Over the Christmas break, one of the many things I read was Hannah Fry’s book about algorithms, Hello World. It’s a cracking exploration in non-technical terms about the world of algorithms, artificial intelligence and machine learning.
Early on in the book, Hannah describes how all algorithms do one or more of the following things:
- Prioritization – creating an ordered list
- Classification – picking a category
- Association – finding links
- Filtering – isolating what’s important
It’s a wonderfully simple explanation of what algorithms (and, to an extent, automation systems) are all about.
It’s got me thinking. What if we asked people in organisations to think about how their organisation operates in terms of the algorithms that are necessary? Not to actually define them in detail, but just to talk about the acts of prioritization, classification, association and filtering that are necessary for the business to operate. A lens to look through that helps to bridge between the worlds of technology and not-technology.
So, for example, what are some of the algorithms that are important for a Housing Association to be able to operate that we could identify? Here are a few…
For Prioritization, what building works should we prioritize to both maximise the safety and comfort of our customers and also get the best return on our investment?
For Classification, how can we understand from the report of a fault from a customer what sorts of building tradespeople might need to be sent out to deal with the fault effectively?
For Association, can we spot if a certain pattern of requests come from customers (or from their neighbour) might mean that we could take preventative action to help stop further problems arising in the future?
For Filtering, can we identify if certain patterns of payment (or not-payments) mean that we should provide additional support to particular customers?
It feels like Hannah’s four categories could form the basis of a useful exercise in business analysis with groups within a business to help describe how a business works in ways that can be more easily translated into technology.
Open up with some sharing of what people think algorithms are.
Then explore the categories with general examples (for example, the one’s in Fry’s book include Google Search results ranking for Prioritization, Netflix recommendations for Classification and so on).
Then get people to identify some more examples from outside of their work experience.
Then get them to identify those that form the basis of how either they or other parts of the business work and in turn understand whether those rules are currently in systems, processes, heads (or are just nice to have).
I’ll be experimenting soon…