The Power of Diversification

Finance has been a big part of my life. My father was a banker, my sister owns a money management firm, and I was a finance major in college. At one point, I managed the website for the world’s largest futures exchange. One of the main things I have learned about finance and investing is the power of diversification. You have many choices if you’re fortunate enough to have enough money to invest. You can invest in stocks, bonds, real estate, crypto, etc. While it is tempting to go “all-in” on one stock or type of investment, it has been statistically proven that spreading your money across multiple stocks and investment types outperforms putting your eggs in one basket. While you sometimes hear about the person who put all of their money into Apple, Tesla, or Nvidia and made a fortune, there are likely ten times as many people who put all of their money into one stock and lost most of it (those people are probably quiet about it!). In the long run, diversified portfolios tend to outperform individual stocks.

So what does this have to do with technology? When your organization purchases software, it’s a form of investment. The organization agrees to pay a vendor (say $1,000,000 per year) with the expectation that they will receive value. Just as investors have investment options, organizations have different technology architectural options. Organizations can purchase “suite” platforms with many bundled products or a composable/warehouse-first approach that uses multiple vendors for various functions. The following will share thoughts on how I see this technology architectural choice through the lens of investment types.

Many organizations pick one vendor or “suite” to do everything they need. The “suite” approach is especially popular in my particular area of marketing, where Salesforce and Adobe command most of the market. I view the “suite” approach as similar to putting all of your money into one asset type (for example, a home). When organizations go “all-in” on a suite, they are beholden to that suite (for many years!). Will the suite work? Will the suite innovate and add new essential features? Will the suite vendor provide good support? Will prices go up a lot each year? When you go all-in on a suite, you place a pretty big bet. Suites can also take years to fully implement and often require outside consulting expertise due to their complexity.

The other architectural option is the composable/warehouse-first approach. In this approach, organizations choose a cloud data warehouse and then add different “composable” vendors on top of the warehouse for various functions. For example, an organization might put Databricks at the center and then add Hightouch for CDP, Iterable for email, Kubit for analytics, etc. In a composable/warehouse-first architecture, all data goes in and out of the warehouse. Therefore, vendors act as the last mile of a specific function. In this model, there’s a lot of value in both the warehouse and the vendors connected to the warehouse, but the latter can be changed more easily than in the suite model. For example, if you’re using the Salesforce suite and don’t like its email tool, it’s more difficult to plug in another email vendor since you must integrate that vendor and the suite. However, if you use a composable model, swapping vendors is easy as long as the new vendor can work in a composable manner from the warehouse. Diversifying your tech stack through composability may require some work to pick and set up multiple vendors, but in my experience, this pales in comparison to setting up monolithic suites.

Returning to the investment analogy, using a composable/warehouse-first architectural model is a form of technology diversification. While you are placing a big bet on the cloud data warehouse you choose (Databricks in my example), all of the other SaaS vendor bets are relatively minor. Instead of putting a lot of customer data in multiple SaaS tools, you are placing most/all customer data in the warehouse, so if you ever need to swap SaaS tools, you’re just swapping features, not data. Even if you eventually need to switch data warehouses, that can often be less challenging, costly, and time-consuming than switching suites. You should have the flexibility to try new tools to fuel growth and beat your competitors. Placing several smaller bets that can be changed at any time is similar to purchasing many different stocks and index funds that you can buy and sell at any time, versus tying up most of your money in an illiquid asset like a house!

Another benefit of the composable architecture is pricing power. When you go all-in on a suite, the vendor knows they have you and can raise prices as they wish. The fear of switching costs is a significant factor. When you use a composable architecture, each vendor knows they are expendable, giving you pricing power and leverage in contract negotiation. The fear of being replaced is also a key motivator for composable vendors to deliver value and continue innovating.

So, when choosing a technology, consider the parallels between investment diversification and technology architectural diversification. I am a proponent of technology architectural diversification, especially in today's rapidly changing landscape. Over the next few years, AI may create an environment where new technologies arise from nowhere. Consider diversifying your risk and keeping your options open through a composable technology architecture.