Not Invented Here, usually abbreviated NIH, is the disposition to avoid using work that originated outside one’s own group, preferring to design and build an equivalent in-house instead. In software it shows up as teams that write their own logging library, their own date parser, or their own framework rather than adopt an established external one, often justified by a belief that the outside option is not quite good enough, not quite trustworthy, or not quite theirs.
The phenomenon was given empirical grounding by Ralph Katz and Thomas J. Allen in their 1982 R&D Management paper, “Investigating the Not Invented Here (NIH) Syndrome: A Look at the Performance, Tenure, and Communication Patterns of 50 R&D Project Groups.” Studying fifty project groups in an industrial laboratory, they found that as a group’s members stayed together longer, their communication with outside sources of information declined, and their performance eventually fell off. The NIH syndrome, in their framing, is the tendency of a stable, long-tenured group to come to believe it holds a monopoly of knowledge in its area, and so to discount ideas from beyond its boundary.
In engineering practice, NIH is usually treated as a cost. Rebuilding a solved problem consumes time that could go to the product, produces code that must be maintained forever by a team that now owns a problem the wider world has already solved, and forgoes the battle-testing that a widely used external component has received. The rise of open-source ecosystems and package registries made the cost more visible: when a robust, free, well-maintained library is one install away, choosing to reimplement it looks harder to defend.
Yet NIH is not always irrational, and the reaction against it can overshoot into the opposite failure, sometimes called “proudly found elsewhere,” where teams pull in heavy dependencies for trivial needs and inherit other people’s bugs, security exposure, and supply-chain risk. There are legitimate reasons to build in-house: an external component may be a poor fit, may carry an incompatible license, may be abandoned, or may sit at the core of a product where deep control and understanding are worth the cost. The much-cited left-pad incident, where removing a tiny dependency broke builds across the internet, is regularly invoked by both camps.
The judgment, then, is contextual rather than dogmatic. Healthy engineering reuses external work for commodity problems and reserves in-house effort for what is genuinely differentiating or genuinely unmet. NIH becomes a syndrome, in Katz and Allen’s sense, only when the decision stops being a deliberate trade-off and becomes an automatic reflex driven by insularity, the group rejecting the outside world not because it has weighed the options but because it has stopped listening.