Data-Oriented Design: Thinking About Memory, Not Objects

Data-oriented design is a shift in how you think about performance-critical code: instead of modeling the problem as objects with behavior, you model it as data and the transformations applied to it, organized for how the hardware actually accesses memory. For the hot paths of a game, this perspective can unlock performance that object-oriented thinking leaves on the table.

The CPU is waiting on memory, not math

A counterintuitive truth about modern hardware is that processors are often far faster than memory, so a huge amount of time in performance-critical code is spent waiting for data to arrive from memory rather than doing computation. The CPU loads memory in chunks into a fast cache, and code that accesses data which is already in cache runs dramatically faster than code that has to fetch from main memory. This is the central insight of data-oriented design: the layout and access pattern of your data, not the cleverness of your algorithms, is frequently what determines performance, because cache misses dominate the cost. Object-oriented designs that scatter related data across many separately-allocated objects, chased through pointers, produce exactly the cache-unfriendly access patterns that leave the CPU stalling.

Data-oriented design organizes for cache-friendly access in the hot paths. The practical move is to lay out data so that the things you process together are stored together, contiguously, and to process them in tight batches—iterating over a packed array of the data you need rather than jumping around memory through object references. This keeps the data the CPU needs in cache and the processor fed, which can be many times faster for the hot loops that run over thousands of items every frame. You don't apply this everywhere—most code isn't performance-critical and benefits more from being clear and object-oriented—but in the hot paths where you process large amounts of data every frame, thinking about memory layout rather than conceptual objects is what unlocks the big performance wins. This is closely related to why ECS is fast, and the underlying principle is the same: feed the cache by keeping the data you process together, contiguous, and accessed in order.

Scope is a decision, not an accident

Almost every overscoped game got that way one reasonable addition at a time, with no single decision ever feeling like the mistake. The finish line recedes a little with each new feature, and because the project always feels nearly done, the developer rarely notices how far the goal has drifted until they're exhausted and the game still isn't out.

Treat scope as something you actively decide rather than something that happens to you. Write down what the finished game contains, make every addition a conscious trade against that, and keep most new ideas in a backlog where they belong — because a small game you finish beats a large one you abandon.

Measure before you optimise

Intuition about what's slow, what's confusing, or what's driving players away is usually wrong, and acting on it wastes effort on problems that don't matter while the real ones persist. The developers who improve their games efficiently are the ones who measure first — profiling performance, watching real sessions, capturing actual errors — and let the data set their priorities.

It's slower than trusting your gut, but it's the only approach that reliably improves the game instead of just changing it. Find the biggest real problem, fix that, and measure again, rather than optimising guesses.

The first impression is most of the battle

More players leave in the opening minutes than at any other point, which makes the first few minutes the highest-leverage stretch of the whole game — and also the part the developer can least see clearly, having played it a thousand times. What feels obvious to you is often confusing to someone seeing it fresh, and that gap quietly costs you players before they ever reach the good part.

Get the player into the interesting part fast, let them feel competent quickly, and watch first-time players go through the opening without helping them. Nobody quits a game they're enjoying, so making the early minutes land is most of the battle for retention.

Small and finished beats big and abandoned

A folder of impressive unfinished projects teaches far less than a single small finished one, because finishing is where the hardest and most valuable lessons live — the unglamorous final stretch of bug-fixing, polishing, and shipping that ambitious abandoned projects never reach. Each completed game, however modest, builds the finishing muscle and the confidence that make the next one achievable.

So resist the pull of the dream project until you've shipped a few small ones. Scope to what you can actually complete, finish it, and let the experience of shipping make your bigger ambitions realistic.

Trust behaviour over opinions

People are unreliable narrators of their own experience — they're polite, they rationalise, they suggest fixes that miss the real problem. What they do tells the truth that what they say obscures: where they hesitate, where they get stuck, what they ignore, where they quit. The most valuable feedback is usually the behaviour you observe, not the opinion you're offered.

This is why watching beats asking, and why real data about what players actually do beats any amount of speculation. When several people stumble at the same spot, that's a problem worth fixing, regardless of whether any of them mentioned it.

The CPU waits on memory, not math. Lay out hot-path data contiguously and process it in batches to feed the cache.