Optimization Events

Initiate Checkout Optimization vs Purchase Optimization: The Last Proxy Before Real Sales

In short: Both live in the Sales objective, with Initiate Checkout acting as the tightest common proxy for the real goal, Purchase. Initiate Checkout rewards reaching the payment or shipping step, which correlates with an eventual sale far more closely than a cart add does, as long as checkout itself has no major leaks; Purchase rewards the completed, paid order and is the only one of the two with real value attached in reporting. Because the two audiences overlap heavily on a healthy checkout, moving between them is usually a smaller relearning shock than moving from Add to Cart. Use Initiate Checkout while purchase volume is close but not quite sufficient, and move to Purchase once it is.

By the AdFlint research team · Fact-checked against current Google and Meta platform behavior · Last reviewed July 2026

Initiate Checkout Optimization

Targets people likely to begin the checkout flow, sitting between add to cart and purchase in both intent and event volume.

The event fires when a shopper enters checkout, so it carries stronger buying intent than a cart add while still firing more often than a completed order. It is the sensible middle step for stores that cannot sustain purchase optimization but find cart signal too loose. The trap is a leaky checkout: if payment or shipping steps kill conversions, you will buy abandoners efficiently and sell nothing.

Full definition

Purchase Optimization

Points Meta's delivery system at people predicted to complete a checkout, bidding for buyers rather than for clicks, sessions, or cart activity.

Meta needs a steady stream of purchase signals, from the pixel, the Conversions API, or both, to model who buys. It is the correct default for any store with real sales volume, since every shallower event is a proxy. The failure case is running it at a budget that cannot generate enough weekly purchases per ad set: delivery starves, the ad set never exits learning, and costs swing wildly.

Full definition

Side by side.

The differences that actually change what happens in your account.

	Initiate Checkout Optimization	Purchase Optimization
What has to happen for the event to fire	Checkout begins - payment or shipping fields opened.	Checkout completes and payment goes through.
Position in the funnel	One step upstream of Purchase, one step downstream of Add to Cart.	The final, ground-truth revenue event.
Gap to Purchase	Usually smaller than the Add-to-Cart-to-Purchase gap, since fewer people bail between checkout entry and payment.	No gap - it is the event itself.
What breaks the correlation	A leaky checkout - a broken payment step, a late shipping-cost reveal, forced account creation - makes it diverge badly from real sales.	Nothing to break; it only counts if payment actually clears.
Reporting	No direct order value attached.	Real order value and ROAS.
When to reach for it	Purchase volume close to sufficient but not quite there, and checkout itself is already solid.	Weekly purchase volume can sustain Meta's roughly 50-events-a-week threshold for stable learning.

What actually separates them.

Switching between Initiate Checkout and Purchase optimization is a setting change within the same Sales campaign, but it still triggers a learning-phase reset, so expect a short delivery dip regardless of how similar the two audiences turn out to be.

Initiate Checkout is the tighter of the common proxy events precisely because fewer people abandon between checkout entry and payment than between cart and checkout, which is why it is often preferred over Add to Cart as the fallback when Purchase volume is thin.

Purchase optimization carries real order value in its reporting; Initiate Checkout carries none, so a chain of Add to Cart, Initiate Checkout, and Purchase counts is the only way to diagnose exactly where a leaky funnel is losing people.

A checkout with genuine friction - a broken payment step, a late shipping-cost reveal, forced account creation - makes Initiate Checkout optimization actively counterproductive, since it rewards reaching a step people then bail from, worse than not optimizing on checkout behavior at all.

Because the two audiences overlap heavily when checkout is healthy, some advertisers run Initiate Checkout as a permanent floor under Purchase optimization rather than a temporary stepping stone, accepting a slightly softer signal in exchange for steadier volume.

Which one should you use?

Use Initiate Checkout Optimization when

Your Purchase volume is close to sufficient but not quite there yet.
You already tried Add to Cart and it felt too loose for your audience.
Your checkout flow is solid, so checkout entry correlates well with an eventual sale.
You want a tighter proxy than Add to Cart while building toward full Purchase optimization.

Use Purchase Optimization when

You have enough weekly purchase volume to optimize on it directly.
You have already fixed checkout leaks and no longer want to reward checkout starts that don't convert.
You want real order value and ROAS reflected in what delivery optimizes for.
You are past the stage of needing an intermediate funnel signal at all.

Run the numbers yourself

Meta Learning Phase Budget Calculator

Check whether your budget can exit Meta's learning phase, and see the maximum number of ad sets it can actually support.

Open the free calculator

Common questions.

Is Initiate Checkout a safer middle step than Add to Cart before moving to Purchase?

Often yes, since checkout entry correlates more closely with an actual sale than a cart add does, so the audience Meta finds under Initiate Checkout optimization looks more like your eventual buyers. It's a reasonable rung if Add to Cart still feels too loose and Purchase volume is not quite there yet.

My Initiate Checkout numbers look great but purchases didn't follow, what should I check first?

Look at your actual checkout flow for friction - unexpected shipping costs, a forced account-creation step, or a broken payment gateway will all buy you cheap checkout starts that never finish, since the algorithm is only rewarded for the start of checkout, not the end.

Should I run Initiate Checkout and Purchase optimization on the same product at the same time?

You can split them across separate ad sets or campaigns to compare, but a single ad set only optimizes for one event. Running both at once on identical audiences mainly tells you whether your checkout has a leak, not which setting performs better long-term.

Or stop choosing between them.

AdFlint picks the setting, writes the ads, and keeps optimizing inside the Google and Meta accounts you already own.

Try AdFlint free See how it works

Related comparisons

All comparisons