Optimising for value versus purchases: when value bidding lifts your AOV

Optimising for purchases teaches Meta to find as many buyers as possible, optimising for value teaches it to find valuable buyers. Value bidding can raise your average order value, but it needs enough signal volume to work.

Optimising for purchases tells Meta: find me as many buyers as possible. Optimising for value tells Meta: find me the buyers who spend the most. That difference can raise your average order value considerably, because the algorithm goes looking for the people who place big orders instead of simply the most orders. But value bidding only works if you have enough signal volume. Without it the algorithm has nothing to steer on.

What exactly is the difference?

When optimising for purchases, every conversion counts equally. An order of twenty euros and an order of two hundred euros are the same to the algorithm: a purchase. So Meta goes looking for people likely to buy, regardless of how much they spend. When optimising for value, you attach an amount to every conversion, and the algorithm learns from that. It no longer looks for the most buyers, but for the buyers who together deliver the highest value.

That sounds logical and attractive, and it is, provided the conditions are met. The algorithm can only steer on value if there really is a difference in what your customers spend, and if enough orders come in to extract patterns. Value bidding is not a button you flip for free revenue, it is a strategy with requirements. And those requirements are hard: an account that switches to value too early loses stability instead of winning AOV. So it pays to look honestly at your own order volume before you flip the switch.

When does value bidding raise your AOV?

Value bidding pays off most when your catalogue has a wide price range. If you sell products at both twenty and two hundred euros, there is genuinely something to win: the algorithm can learn who places the expensive orders and look for more of those people. If you also have bundles, higher price points or customers who buy a lot at once, value optimisation gives the system something meaningful to aim for.

You see the effect back in your average order value. Where optimising for purchases can bring you cheap buyers who erode your margin, value optimisation attracts the customers who spend more. For brands with a low margin per order that can be the difference between profitable and not, because every euro of ad budget now delivers a higher average order.

Optimise for purchases and you buy buyers. Optimise for value and you buy spend.

How much signal volume do you need?

This is the trap. Value optimisation learns from your conversion values, and to learn it needs enough data. If you are at a handful of orders a week, the algorithm has too little signal to reliably steer on who places the big orders. It then guesses more than it learns, and your results become erratic. In that case optimising for purchases is more stable, because it steers on a simpler goal.

• Enough orders: you need sufficient conversions per week so the algorithm can find patterns in the value.
• Real spread: there has to be a difference in what customers spend, otherwise there is nothing to optimise.
• Clean value signals: the conversion value you feed back has to be correct, otherwise the system learns the wrong thing.
• Enough budget: value optimisation needs room to learn before it pays off, like any learning phase.

In short, value bidding is powerful but hungry. It rewards accounts with volume and spread, and it punishes accounts that switch it on without the data to feed it. Know your position before you switch.

How do you choose between the two?

Look at your catalogue and your volume. If you have a wide price range and enough orders, value optimisation is worth trying, because there is room to raise your AOV. If you sell mostly one price point or still run limited volume, you are better off on purchases, because the algorithm then has a clearer and more achievable goal. You can also grow toward value optimisation: start on purchases, build volume and switch once you have enough signal.

That is how we approach it at AdSplicit. With 15M+ in profitable ad spend across 65+ brands we choose the optimisation goal based on the data the brand actually has, not on what sounds best in theory. The right choice is the one that fits your catalogue and your volume today, and it evolves as you scale. A good rule of thumb: if you doubt whether you have enough volume, you probably do not yet, and then you build further on purchases first.

Conclusion

Optimising for value can raise your average order value because the algorithm seeks the buyers who spend the most, but it needs enough signal volume and real spread in order value. If you have those, value pays off, if you do not, you stay on purchases. Want to know which optimisation goal fits your catalogue and volume, or whether you are ready for value bidding? Book a call and we will gladly look at it with you.