An algorithm is simply a set of instructions for solving a specific problem. It is commonly associated with math or computers, but applies to all problems. A troubleshooting guide is a form of algorithm, as is a recipe. Algorithms make life easier by standardizing the method to solve a problem, and help us avoid having to reinvent the wheel every time we encounter some obstacle.
To truly be an algorithm, the set of instructions must deliver the same results for the same set of conditions, each and every time.
One of the most famous, most complicated, and most secret algorithms that you are probably using every day is in your search engine. (That algorithm is processing constantly changing data, though, so you will see some variations in the results. The variation is due to the fluctuation of millions of inputs, though, not the algorithm itself.)
In Lean, a process is an algorithm. It is a set of instructions that, when followed, should provide a predictable result. Common Lean examples involve tools for calculating kanban determining the number of operators required in a work cell.
Because of the up-front cost of developing an algorithm, they are most useful for high volumes of problems, or when the result is critical.
One pitfall of algorithm use is when they are misapplied. The instructions are based on specific conditions. Most algorithms don’t work when the conditions the instructions were designed for don’t exist.