Optimize Meta Ads to LTV, Not First-Order ROAS

Why first-order ROAS is the wrong bar for repeat-purchase brands

A ROAS target measures how much a customer paid back on one transaction. It does not measure what that customer is worth. For a brand where repeat purchase or subscription revenue makes up a meaningful share of total revenue, that gap is not a rounding error. It is a structural mismatch between the metric and the economics. First-order ROAS measures transaction quality. LTV measures customer quality. Most Meta accounts are still built entirely around the first number.

Here is what that mismatch does in practice. Say a brand sets a flat 2x ROAS target because that number felt safe, or because it matched the account's overall gross margin. Every campaign that lands below 2x on the first order gets paused or defunded, regardless of what those same customers go on to spend. If the brand sells a product people reorder, or runs a subscription, the first order is rarely where the profit lives. It is the entry ticket. A target built only on the first order will reject exactly the customers a repeat-purchase brand most wants: the ones who come back.

This same gap shows up in the metrics operators already track side by side with ROAS: CAC, MER (marketing efficiency ratio), and increasingly NC-ROAS (new-customer ROAS), which isolates first-order performance on new buyers specifically. All three share the same blind spot as a flat ROAS target: they describe the acquisition moment, not the customer relationship that follows it.

This is not an argument for ignoring ROAS. It is an argument for measuring it against the right number, at the right horizon. First-order ROAS measures profitability at the acquisition horizon, the first transaction. The LTV-adjusted floor measures it at the customer horizon, across however long that customer keeps buying.

How to optimize Meta ads to LTV: the LTV-adjusted ROAS floor

The formula:

LTV-adjusted ROAS floor ≈ Break-even ROAS ÷ (LTV ÷ AOV)

Break-even ROAS is the standard 1 ÷ gross margin figure most media buyers already calculate. The LTV-to-AOV multiple is the customer's lifetime value over a defined window, commonly 12 months, divided by the average first order value. Divide one by the other and the result is the real floor: the first-order ROAS a campaign needs to hit for the brand to break even once repeat purchases are counted, not just the first sale.

Each piece of this connects to the next. Average order value and gross margin set the break-even ROAS. Lifetime value divided by average order value sets the multiple. Break-even ROAS divided by that multiple sets the floor. The floor, in turn, sets the real target cost per acquisition the campaign should be judged against, rather than a flat ROAS number chosen without reference to how the customer actually behaves after the first purchase.

One nuance worth stating plainly: the margin figure that belongs in this formula is contribution margin, gross profit after fulfillment and variable costs, not the topline gross margin percentage some brands default to. Using contribution margin keeps the break-even ROAS honest about what actually reaches the bottom line.

The table below works through the math with illustrative numbers. This is a worked example to show the calculation, not a claim about any specific brand's actual results. Illustrative worked example, not a benchmark.

Read against the naive 1.82 break-even bar, a campaign delivering a 0.9 first-order ROAS looks like a clear loser. Read against the 0.51 LTV-adjusted floor, that same campaign is well above the line, and it is likely funding customers who become highly profitable by their second or third order. The gap between 1.82 and 0.51 is the whole argument: a brand using the wrong number would have paused a campaign that was actually working.

The exact multiple depends on a brand's real repeat-purchase and subscription data, and it moves over time as retention changes. Treat the table as a way to see how the calculation behaves, not as a universal target. In practice, brands arrive at their multiple through cohort analysis: grouping customers by acquisition month and tracking each cohort's retention curve and repeat purchase rate over time, rather than reading a single blended average across the whole customer base.

How Meta's value optimization actually uses your LTV signal

Meta's value-optimization and Highest Value bid strategies already try to predict which users are worth the most and bid accordingly, using Meta's own documentation on maximizing the value of conversions as the reference for how the system behaves. The system, though, can only optimize toward the value signal it is given. Meta's bidding is bounded by the event value it receives: it cannot optimize toward a value that was never passed to it, not because the system is incapable, but because it has nothing else to bid on.

If the conversion value sent to Meta is first-order revenue, Highest Value bidding will likely spend toward people who spend the most on their first order, which is a different population than people who are worth the most over 12 months. Meta's own integration guidance is explicit that a margin-adjusted or predicted-LTV value, often shorthanded as pLTV, should be added as the value and currency parameters on both the Meta Pixel and the Conversions API, kept consistent across both sources rather than sent through one and not the other. Setting up value optimization this way redirects delivery toward the higher-lifetime-value population instead of the biggest first-order spenders. In plain terms: the algorithm is not wrong, it is answering the question it was asked, and most accounts are asking it the wrong question. The quality of your value signal tends to set a ceiling on the quality of Meta's bidding decisions. Changing a bid strategy without changing the underlying value signal rarely changes who Meta actually buys. This is also the mechanism behind the scaling guidance in how to scale Facebook ads without killing ROAS: scaling too fast on the wrong value signal tends to compound the mismatch, not just the spend.

This is also where an approval-based, margin-aware AI media buyer does real work rather than just adding automation for its own sake. A system that reads AOV, repeat rate, and LTV can compute the LTV-adjusted floor and propose the value signal and target CPA for a human to approve before it goes live, rather than leaving a media buyer to guess at the right number by hand. AdAdvisor, built by a team with over seven years in paid ads management, more than $60M in managed ad spend, and an ex-Meta data engineer who has built and shipped AI products, uses this kind of calculation inside its Nova approval workflow, proposing the LTV-adjusted target rather than asserting a flat ROAS number.

Where this applies fastest: subscription beauty and supplements

Two categories feel this mismatch earlier and more sharply than most.

Subscription and replenishment beauty, think skincare and haircare consumables sold on a recurring cadence, frequently sets first orders near break-even by design, often through an introductory offer or a trial size. A flat ROAS bar filters out exactly the subscribers the whole model depends on, because it judges the loss-leading first order in isolation.

Supplements carry the same pattern from a different angle. Repeat purchase and subscription replenishment routinely make up the majority of a supplement brand's lifetime revenue, so the true ceiling on what a brand can pay to acquire a customer is the LTV, not the first order. This is the same position the supplement-brand compliance and LTV piece makes in more industry-specific detail, including how policy-safe creative scaling interacts with this same LTV math.

These are two related but different numbers, worth keeping separate. One 2026 DTC benchmark set reports subscription LTV running roughly 3 to 5 times higher than one-time-purchase LTV at the same gross margin, with supplements specifically cited around 3 to 4 times and beauty around 3 to 5 times, driven mainly by order frequency rather than any pricing difference. That is the subscription-versus-one-time LTV multiplier, not the LTV to CAC ratio. On LTV to CAC specifically, the same source's companion benchmark puts the general ecommerce sweet spot at roughly 3:1 to 4:1, with Beauty & Personal Care reported around 3.2:1 and Pet Supplies around 3.8:1 in its 2026 vertical table. Reported figures vary by source, vertical, and cohort, so treat these as directional ranges rather than fixed targets, but the pattern is consistent with the size of the gap the LTV-adjusted floor is built to correct.

When the LTV-adjusted floor does not apply

This framework assumes repeat purchase or subscription revenue is doing real work in the business. Where that assumption breaks, first-order ROAS is likely still the right bar, or close to it.

Low-repeat and one-time-purchase categories, big-ticket items like furniture or major appliances, and genuinely seasonal businesses with no meaningful repeat cadence all have an LTV-to-AOV multiple close to 1. In that case the LTV-adjusted floor and the naive break-even ROAS converge, and there is little to gain from the extra calculation.

The framework also depends on having margin and LTV data a brand actually trusts. A retention curve built on six weeks of data, or a contribution margin figure that has not accounted for returns and fulfillment costs, will produce a floor that is confidently wrong rather than usefully imprecise. Brands without at least a few months of cohort data are generally better served tightening their first-order ROAS discipline before loosening it based on a projected multiple.

One more honest caveat: shifting the value signal Meta receives changes who the algorithm targets, but it does not by itself prove those customers were incremental, meaning they would not have converted anyway. Brands making a large LTV-based shift should still sense-check it against incrementality testing rather than assuming every gain in predicted LTV is a real gain in profit.

Frequently asked questions

The bottom line

The number most brands use to judge their Meta ads is measuring the wrong transaction. The mistake is not that marketers use ROAS. The mistake is assuming the first order is where profitability should be measured. For any retention-driven business, the customer's lifetime, not their first checkout receipt, is the economic unit that actually matters. The LTV-adjusted ROAS floor, the break-even ROAS divided by the LTV-to-AOV multiple, is the number that reflects that reality, and it is usually a lot more forgiving than the flat target most accounts are still using. Supplement and subscription-beauty brands feel this gap earliest, since that is where repeat and subscription revenue does the most work.

If you want the deeper, compliance-specific version of this argument for supplements, read the supplement-brand piece next.

Sources

• Meta Business Help Center, "About Highest Value"
• Meta Business Help Center, "Set Up Value Optimization"
• Meta Business Help Center, "About maximizing the value of conversions"
• Eightx, "LTV:CAC Ratio Guide (2026)"
• Eightx, "Subscription LTV vs One-Time Purchase: 2026 DTC Benchmarks"
• Meta for Developers, Conversions API, "Integration Guidance: Value Optimization"