Lots has been said about the business of Payment for Order Flow over the past five years, and even more has been discussed in the past week as Robinhood has come under fire once again.
PFOF has become one of the businesses of the retail brokerage, and very much the transactional component for zero commission brokerages. The way it works is you send an order into your brokerage. The order then gets routed to market makers, and potentially combined with other orders and filled. The broker has an arrangement with the market makers to get paid a small fraction of a penny per share to give them that order flow, which market makers need in order to make any money providing liquidity.
The proponents say that this is good for smaller brokerages as it gives them benefits of scale and market makers competing of their order is good for the customer. Plus they need to prove best execution to the regulators, so there is nothing to worry about here. The worst critics call it "front-running" and very anti customer. They say by showing your 10 share purchase of $GME to Citadel, they are going to front run you, which technically is the case as they must get the stock to fulfill your order. I fall in the former camp, but I understand the tendency to be in the latter. It feels wrong.
What if we just looked at PFOF as your standard ad business. The brokerages are your Supply Side Platforms (SSP's) - your publishers who attract trades (the pageviews, impressions, clicks in this case). The market makers are the demand side and the tools and the order routing system is basically a DSP - ad tech for trades.
In this world, PFOF is merely ad revenue. The brokers provide purchase intent to people who supply stock to the their users. But what about front-running. Let's examine a case of an Amazon promoted listing. Let's say that Amazon lists an item that is a completely commoditized product where there is not infinite supply even in the long run, and there are many sellers who provide it - let's say its a Tendies T-shirt that only has a limited print run. At this point the only differentiation between Tendies T-shirts to the customer is price.
Amazon wants to give you the best price, because that's what Amazon tries to do. In turn, they create a Sponsored listing where the sellers can pay a little bit of money to be the first listing for your search for a Tendies T-shirt. The point of such a listing is that at any given moment each seller could have different cost functions and ability to deliver a Tendies T-shirt as their relationships with the supplier of Tendies T-shirts could be very different. Some can sell the Tendies Shirt at lower prices than others and this may be a temporary condition. Those that could afford to compete on price at this moment will bid for a sponsored listing to make sure that they come up first, and more likely to make a sale.
This begs the obvious question - why should Amazon charge a sponsored listing at all doesn't that hurt the customer? They could just sort all offers by lowest to highest. Presumably the customer would pay a little less for a Tendies T-shirt if the seller didn't have to pay Amazon such a fee, and they just gave you the lowest price up front. The simple answer is Amazon is a business operating in a commoditized market therefore it needs to find a way to make money somewhere, and it has enough order flow to do so. The more complex answer is that this ad slot represents an option of sorts, and actually enables it to attract more sellers. Instead of having to compete on price with other sellers with an unknown order flow, risking selling lots of inventory way too cheap or getting scalped by another seller, they can all compete to buy an option to show up first should that order flow come in. That way they all can compete more on pricing that option vs. on the price of the underlying shirt, which in turn makes them more likely to list their inventory in the first place.
Of course, the reality on Amazon is products are not fully commoditized and can be produced in the long run to match demand. We might have seen the closest thing to this, during the early stages of the pandemic we did see some of this at play when we ran out of certain items such as hair clippers and you saw many manifestations of the same core, commodity "Alibaba" product listed across Amazon. The sponsored listing usually was the one that could get delivered to you in a decent enough time, which makes sense. In normal times the search algorithm interplays with the sponsored listing to both make people want to buy through them and sellers want to list through them without getting undercut through price in the long run, but compete on price if they see an opportunity to do so. It's not the perfect market by any means but it works and most people don't consider it inherently evil.
PFOF is the ad model for brokerage. People are right to be suspicious of it as most people don't like this model for media, but I would say its more like Amazon or Google ads than traditional display, or even super well-targeted Facebook or Instagram ads. The more troublesome thing to worry about is that for most of these companies it isn't enough alone to pay the bills, and increasing supply of commission free trades could bring down rates like the ever increasing supply of page views brought down ad rates. The practices brokers may engage in in order to cover such a gap, or their long term inability to find something viable to do so should be more concerning than PFOF itself