Data Strategy ~= Business strategy

There is an inherent link between your data strategy and your business strategy, in that to quickly execute your business strategy, your data needs to be in great shape. So therefore — ideally — you know your business strategy already and you can leverage that knowledge to prioritize which dataset to improve first.

Here are some example business strategies within the financial industry:

• Improve access to data to iterate business capabilities more quickly
• Sell data
• Expose data & APIs to customers (whether internal or external) so they can self-serve
• Build a platform; let other companies build systems on top of your infrastructure

The first — ‘improve access to data’ is the hardest to measure and is often what sprawling enterprises think of. They want to move ‘faster’ but don’t necessarily have the vocabulary to say in which areas. Ideally you drill into it to narrow down on where to begin: “are you looking to build a full picture of your clients?”; “are you looking to improve post-trade operational efficiency?’, and so on. Though a company may want to make improvements everywhere at once, it is best to prioritize to avoid boiling the ocean too much.

Selling data. This has become a popular but very overhyped business model. Ask yourself if you really have ‘new’ data; or is it similar to other established/free datasets. E.g. if you’re trying to sell bond data, how is it better than TRACE data? If you’re selling swaps data, how is it better than SDR data? I won’t be covering this strategy in much detail because the details will vary depending on the use case.

Expose data / APIs. Letting customers self-serve can be the difference between a happy customer and one that wants to get off your system. The old school charging model was to charge by the eyeball, e.g. licenses, or desktop installs. This led to a perverse situation where the software vendor doesn’t want to let you automate their work because it reduces income. This model is going extinct as customers get used to being able to automate run-of-the-mill operations with Python scripts, no-code solutions and beyond. Instead, software companies need to build good data formats and high-quality APIs and charge by monthly subscription or by API hits — which will better correlate your revenue to the business value you’re providing to the customer, thereby aligning incentives between you and your customer. To hit the point home let’s imagine 2 systems: one which helps an investor decide how to position their investment portfolio, and another which executes trades. If they are unconnected without APIs, a user has to sit and physically click in systems and perform and data translation. If the first offers an API to source the ‘target portfolio’ data, it can be retried by a script and uploaded to the execution systems trading API. As organizations scale — this optionality to script capabilities that are not natively supported is MASSIVE.

Build a platform. This is the mecca of the SaaS business model. Imagine you’re a trading exchange, and you get 10 big feature requests from 10 big customers. You simply won’t be able to respond to all requests, build them with excellence, and integrate them into your architectures without picking up some unanticipated technical debt. If 10 customers all ask for the same feature, then sure, build it. But when you have esoteric requirements from different customers you become the ‘clearing mechanism’ for good ideas, and there is no guarantee you’ll recognize the best. If your data model and API strategy is great, you can let third parties plug the gaps. Perhaps these are employees at your customers’ companies, perhaps they are budding entrepreneurs who will charge for the service, but this gives an additional outlet for users to meet their needs within your ecosystem. This is basically the new version of open source; but instead of committing to a framework, a customer is committing to an ecosystem. As the ecosystem grows you let third parties earn money from the fruits of their labour, whilst taking a cut for yourself. You do need to be considerate operating this approach; you have to ensure there are enough customers for the platform to make economic sense to build. Building a platform for 4–5 customers doesn’t make sense. You need to be confident the ecosystem will attract many people, so that you can invest enough in your core data infrastructure, mostly: a) good data model, b) plenty of security, c) elastic scalability as you get more popular.

Pick your business strategy first

By picking the business strategy you’ll be able to invest an appropriate amount in the appropriate area. Below is a little cheat sheet to help you flesh out your thinking.

There is a key decision to make; what is the type of return you’re seeking. Is it immediate return on cashflow, or is it to build a capital asset (the software) that is independently valuable line item in your balance sheet.