Strother Martin in Cool Hand Luke.

When a lot of people mention Information Theory (Claude Shannon's contribution to C20), they think of comms channel capacity and number of bits per second. What Shannon was developing at a deeper level was Information = what changes beliefs and expectations (in the receivers, ie, other people). Specific beliefs and expectations, not how many words it was, but what influence it had. I'd say comms is (are) the skills of achieving that influence. Failure = a skills gap or an awareness gap about the receiver. Tom.

The important question to ask is why do we have conversations in the first place - why do we need them? To build strong reliable relationships - to belong. Relationships don't suddenly blossom, they blossom one conversation at a time over time. Relationships don't suddenly fail, they fail one conversation at a time over time. So the answer to your question - why do communications fail so often - is because conversations need to build on each other, be relevant, be consistent, be true, be worthwhile. Without these elements, failure will loom.

Quickish comment about modelling and analytics: Modelling, data analysis and such don't exist in isolation from humans, perceptions, insights and goals. Models (etc) are there to reflect relationships that can be found by measurement in CRM or survey data. The motivation to even look for those relationships come from people making decisions about running a business. It's for their benefit. Lots of relationships unfortunately (especially responses to unusual situations) are very hard to model, but reasonably easy to guess about. Those guesses improve with experience and having "a good way to look at the situation". Indeed, relationships which are easiest to model are often those a client says "I knew that already". Modelling can sharpen precision, and reveal sneaky but important things in extensive data that aren't apparent to unaided humans. The very least that can be said is that ignorance is the enemy. Tom.