# Crossing the Spaghetti Junction source: https://developer.mastercard.com/blog/crossing-the-spaghetti-junction/index.md Authors: Oran Cummins. It's a task every company is faced with. Even if you're a one year old tech start-up running in a green field, you'll have accrued some level of technical debt by now. However, if you're a large successful company then years (decades) of product evolution, millions of lines of code, countless customer / market customizations, shoehorning regulatory compliance and multiple acquisitions inevitably scar you with an undesirable level of system complexity. This spaghetti infrastructure can be a huge weight around your neck; slowing down new product development, undermining your competitiveness and reducing morale. The bad news is that this spaghetti doesn't untangle itself. The good news is that it's fixable but it requires prioritization, investment and a stomach for making tough decisions. Create a real imperative All too often, I hear about how executives are convinced that they have a great product roadmap but the inability to execute on technology (both build \& run) is the reason they haven't achieved large scale success. The reality is that the product / strategy (or their inability to sell it) is the reason why the business imperative isn't there. Without a business imperative, the level of investment is only enough to keep the lights on for existing stuff and add to the spaghetti in an incremental way rather than addressing the mess and making real strides towards the vision. First things first -- ensure there's a vision and strategy to support that eye watering investment. Make personnel changes It would be unfair to blame current employees on the current state of affairs, however you must accept two things: Martians didn't land and deposit this mess -- it was people working for your company who got you here It will take people to get you on a trajectory out of it but probably not the same folk you have on board today A beefed up imperative can only go so far - an injection of fresh eyes and pedigree will undoubtedly be required. Steal shamelessly Your problems aren't unique to you -- they've been solved multiple times now. It's absolutely fine to take ideas (and people) from other companies - where do you think they got theirs from? Show what good looks like The bigger you are, the harder this problem is to solve and can only be achieved over time (multiple years), no matter what the investment. An efficient way to achieve it is to create an example, a proof-point for what 'good' looks like. This way you can direct a relatively small investment towards bringing to life a reference implementation which you can now point to as what you want across all products / departments. Be careful here, a poor execution and you've got the wrong template for replication and also fuel the inevitably crop of 'Nay Sayers' who feel threatened by a deviation from the status quo. But that's why you're doing a reference -- it's to fail fast and iterate to a winning proof point that you can extrapolate from. APIs are your friend This point is probably the whole reason why I'm writing this blog. Many companies / products have external facing APIs and you see the internal complexity/spaghetti manifesting externally. With APIs, you have the opportunity to manage that façade and create the doorway that you would like to expose to the world. Obviously, you need to do work to map from those intuitive interfaces to internal ones but this buys you time to fix the spaghetti behind it. Also, internally, it allows you to create good separation and compartmentalize the effort, enabling the company to address the spaghetti one bite at a time. You could conclude that this is not for you. You could convince yourself that this level of complexity was created by a bunch of idiots and you deserve better. However, given the increasing complexity of doing business today, the speed at which we're having to operate and the inevitable legacy that builds up over many years, this is something that is pretty much unavoidable. So, if you're a technologist, you better get used to eating spaghetti or expect to go hungry.