Get this from someone else? Click here to subscribe yourself and get future content delivered directly to your inbox

It feels like we’re at an interesting inflection point in the capital markets discussion of build vs. buy. Ten years ago, maybe even just five years ago, the industry was split into two buckets of firms: those who built all the technology they needed in house, and those who didn’t have all the technology they needed, because they didn’t have the budget to build it in-house.

The largest brokerages in the world leveraged their size and global platforms to get economies of scale and build the solutions they needed in-house. These could range from trading tools and applications, through to compliance solutions, and CRMs. If you didn’t have the budget to build it in-house, you had to buy something off the shelf, or make do with ad hoc solutions – but the unfortunate reality was that there weren’t too many ready products available off the shelf. The build vs. buy discussion was usually relatively straight forward: if you want it to work well, you build it. If you need to compromise, or don’t have the budget internally, you buy it.

Fast forward 5-10 years, and we find ourselves at the start of 2020. There are two major themes that are radically re-framing the build vs. buy discussion. First, margin compression. In an industry that is increasingly under margin pressure, having massive technology teams building everything bespoke becomes more challenging. Firms are being forced to re-evaluate what is a core offering and truly differentiates, which they should continue to build, versus what isn’t a differentiating product, and what they should think about replacing with an off the shelf solution.

The second theme is perhaps more important though: machine learning (data scale) and network effects are making the build discussion ever more difficult. Applications built in house will never have access to the same scale and scope at market solutions. It’s just a reality of building something. The potential client base is limited to one. In a world that is increasingly driven by machine learning and artificial intelligence, that means that those in-house solutions will never be as ‘smart’ because they will never reach the same scale as a market solution. Build vs. buy becomes challenging when some parts of the problem can’t be solved with just resources – they need scale.

The second part of that challenge is the network effect. As products are reaching more data scale, they are also becoming networks. Networks inherently benefit from increasing the number of participants, as they centralize value for participants. Look no further than an Exchange for an industry example. Thus, it is very hard for each firm to ‘build’ their own network.

Reframing the build vs. buy conversation

Around a month ago, I was in with one of our largest prospective clients (who is now a client) and was having a discussion with one of our sponsors. The topic of build vs. buy came up, and they asked me: how do you think about the trade-offs?

I thought to myself for a moment, then framed the question in a different way, which I think really captured the thought for me:

“If you build what we have, you’re essentially choosing to start a competing company who will only ever have one customer. Sure, you get to control the product roadmap, and you can make it a little more bespoke, but you’ll never have more than one customer (yourselves) so you’ll never get any scale, and you won’t benefit from any network effects. If you think this is a core competitive advantage or differentiator, I would encourage you to consider that option, but if you think this is infrastructure or a commodity, then I would consider you to re-evaluate. It’s like deciding to build your own airplane versus building your own airport as a plane manufacturer. Building your own airplane makes sense. That’s your differentiation. Building your own airport doesn’t. That’s infrastructure.”

That framing and analogy seemed to really resonate.

It makes sense though: how can you expect to build a competitive product, when you are inherently limiting the customer base to just yourself, and then expect to do so for cheaper or more efficiently? There is no way that Airbus could afford to build their own network of airports. Instead, you can leverage the existing network – thus splitting the cost across all the clients, but also get the scale and scope when it comes to intelligence and machine learning.

The build vs. buy conversation continues to evolve as the industry evolves. More and more we’re encouraging our clients to think about whether or not the product in question provides a unique advantage or differentiation. Should you build core infrastructure, plumbing, and applications in house? Probably not. Should you build intelligent recommendation engines, analytical tools, or other personalization engines in house? Probably.

It feels like we’re reaching that tipping point where the conversation is starting to tilt away from “build” in the majority of cases, towards buy. The true value will come from stitching together all the ‘buys’ into one central view in house, and providing that truly unique analysis and holistic view of all the data sets and insights together.

Perhaps the conversation will evolve to “buy, then build.”

Blair Livingston
CEO
Street Contxt