Using word puzzles to replace an application
A friend and colleague called my attention to an article describing the Strangler Fig: a plant that starts its life in the branches of another tree and grows roots downward, strangling its host in the process.
The article compares the Strangler Fig to replacing an application. I’ll link to the article in the comments, but the basic premise is that rather than doing a massive, monolithic replacement with a hard cutover, it’s better to replace a system one little bit at a time. The new application starts as little bits in and around the old application, and eventually replaces it entirely. That approach is cheaper, more agile, and far less risky.
It’s a great analogy, and I have another one that I’ve used for years. There is a word puzzle, sometimes called “Changeaword”, that provides you with a start and end word. Your task is to morph the start word into the end word, by changing only one letter at a time. The catch? Each letter change must result in an actual word.
So you can do “PLATE > PLANE > PLANT > SLANT” but you can’t do “PLATE > PLANE > SLANE > SLANT”, because slane isn’t a word.
Your job in planning a system replacement is the same: make one change at a time (or realistically, a few changes), and each change must result in a viable, usable product. Plan out where you want to end up, in broad strokes, and establish a game plan that makes smaller, discrete changes until you’ve reached your destination.
Creating a whole new product from scratch while still operating the existing one is fraught with risk, logistical challenges, and expense: you'll need two wholly separate teams, the new product becomes a behemoth that can't respond to changing needs, and either the existing product has to become stale or the new product has to keep up with changes to the existing one, while still being built.
On top of that, users will have to deal with a wildly new environment overnight, as everything they know about the product changes at once.
Gradually replacing pieces of the original product is not easy and carries its own challenges. But it will cost less, require less detailed project tracking, and be gentler on users.
Much like the Strangler Fig and the Changeaword puzzle, your ultimate goal is not to create a whole new application and then destroy the old one. Rather, you want to slowly grow the new application while supplanting bits and pieces of the old. Until eventually, all that’s left is the new one.
👋 I’m Laura Stringer.
💻 Product Management Expert for B2B SaaS.
📈 I help with strategy, roadmapping, process, experience, & startup growth pains.
☕ Need help or not sure “what now”? Hit me up and we’ll talk through it!