What is value optimization on Meta?
Standard purchase optimization tells Meta: “Find people who will buy.” Every purchase counts the same regardless of order value. A $15 impulse buy and a $300 premium bundle are equal conversion events in the algorithm’s model.
Value optimization tells Meta: “Find people who will buy the most.” The algorithm receives purchase value data through your Pixel and Conversions API and learns which user signals (demographics, behavior, interests, device, time of day) correlate with higher-value purchases. Over time, it shifts delivery toward users more likely to place larger orders.
Meta’s Andromeda engine processes these value signals alongside creative signals to find high-value buyers at scale. The more accurate your purchase value data, the better Andromeda can optimize. This is why clean Pixel/CAPI implementation is a prerequisite. If your purchase values aren’t passing correctly, value optimization has nothing to learn from. See our Pixel and CAPI setup guide for the technical requirements.
For the full strategic framework, see our Meta Ads for eCommerce: The Complete Guide.
Our finding: Value optimization produces meaningfully different results than purchase optimization for brands with high AOV variance. Across accounts where the AOV range spans 5x or more (e.g., $30 lowest to $150+ highest), value optimization typically shifts the average order value up by 15-30% while CPA increases by 10-20%. The net effect is higher revenue per dollar spent, even though the cost per individual conversion is higher.
When does value optimization work best?
Value optimization isn’t universally better. It’s situationally better. These conditions determine whether it’ll outperform standard purchase optimization for your brand.
Wide AOV range. If your cheapest product is $20 and your premium bundle is $200, there’s a 10x value difference between your lowest and highest purchases. That variance gives the algorithm something to optimize toward. Brands selling single-product lines at a fixed price point ($49 for everyone) won’t see meaningful benefit because there’s no value gradient to learn.
Catalog diversity. Brands with many SKUs at different price points (fashion, beauty, home goods) give the algorithm more levers. It can learn that certain user profiles gravitate toward premium products and prioritize those users. Single-SKU brands or brands with a narrow product line have less for the algorithm to work with.
Strong conversion volume. Value optimization needs more data than purchase optimization because it’s learning a more complex pattern (value prediction, not just conversion prediction). Aim for 100+ purchases per week before testing value optimization. At lower volumes, the algorithm can’t reliably distinguish high-value from low-value user signals.
Accurate purchase value tracking. Your Pixel and CAPI must pass the correct purchase value for every transaction. If values are rounded, missing, or inconsistent (some orders pass value, others pass zero), the algorithm learns from corrupted data. Check your Event Manager quality score before enabling value optimization.
Products that upsell or bundle. If your site has effective upselling (cart page add-ons, bundle offers, subscription tiers), value optimization can learn to find users who are more susceptible to those upsell flows, further amplifying the value gap.
When should you stick with purchase optimization?
Value optimization isn’t a free upgrade. In some scenarios, standard purchase optimization performs equally well or better.
Narrow AOV range. If 80%+ of your orders fall within a $10-15 range (e.g., $45-60), there’s not enough value variance for the algorithm to optimize meaningfully. It would need to find users who consistently place $60 orders instead of $45 orders, a subtle distinction that requires enormous data to learn reliably. Purchase optimization is simpler and equally effective here.
Low conversion volume. Below 50-75 weekly purchases, the algorithm struggles to exit the learning phase for standard purchase optimization. Value optimization requires even more data because the model is more complex. Don’t add complexity when you don’t have the data foundation to support it.
New accounts or products. If you’re still finding product-market fit through Meta Ads, optimize for purchases first. You need volume and learning before you can optimize for value. Value optimization is a refinement, not a starting point.
Subscription-first brands. If your primary business model is subscription and the initial purchase value is relatively uniform ($29/month for everyone), there’s no value variance on the first transaction. Value optimization can’t see future LTV. It can only optimize for the purchase value it receives at the point of conversion. The exception is if you offer multiple subscription tiers with different initial prices.
Our finding: The most common value optimization mistake is enabling it too early. Brands hear “optimize for value” and assume it’s strictly better than optimizing for purchases. But value optimization with insufficient data (under 100 weekly conversions) or insufficient value variance (narrow AOV range) actually performs worse because the algorithm is trying to solve a harder problem with inadequate signal. We test value optimization as an A/B against purchase optimization for 2-3 weeks before committing. In roughly 40% of the accounts we test, purchase optimization wins because the conditions for value optimization aren’t met.
How do you set up value optimization?
Setting up value optimization is straightforward in Ads Manager, but the prerequisites are what determine whether it works.
Prerequisites checklist:
- Pixel and CAPI both active and passing purchase values accurately
- Event Manager shows “Good” or “Great” quality score for purchase events
- Purchase values match your actual order values (check for discrepancies between Meta-reported revenue and your CMS)
- 100+ weekly purchases in the account (not per campaign)
- AOV range spans 3x+ between your lowest and highest typical orders
Setup in Ads Manager:
- Create a new campaign or edit an existing one
- Under “Performance Goal,” select “Maximize value of conversions” instead of “Maximize number of conversions”
- Optionally set a minimum ROAS target. Start without one and let the algorithm learn freely. Add a minimum ROAS floor only after 2+ weeks of data if you need to enforce a profitability threshold
- Keep all other settings the same (broad targeting, Advantage+ Placements, etc.)
The ROAS floor decision:
Setting a minimum ROAS target (e.g., “don’t deliver below 3x ROAS”) constrains the algorithm’s delivery. It will only show ads when it predicts the conversion value will exceed your threshold. This can limit scale because the algorithm passes on opportunities that might have been profitable but fall below your floor.
Start without a ROAS floor. Let the algorithm learn for 2-3 weeks. Then evaluate: if revenue per conversion is strong but some individual purchases are far below your acceptable threshold, add a conservative ROAS floor (lower than your target, not equal to it). A 2x floor when your target is 4x gives the algorithm room to optimize while preventing truly unprofitable delivery.
How do you measure value optimization performance?
The metrics that matter change when you switch from purchase to value optimization. CPA and ROAS tell an incomplete story.
The metrics that matter:
| Metric | What to Watch | Why |
|---|---|---|
| Revenue per conversion | Should increase vs. purchase optimization | This is the core value optimization benefit |
| Total revenue | Should increase even if CPA rises | Higher-value orders compensate for higher acquisition costs |
| Average order value | Should shift upward | Algorithm is finding users who spend more |
| nCAC | May increase 10-20% | Expected tradeoff: paying more per customer who’s worth more |
| MER | Should improve or hold steady | Total efficiency should be net positive |
What not to worry about:
- CPA increasing. Value optimization deliberately targets higher-value users, who are often more expensive to acquire. A CPA increase of 10-20% paired with a 20-30% AOV increase is a net win
- Conversion volume decreasing slightly. The algorithm may deliver fewer but higher-value conversions. If total revenue increases, this is working as intended
- Daily fluctuation. Value optimization has higher variance than purchase optimization because it’s optimizing for a continuous value, not a binary event. Give it 2-3 weeks of stable data before evaluating
The A/B test framework:
The cleanest way to evaluate value optimization is to run it as an A/B test against purchase optimization in parallel campaigns for 2-3 weeks. Keep creative, budget, and targeting identical. Compare total revenue, revenue per conversion, and MER. If value optimization wins on total revenue and MER, switch. If it doesn’t, stick with purchase optimization.
We measure all of this through first-party attribution tied to the P&L using CAPI + Blotout, not through platform-reported metrics. Platform ROAS can be misleading when comparing optimization strategies because attribution windows treat high-value and low-value conversions identically.
Our finding: The accounts where value optimization produces the biggest lift share one characteristic: strong catalog depth with clear price tiers. A fashion brand with $30 basics, $80 mid-range, and $150+ premium lines saw a 25% increase in average order value after switching to value optimization, with only a 12% increase in CPA. Total revenue increased 18% at the same budget. But a single-product brand selling a $65 item saw no meaningful difference because there was no value gradient to optimize toward. The product catalog determines whether value optimization has room to work.
Frequently Asked Questions
Does value optimization work in ASC campaigns?
Yes. ASC supports value optimization through the “Maximize value of conversions” performance goal. The setup is the same as manual campaigns. ASC’s broad targeting combined with value optimization is a strong combination because the algorithm has the largest possible audience pool to find high-value buyers within. See our ASC playbook for campaign setup.
Can I use value optimization with a cost cap?
You can set a minimum ROAS target, which functions similarly to a cost cap for value optimization. However, adding constraints limits the algorithm’s flexibility. Start without constraints, let it learn for 2-3 weeks, then add a conservative minimum ROAS if needed. Don’t set an aggressive ROAS floor from day one because the algorithm needs room to explore.
How long does value optimization take to learn?
Longer than purchase optimization. Expect 2-3 weeks for the algorithm to build a reliable value prediction model, compared to 1-2 weeks for purchase optimization. During this period, performance may be inconsistent. Don’t switch back to purchase optimization before giving value optimization its full learning period.
Does value optimization change which creative performs best?
Sometimes. Value optimization may shift delivery toward creative that attracts higher-value buyers, which isn’t always the creative with the lowest CPA. A premium lifestyle video might underperform on CPA but attract $120 average orders. Under purchase optimization, it gets deprioritized. Under value optimization, it may become your top performer. Monitor creative performance shifts after switching.
Should I use value optimization for retargeting campaigns?
Generally no. Retargeting audiences are small and already high-intent. The algorithm doesn’t have enough audience variety within retargeting pools to meaningfully optimize for value. Purchase optimization is sufficient for retargeting because the goal is converting warm audiences, not selecting among them by value.
What to Read Next
- Meta Ads for eCommerce: The Complete Guide (2026) — The full strategic framework including campaign architecture, creative, and attribution
- Meta Ads Benchmarks for eCommerce: ROAS, CPC, CPM & CPA by Industry (2026) — Industry-specific performance data to set your targets
- Meta Pixel and Conversions API: The Complete eCommerce Tracking Guide — The technical foundation value optimization depends on
- How to Scale Meta Ads for eCommerce Without Killing Your ROAS — Scaling framework that works alongside value optimization
