A 10% increase in average order value is worth more than a 10% cut in acquisition cost. That holds on almost every account I have run. The arithmetic is short. Cutting cost saves you a slice of what you spend. Raising the basket earns you a slice of what you sell, and on a profitable order that slice is bigger. The basket route also asks nothing of the auction, which matters when ad prices climb every quarter. This one is for you if the only lever anybody in your business talks about is cost.
Why an increase in average order value is the cheaper lever
Contribution on one order is basket times margin, minus whatever you paid to acquire it.
Raise the basket by a tenth and you gain a tenth of basket times margin. Cut acquisition by a tenth and you gain a tenth of acquisition cost.
So the basket route wins whenever basket times margin is larger than acquisition cost. Which is the same as saying it wins whenever the order was profitable in the first place.
That is a strong claim, so read it twice. On any order that makes money, a 10% increase in average order value beats a 10% cut in CPA. The more profitable the order, the wider the gap between the two.
There is one exception. An account already underwater has nothing on the margin side to multiply, so cost is the only lever that reaches.
The arithmetic, with numbers you can check
Take an order at $100. Gross margin 45%, which sits inside the 20% to 50% range I see across ecommerce clients. That gives $45 of gross profit.
Now take off the costs that do not shrink when the basket does. Picking, packing, the delivery subsidy, the fixed part of the card fee. Call it $8. Contribution before media is $37.
Acquisition cost $25. Contribution after media is $12.
Raise the basket a tenth. The order is now $110, margin still 45%, so $49.50 of gross profit. The same $8 comes off, then the same $25. Contribution: $16.50, up 37.5%.
Instead cut acquisition a tenth. The basket stays at $100, so contribution is $37 minus $22.50. That is $14.50, up 20.8%.
Same 10% on both sides. One of them is worth almost twice the other.
What an increase in average order value does to break even ROAS
Most pages get this part wrong, so here is the careful version.
Break even ROAS is one divided by your contribution margin. In the example above that is $100 over $37, or 2.70x.
Raise the basket to $110 and contribution becomes $41.50. Break even ROAS is then 110 over 41.50, or 2.65x.
The move is small and it is free. It happens only because some order costs are fixed per order rather than proportional to revenue. If every cost scaled with revenue, a bigger basket would leave break even ROAS exactly where it started.
Cutting CPA does not move that threshold at all. It changes whether you clear the bar, not where the bar sits. Both are worth doing. Only one of them lowers the bar.
What the auction does when you cut CPA instead
Cutting acquisition cost means winning the same conversions for less, in a market moving the other way.
Meta reported its average price per ad up 12% year on year in both the first and second quarters of 2026. That is the company’s own investor reporting rather than an agency estimate.
Peaks are worse. Gupta Media’s social CPM tracker recorded Meta CPM at $17.70 on Cyber Monday, 138% above the 2024 annualised average, in data published in October 2025. Thanksgiving week averaged $13.42 against a 2025 Meta average of $8.19. The tracker runs on tens of billions of impressions as a 14-day trailing average.
So a 10% cost target cut is a negotiation with an auction that is not negotiating back. Usually you get less volume rather than a lower price.
The basket lever sits entirely inside your own store. Nobody is bidding against you for it.
The five things that actually increase average order value
Ranked by how often they work in the accounts I run. Every one of them moves AOV without touching a bid.
Bundles and multipacks come first. A three-pack priced above the single unit raises the basket with no discount code anywhere near it.
Delivery thresholds are second, and they need the margin checked rather than guessed.
Owned media is third, and it is the one people forget. When I scaled a client from $10,000 to $50,000 a month, I did not simply raise the Meta budget. I added objectives, awareness campaigns, custom audiences and lookalikes, plus channels the client was not using, including TikTok and Klaviyo. Klaviyo is email rather than an ad platform, so post-purchase and cross-sell flows there cost almost nothing per send.
Product mix is fourth. Stop bidding hardest on your cheapest SKU.
Value based bidding is fifth, and the ad account section below covers how to get there.
Free shipping thresholds and the margin trap
A threshold only pays when the extra margin beats the delivery you gave away. It is the second cheapest way to increase average order value and the easiest one to get wrong.
Work it through. Your basket sits at $100 and you set free delivery at $110. A customer adds $10 to qualify. At 45% margin that $10 earns you $4.50.
Now, what did the parcel cost? On most stores, considerably more than $4.50. So every order that only just clears the threshold has cost you money to win.
That does not make thresholds wrong. It makes a lazily set threshold wrong.
Set it far enough above your current basket that the incremental margin covers a real parcel. Price it against your actual carrier invoice rather than a round number somebody liked. Then watch how many orders land just above the line, because that tells you whether the threshold changed behaviour or simply discounted your postage.
The trade nobody measures: basket size against conversion rate
Anything that adds friction to raise the basket can cost you orders, and hardly anybody measures both sides.
Run the numbers. Suppose an upsell step lifts basket size by 3% and costs you 5% of conversions. Revenue per session becomes 1.03 times 0.95, which is 0.9785. You are down roughly 2%.
I have not run a controlled test of upsell popups against bundles, so I am not going to quote one. What I will say is which levers survive in my accounts. The ones built into the product and the offer last. The ones bolted onto the checkout rarely do.
So judge these changes on contribution per session rather than contribution per order. The per order view flatters anything that raises baskets by quietly annoying smaller customers into leaving.
How to increase average order value in the ad account itself
Three moves, and none of them touch the creative.
Split by price band. When cheap and expensive products share one campaign, budget follows whatever converts most easily, which is usually the cheap one. Separate them and each can be funded on its own economics.
Clean the feed. Excluding loss leaders from the catalogue is the fastest basket change available in most Shopping and catalogue setups.
Then move from a flat cost target to value based bidding, once volume allows. Google states that target ROAS on Search and Shopping needs at least 15 conversions in the past 30 days, with Demand Gen requiring 50 in 35 days. Google also advises waiting four weeks or three conversion cycles after a reporting change before judging results.
Audience matters as well. In my Meta flag and banner account, the best ad set returned 4.84 at $29.37 a purchase, while the worst active one returned 0.98 at $93.96. Different audiences buy different baskets.
What to change this week
Three things, in order.
First, work out contribution per order: basket times margin, minus the per-order costs that do not scale. Without that figure you cannot tell which of these two levers is bigger for you. I still cannot calculate it for one 2026 client, because their margin was never disclosed to me.
Then price one bundle above the single unit and put it in front of traffic you already buy. No new spend, no new auction pressure, and the answer arrives inside a fortnight.
Finally, check your delivery threshold against a real carrier invoice. That single line has been worth more than a bid change on more than one account I have looked at, and it takes an afternoon.
Landing pages decide what a click is worth covers the on-site half of this, two real acquisition numbers at disclosed volume gives you the cost side of the comparison, the number behind the number sets out what survives goods and fulfilment, and benchmarks worth trusting covers where the basket figures actually come from. On Meta this is a structure and feed decision as much as a merchandising one, which my Meta Ads work handles directly. To have the arithmetic run on your own account, book a teardown.