Whoever decides transaction order within a block can extract value from that choice. That fact reorganized how Ethereum blocks are produced.
Ordering has monetary value
Transactions in a block execute sequentially, and the result depends on sequence. A trade executed before another gets a different price.
Someone who observes a pending large trade can place their own transaction ahead of it and profit from the price movement it causes. Arbitrage between venues works similarly.
This value is available to whoever controls ordering, which under the original design meant the block producer directly.
Concentration was the danger
Extracting this value well requires sophisticated software, low-latency infrastructure and continuous market monitoring. These are not things a home validator can provide.
If capturing it required operating at that level, small validators would earn systematically less and the network would consolidate toward professional operators.
That outcome would undermine the distributed validator set the network depends on, which is why the problem was treated as structural rather than as a nuisance.
Separating proposal from construction
The response was to split block production into two roles. Specialized builders assemble complete blocks optimized for value, and validators simply choose the most profitable one offered.
A relay sits between them, holding the block contents until the validator has committed to the header. This prevents the validator from stealing the builder's arrangement.
The validator needs no sophistication at all. It accepts the highest bid, which lets a small operator earn comparable rewards to a large one.
Users can route around the public mempool
Private transaction channels let users send transactions directly to builders rather than broadcasting them publicly, so the intention is not visible in advance.
Some of these arrangements return a share of extracted value to the user whose transaction created it, turning a cost into a partial rebate.
Adoption of private routing has grown substantially, which changes what remains visible in the public mempool and complicates analysis based on it.
The market has its own concentration problem
Solving validator centralization created a builder market where a handful of firms produce a large share of blocks, because scale improves the ability to bid.
Relays add another intermediary that could censor or fail. Their number is small and their operation depends on continued voluntary provision.
Research into designs that distribute these roles further is active and unresolved. The current arrangement is understood as a working compromise rather than a settled answer.