cycles. Decisions made on 1–3 day windows lead to premature optimization and over-correction.
• Run incrementality tests before cutting budget on apparent low performers. Conversion lift studies and geo holdout tests frequently reveal that ASC is driving more incremental revenue than last-click attribution captures.

Common Mistakes That Kill ASC Performance

Even accounts with strong fundamentals routinely make structural and operational errors that suppress Andromeda's optimization. These are the most damaging patterns observed across e-commerce accounts in 2026.

• Running ASC alongside overlapping retargeting campaigns. If you have a separate retargeting campaign targeting the same pixel audience on the same catalog, you are competing against yourself in every auction. Consolidate or exclude audiences explicitly.
• Too few creatives at launch. Launching with one or two ads forces Andromeda to over-deliver on a narrow creative set, accelerating fatigue and limiting the user segments the model can efficiently reach.
• Ignoring catalog data quality. Disapproved products, missing GTIN data, inaccurate pricing, and low-resolution images reduce catalog reach and cause delivery errors that suppress overall campaign volume.
• Making daily bid changes to chase short-term ROAS fluctuations. Every bid strategy change triggers a new learning phase. Accounts that change bids more than once per week consistently underperform those that hold settings stable for 2–3 weeks.
• Misreading the existing customer budget cap. Many advertisers set this cap too high (60–80%) thinking it maximizes ROAS. In reality, it limits new customer acquisition and inflates reported ROAS with easy retargeting conversions that would have happened organically.
• Launching without CAPI configured. Browser-only pixel tracking loses a meaningful percentage of conversions due to browser privacy restrictions and iOS signal loss. CAPI server-side tracking recovers those signals and gives Andromeda accurate optimization data.

Expected Results and Timeline

ASC performance follows a predictable progression when the campaign is structured correctly. Understanding the timeline prevents premature campaign shutdowns during the critical learning and ramp phases.

Accounts with sufficient purchase volume (50+ weekly conversions) and strong creative variety routinely exit learning within 7 days. Accounts with thin conversion signals may take 2–3 weeks to stabilize. Industry practitioners commonly report 15–35% ROAS improvements when migrating from fragmented manual campaigns to a properly structured ASC setup — though results vary significantly by vertical, catalog size, and baseline creative quality.

Frequently Asked Questions