Raising average order value is usually cheaper than winning more customers, and most shops leave it alone because the tactics feel pushy. Done properly they are the opposite — they answer questions the buyer already has.
Average order value is total revenue divided by number of orders. It is the least glamorous of the three levers a shop has — traffic, conversion rate, order value — and frequently the cheapest to move, because it requires no additional visitors and no advertising spend.
The arithmetic is worth stating plainly. Lifting average order value by ten per cent lifts revenue by ten per cent on the same traffic, and lifts profit by more than ten per cent, because the costs of getting that visitor and processing that order barely change. Very few marketing activities have that shape.
Set the free-delivery threshold from your own data, not from a round number. Look at the distribution of order values, not the average: if most orders cluster just under sixty, a threshold at seventy-five moves a large group and one at a hundred and fifty moves nobody. This single decision does more for average order value than every recommendation engine, and it takes an afternoon with a spreadsheet.
Blanket discount codes, especially the ones a customer finds after filling the basket — they reduce order value and train people to search for a code before buying. Free delivery with no threshold, which removes the mechanism entirely. And an interface where adding a second item risks losing the first, which is more common than shop owners expect and worth testing on a phone before assuming otherwise.
Average order value sits between conversion rate optimization, which is about how many visitors buy at all, and customer lifetime value, which is about how often they come back. All three multiply, and the one most shops have never deliberately worked on is this one.
Ettex Sites carries the shop, the thresholds and the product pages where these decisions are actually made — the payment step is covered in checkout page — and the order data behind the arithmetic sits in Ettex Invoices, with the calculation in Ettex Sheets.
Being clear: there is no recommendation engine, no automated bundling and no personalisation. The suggestions on a page are ones you choose, which for a small catalogue produces better relevance than an algorithm with too little data to learn from.
Total revenue divided by number of orders over a period. Track the median alongside it, because a single large order distorts the mean badly at low volumes.
A free-delivery threshold set slightly above your current typical order, chosen from the distribution of order values rather than a round number.
When they are genuinely relevant — what this needs, what goes with it. Irrelevant suggestions are ignored at best and reduce trust at worst.
Yes, if it suppresses conversion. Tactics that push buyers into bigger baskets can reduce how many buy at all, which is why both metrics have to be read together.
Hreflang is the tag that stops your German page competing with your English one. It is also the tag most often implemented wrongly — and a broken implementation is worse than none at all.
A privacy policy is a disclosure, not a disclaimer. Generated from a template it describes a business that is not yours — and the first person to notice is usually the regulator or the customer who asked a question.
A returns policy is read before the purchase far more often than after it. Written as a defence against fraud it costs you sales; written as a promise it becomes one of the strongest reasons to order.