Offer timing versus offer content: which moves more

Offer optimisation work is overwhelmingly about contents and price. What goes in the bundle, how it is presented, what it costs, how the discount is framed. These get tested repeatedly.

When the offer appears is usually fixed early — after a level, on session start, on a schedule — and then never revisited. That asymmetry is backwards relative to how much each variable actually matters.

Why the moment carries more weight

An offer is an answer to a question the player may or may not be asking.

If a player has just failed the same level four times, they are asking how to get past it. An offer that helps with exactly that lands on a question already in their head. They do not have to be persuaded that the problem exists — they were just experiencing it.

The same offer shown at session start, to the same player, is answering nothing. There is no active problem. The player evaluates it abstractly, decides they might need it later, and closes it. Same contents, same price, same person, very different outcome.

This is why timing dominates. Contents determine whether an offer is relevant to a problem. Timing determines whether the problem is currently on the player's mind. The second condition gates the first — a perfectly matched offer arriving at the wrong moment is evaluated as though it were a random one.

The moments that reliably work

The pattern across most games is that offers convert around friction, not around convenience.

Immediately after a failure is the strongest single moment in most games. The player has just been stopped, the cause is fresh, and the desire to continue is at its peak. This window is short — a minute or two — and it closes when the player either succeeds or leaves.

Running out of a resource is the second. A player who cannot do the thing they were about to do has a specific, immediate, understood need. This is different from having a low balance in the abstract; it is having a low balance at the moment it prevents something.

Reaching a visible threshold is the third. A player who is close to something — a tier, an unlock, a completion — has a goal in view. Offers that close a visible gap convert well because the value is concrete rather than hypothetical.

What these share is that the player already knows what they want. The offer does not have to create demand, only satisfy it, and satisfying existing demand is a much easier sale than creating new demand.

The moments that reliably do not

Session start is the most common placement and among the weakest. The player has just opened the game with an intention, and the offer is an obstacle between them and it. Even a good offer is competing with the thing they came to do.

Immediately after a success is intuitive and mostly does not work. The player just achieved something and is satisfied. Satisfaction is not a buying state — the problem the offer solves was just solved by playing.

Fixed schedules produce the wrong moment for most players by construction. An offer that appears every Tuesday reaches each player at whatever point they happen to be at, which is uncorrelated with whether it is relevant. Some fraction get it at a good moment by chance, and that fraction is the entire conversion.

Repeated appearances at the same moment degrade fast. The fifth time an offer appears at the same trigger, the player has learned to dismiss it before reading it, and it has become a reflex rather than a decision.

See how we trigger offers on what a player is doing →

Testing timing instead of contents

The practical reason timing gets under-tested is that it is harder to instrument. Contents are a property of the offer; timing is a property of the game state when it fired, which has to be recorded deliberately.

The minimum is a field on every offer-shown event describing what the player had just done. Failed something, ran out of something, completed something, opened the game, nothing in particular. Five values covers most games and makes the analysis immediate.

With that field, the first question to ask is whether conversion varies by trigger for the same offer. It nearly always does, often by a wide margin, and the answer tells you where that offer belongs. An offer that converts at one trigger and not others should only fire at that trigger.

The second question is whether any of your triggers convert poorly across every offer. That is a placement problem rather than a catalog problem, and removing the trigger improves everything downstream by not spending player attention on moments that do not work.

The reason this is worth doing before another round of contents testing is economics. Changing what is in an offer requires design and production. Changing when it fires is configuration. The cheaper variable is also usually the larger one, and it is the one most catalogs have never varied.

Talk to us about offer timing →