Why do banks continue to build non-differentiating components?

What could Investment Banks learn from Apple’s approach to the supply chain? How can they use the same approach to drive return on equity and shareholder value?

In any industry there is a supply chain. That supply chain consists of two sets of components, those that differentiate your product and those that do not.

The product that you sell is the result of all the steps in that supply chain delivering into the hands of a customer. That could be a product or service, and increasingly, combinations of both.

Apple’s iPhone is a classic example of this. The supply chain components that differentiate the iPhone, whether you love it or hate it, are the overall design, the user interface, the closed ecosystem, the unique combination of hardware and software components that manifest in a unique user experience.

The supply chain components that do not differentiate the iPhone include the memory chips, the screen, the plastic and metal that bespoke components are made of, the logistics company that delivers it to your door, and more.

There is a simple reason for this. Apple, in common with many other technology companies, has made the strategic decision to concentrate on the areas where it adds value, and buy in components and services where there is no intrinsic advantage.

This has made Apple one of the most successful companies in the world with a cash pile of more than $285 billion dollars. It is also a sign of the maturity in that sector that there are reliable suppliers of these components that allow Apple to focus on what makes it special.

The suppliers to Apple make their money by creating screens, fabrication facilities etc. and them amortising their large investment across multiple customers. Whereas, for a single customer, it would not make sense absorb all these costs for a limited number of products.

Investment Banks operating in the electronic trading space are in a similar industry at a similar stage of maturity but because the sector is smaller, it has only seen the recent evolution of component vendors who are mature enough, with good enough products, to allow investment banks to take a similar approach.

By learning from the likes of Apple the secondary trading industry can improve profitability, increasing RoE and improving CI ratios.

Let’s look at FIX engine space as an example

The initial FIX engine vendors were all in a race to develop functionally complete products. But in this race, and associated grab for market share, they neglected the non-functional elements that are increasingly important to users, such as ease of use, high throughput, and low latency.

Regrettably many of these products are still on the market today, touted by PE backed vendors, who’s investors are simply trying to milk customers for as much cash as possible, many of whom believe it is just too hard to change.

This means that rather than buy in a FIX engine, an investment bank could differentiate itself by creating its own FIX engine that addresses these non-functional requirements.

The same has been historically true of much of electronic secondary trading supply chain, from FIX Engines, through market infrastructure components such as native market data and trading gateways.

What are the components of the electronic secondary trading supply chain?

To a ‘buy-side’ trader at an institutional asset manager a typical ‘sell-side’ supply chain for the Algorithm, sold by an electronic sales trader at an investment bank, probably looks something like this:

  1. The buy-side is enabled for FIX connectivity to the sell-side.
  2. The buy-side and sell-side collaborate to ensure their usage of FIX is common, whilst FIX is a standard there are still many dialects, and other variations such as choice of symbology.
  3. The buy-side OMS routes an order to the sell-side using a FIX connection, either as a direct connection of over an order routing hub.
  4. The sell-side receives said order and performs any mapping and normalisation that has been agreed earlier.
  5. The sell-side passes the order into its algorithm.
  6. The algorithm breaks the order down into many smaller child orders and determines the size, timing or other order attributes for each of the children.
  7. The orders are passed down to an execution venue, often in a binary protocol.
  8. As child orders are executed, they are reported by to the gateway that originally delivered the order.
  9. Child order executions are reassembled and mapped back to the original order.
  10. FIX executions are reported by to the buy-side against the parent order.

There are typically other steps that have been omitted from this simplified narrative such as state mapping between the FIX and Exchange Native Protocols, risk handling etc, but to make the point this simplified illustrative flow is enough.

How many of these 10 steps are differentiating for the investment bank. Only step 6 is IPR that the investment bank owns and cannot be achieved most cost effectively by purchasing a vendor solution.

The other 9 steps could be described as “core infrastructure plumbing” – every bank needs it, it is expensive to do well and it is non differentiating.

Firms like Rapid Addition replicate this spend, create the best possible solution to the challenges presented in creating this infrastructure, that must work, must work well, must be reliable, and amortise their costs by selling it to multiple banks.

Building infrastructure of this type is not cheap – RA knows this well given the time and money of investment and research that has been put into the electronic trading infrastructure.

The economics

We discussed many of the emotional and self-belief issues in the earlier blog, Developing Trading Algorithms – Buy Vs Build but now let us take a look at the cold hard economics and a very simplistic example (but it goes a long way to illustrating the point)

Banks are owned by shareholders – shareholders want a good return on their investment; unnecessary expense in building expensive complex software will damage the bottom line.

RA build software – RA does it very well, a team of experts in infrastructure development – given time to research the right technology, the right methodologies and time to write great software with awesome performance (advert over – for now).

Say our software has cost us £5 to develop – with that we have taken the risk on never selling it, but of course we do – we don’t sell it for £5 because if we did our buy vs build arguments fall flat. So, say we sell it for £1 – for which we support and keep it up to date with upgrades, new functionality and new technology enhancements. You have just saved £4

This not only has a direct impact on potential money available for further differentiating developments that could have revenue enhancement potential, but it could, if the savings are not fully spent elsewhere, also impact on the bottom line for the firm and hence on the Return on Equity for the firm as a whole.

There is also the question of opportunity cost to consider. A vendor, such as Rapid Addition, has their components ready to go, tested and just needing your differentiating business logic to be added. To build the same infrastructure could easily take your inhouse team two years to create something that ultimately does not differentiate you from your competition, and the ongoing maintenance of it will delay projects that could differentiate your firm.

In conclusion

Whilst there may be times when there is no option to build due to outside factors, banks should focus on what they do well, providing services and shareholder value.

Buy where you cannot add value over the market place offering and leave the non-differentiating components to the experts.

So, to answer the question we asked at the beginning of this piece, we don’t know!