Eg if "search" quality is a competitive differentiation, make sure you implement search yourself. Your goal is to beat the rest of the market in search quality so buying a commodity search product that anyone else can buy doesn't make sense. You are directly rewarded for paying the extra NIH cost to go the extra mile for your customers.
For everything else, there's little benefit to paying the extra cost of NIH. An excessively fancy/powerful support ticketing system, for example, isn't going to help you get more customers. Few people might care how much nicer the checkout/payment page is than on other products, so you can outsource that.
Most areas in real life don't benefit from extra quality either. For eg, the extra quality you might get from hunting/farming your own own food makes little difference to most people, so most choose to outsource that.
This sounds similar to the NIH syndrome in some software companies. I’m still trying to figure out when it makes sense and when it doesn’t. In software, I tend to err on the side of NIH. In life, I am almost never DIY. I suspect I’m out of balance in both regards.