Hi Dennis! Short answer: to me, ad-delivery doesn’t seem easily optimisable by decentralisation (existing solutions work well). Trade execution does.

The piece is more concerned with the latter. Matching proofs with orders. In the scheme presented, where relayers are those who suffer from “unfinished trades” (put up gas fees of injecting orders), I assume they would be the most interested in making sure ads actually get delivered and the proper proofs are generated, fast enough (i.e. the optimisation you talk of) - so they complete their trades.

Ads for me have been showing up pretty fast on destktop & mobile lately =/, I don’t see how 0x could help this piece of the stack perform better. The problems I see, that can be tackled, are more on the aggregation of inventory and trade itself.

Let’s consider such a world where there’s a relayer just for bidding on video overlay ads. I feel it could maintain an IPFS node optimised to work as an ad server, and if its implementation happens to be the most reliable one around, it’ll probably outcompete other relayers who are on the same field. Or it can pay a traditional hosting & CDN service to handle advertiser’s media, and if it works better, its order book will have higher activity than competitors. And so on: relayers can choose infinite strategies for whatever they’re delivering beyond the minimum functionality of a relayer (trades!). The most effective strategies will accrue volume.

In ads, borrow/lending, commodity auctions, microinvestments, or whichever verticals where relayers come to flourish on, specialisation will be a prerequisite for survival.