Average order value: what range is healthy and why raising it beats raising conversion

When a seller wants to grow, the first thing they touch is almost always conversion. Rewrite the title, swap the photos, raise the ad budget, adjust price to win the buy box. All of that is worth doing. The problem is that conversion is the most expensive of the three levers available: raising it means fighting the algorithm, the competition and the cost per click, and every point you win can be taken back tomorrow.

The third lever, after traffic and conversion, is average order value: how much money comes in per order. It is the least attended of the three and, with the cost structure of a Mexican marketplace, usually the one that leaves the most profit per unit of effort. The reason is pure arithmetic: some costs are charged per order, not per peso sold, and those costs do not change whether the order carries one unit or three.

The usual obstacle is that the number is not clean anywhere. Amazon gives you units and sales per period, MercadoLibre gives you its own, and if you run your own store that is a third source with its own definition of “order”. Adding it up and dividing sounds trivial until you ask yourself whether you count cancelled orders, whether you include the shipping you charged, whether VAT is in or out, and whether an order with three units is one order or three.

This post is about calculating average order value so it is actually decidable, what ranges the industry reports and which markets they come from, and which levers really move it when the checkout is not yours.

what average order value is and which version you are looking at

The formula is revenue divided by number of orders over a period. AOV is exactly that. But every decision you make when building the numerator and the denominator changes the result, sometimes by double digits, and that is where the number stops being comparable to itself month over month.

Four decisions you need to fix and never move again:

• With or without VAT. If you sell in Mexico and report VAT-inclusive one month and VAT-exclusive the next, your ticket “changed” 16% without anything happening in the business. Pick one and write it down.
• Does the shipping the customer paid count? If you charge shipping in your own store and include it, your D2C ticket rises artificially against your marketplace ticket, where shipping is almost always baked into the price.
• Cancellations and returns. A ticket calculated over orders created is a different number from one calculated over orders delivered and paid. The second is the one that belongs in a profitability conversation.
• What counts as an order. On Amazon, a customer buying three units of the same ASIN produces one order. On MercadoLibre, depending on how the cart is built, you can receive separate sales that actually travel in the same shipment. If you do not consolidate, your ticket comes out lower than reality and your shipping cost per order comes out better than reality.

Alongside AOV, always carry the two siblings that explain where the movement came from: units per order and average unit price. The ticket is the product of the two. If the ticket rose 12%, knowing whether people are taking more pieces or whether you raised prices is the difference between celebrating and worrying.

the arithmetic: why a peso of ticket is not worth the same as a peso of conversion

An example. You sell headphones at $1,160 on the channel. For every single-unit order:

• Channel commission, 15%: $174
• Shipping cost you absorb: $120
• COGS (what the unit cost you): $520
• Packaging and handling: $20
• You keep $326, 28% of the price

Now the same order with two units, at $2,320:

• Commission at 15%: $348
• Shipping, one heavier package: $140
• COGS for two units: $1,040
• Packaging: $30
• You keep $762

Two units sold in separate orders leave you $652. The same two units in a single order leave you $762. That is $110 more for exactly the same product sold, and the entire difference comes from shipping and packaging being charged per package, not per unit.

Take that to a month. You have 100 single-unit orders: $116,000 in revenue and $32,600 in profit. You want to grow 20% and you have two paths:

• Through conversion: you get to 120 single-unit orders. Revenue $139,200, profit $39,120. You made $6,520 more, but you needed 20% more converted visits, and those get paid for with traffic or with ads.
• Through ticket: you keep 100 orders, but 20 of them carry two units. Revenue $139,200, exactly the same, and profit $41,320. You made $8,720 more without a single extra visit.

Same revenue, $2,200 of difference in profit, and one of the two paths requires the algorithm, the competition and your ad budget to cooperate.

Glossary: real net margin, with everything deducted →

There is an important exception almost nobody mentions, and it changes the size of the effect: if you fulfil through FBA or through Full, the fulfilment fee is charged per unit, not per shipment. The second unit pays its own fee and the dilution shrinks considerably. With self-shipping, FBM or a 3PL charging per package, the dilution is complete. Before chasing ticket, check which of the two worlds is yours, because the prize is not the same.

what ranges the industry reports and why almost none of them are yours

Time to be direct: there is no public, reliable average order value benchmark for marketplace sellers in Mexico. What circulates in 2026 are compilations of data from ecommerce platforms and analytics tools, built almost entirely on direct-to-consumer stores in the United States and Europe, not on Amazon or MercadoLibre listings.

Those 2026 compilations put the global AOV average around 170 dollars on rolling twelve-month windows, with vertical ranges going from roughly 50 dollars in frequently repurchased categories to over 400 in jewellery and automotive. They also consistently report two gaps: desktop versus mobile, with desktop tickets well above mobile, on the order of a third higher, and region versus region, with Europe above the Americas and the Americas above Asia-Pacific. All of those numbers are from Anglo and European D2C ecommerce. None is from a Mexican marketplace.

On the Mexican side, the useful public material is the annual online sales study from AMVO, whose 2026 edition reports a market that grew 19.2% in 2025 to over 940 billion pesos, with 17.7% penetration over retail and more than 77 million digital buyers. It is an excellent picture of the size of the pie, but it does not publish average ticket by category for marketplace sellers. The ticket figure quoted around that study, roughly 1,391 pesos with 14% growth, comes from an own-store platform, not from Amazon or MercadoLibre, which is why it is not a yardstick for your listings.

Practical conclusion: use those numbers to understand the shape of the phenomenon, high ticket where the buying decision is long, low ticket where repeat purchase is frequent, mobile below desktop, and never as a target. The only honest yardstick is your own.

how to build your own baseline

Your average order value baseline is not a number: it is four cuts of the same number, and each answers a different question.

• By channel. Amazon, MercadoLibre and your own store have different audiences, commissions and shipping rules. Comparing them is useful; averaging them into a single number is not.
• By product family. A whole-catalog average rises and falls depending on what sold that month, not on what you did. The family cut isolates the mix effect.
• With and without season. Buen Fin, Hot Sale and December distort the ticket in both directions: some sellers move more units per order, others sell cheaper. Keep the baseline of normal months separate.
• First purchase versus repeat. If you can identify it, which you can in your own store and almost never in marketplace, the returning customer’s ticket usually tells a different story.

With those four cuts, a 6% move in the ticket stops being noise and starts being information. Without them, the same 6% can just mean more expensive items happened to sell this month.

the levers that actually move the ticket on a marketplace

In your own store you own the checkout and can do almost anything. On a marketplace you cannot: the purchase flow belongs to the platform and your levers are fewer, but they exist and they are concrete.

• Quantity per listing. Publishing 2, 3 or 6-unit packs as their own SKU is the most direct lever and the most underused. It is not a discount: it is a different product, with its own per-package shipping cost and its own price.
• Properly built variants. A parent with clean sizes and colors makes the buyer find the combination they wanted instead of leaving. It lifts conversion and, where the category allows, units per order too.
• Real kits. The main product plus the consumable or accessory the buyer was going to get later anyway. A kit works when it solves a purchase the customer already had pending, not when it staples together two things that are never used together.
• Volume price steps. Where the channel allows it, a per-unit price that drops as quantity rises. Watch the floor: the discount has to fit inside the real saving on shipping and packaging, not come out of your margin.
• Trading the catalog up. The higher-capacity, heavier or longer-lasting version. Raising average unit price with product that genuinely delivers more is the most stable lever of all, because it does not depend on the cart.

Glossary: laddered offers, drop to sell and raise to win margin →

And a policy warning: MercadoLibre Mexico requires free shipping above a certain per-unit amount, and that threshold has moved several times. It turns your shipping cost into a step function rather than a straight line, and it completely changes which pack size is worth building. Before designing bundles, confirm the current threshold in MercadoLibre’s help center, and check the category rules for Amazon in Seller Central.

the levers that raise the ticket and lower profit

Not everything that lifts AOV is good business, and this is the mistake that costs the most.

The classic case is the free shipping threshold in your own store. It is the favourite tactic for lifting the ticket and also the easiest to calibrate badly: set it too far above your current ticket and people do not reach it and buy less. There is research on one online retailer, American and not Mexican, documenting close to a 19% drop in consumer spending when the threshold moved from 80 to 100 dollars, with most of the drop coming from cart value rather than frequency. The lesson is not the number, it is the shape: the threshold only works close to where your ticket already is.

The second case is the discounted bundle. A pack that discounts 25% to lift the ticket 30% can leave you less profit per order than selling the pieces separately. The only valid test is the usual one: calculate profit per order for the bundle against profit per order for the standalone product, with real commission, shipping and COGS. If the bundle leaves fewer pesos than the simple sale, you raised AOV and lowered the business.

The third is quieter: raising the ticket by narrowing the catalog. Pushing only the expensive items lifts the average and, on the way, leaves you without the entry products that bring new customers. The ticket looks better and the traffic goes out.

One rule covers all three: any ticket tactic is judged on profit per order, never on revenue per order. If the two do not move in the same direction, the first one wins. How to calculate the real profitability of a marketplace product has the full breakdown of that calculation.

average order value in iqseller

In the panel, average order value does not live alone: it lives next to profitability, which is where you find out whether the lever worked.

The Profitability module calculates net margin per SKU using the COGS you loaded, the commissions coming from the Amazon settlement and from MercadoLibre orders, FBA and Full fees, shipping, advertising and the breakdown of VAT and withholdings. That is the data that turns “the ticket went up” into “profit per order went up”, which is not the same thing. The Parent → Model → SKU tree lets you read it by product family instead of by isolated unit, which is exactly the cut you need when you are testing 2, 3 and 6-unit packs of the same item.

The Catalog module keeps linked the SKUs that are the same real product in different pack sizes and channels, so the comparison between the single unit and the pack is against the same base cost and not against two records living separate lives. Pricing shows which price is live on each channel at each moment, which is what lets you know whether the volume step is still above your floor. And Alerts flags when a price or commission change pushes a SKU below that floor.

None of this raises the ticket for you. What it does is make sure that when you test a new pack size, you have the number that decides whether it stays or goes, instead of waiting for month-end to have a suspicion.

what to review every month

A minimum rhythm that fits in half an hour:

• Average order value by channel, with the same definition as always, and next to it its two components: units per order and average unit price.
• Profit per order by channel. If the ticket rises and this falls, the lever you used is coming out of your margin.
• The five SKUs that pushed the ticket up the most and the five that pushed it down, with their margin beside them.
• The multi-unit packs currently live: how many orders they brought and how much profit per order they left compared with the single unit.
• An eye on the current free shipping threshold on each channel, because when the platform moves it, all of your bundle arithmetic changes that same day.

With that, average order value stops being a reporting metric and becomes a decision: which pack you publish next month and which one you retire.