I’m conscious that relevancy of content is of paramount importance when developing ‘stickiness’ and customer interaction of a website (or indeed a shop I suppose). Having customers trust your site is obviously a priority and trust must partly be built by giving customers relevant content. That is to say, customers easily find what they’re looking for or better still, what they’re looking for is displayed in front of them without requiring a search taking place.
Yet relevancy may differ by customer. Not only that but relevancy changes pending ones mood and feelings. How do you introduce context to content delivery for every individual customer?
I’m beginning to understand more about AI and how theory is applied to business. I am slowly beginning to understand the importance of building a good recommendation engine and the complexity involved. Indeed, I am on the very edge of understanding but I aim to apply a material amount of time to gain a better understanding.
This area is important for a number of reasons:
1. Conversion, generally on websites is 01.%. So, 99.9% of traffic chooses not to purchase, why?
2. If delivering relevant content products and/or content improves conversion then understanding customer demand is the goal
3. I want to make life easier for people, that means building a service which delivers relevant content in each individual
In order to understand each individual one needs information about that individual and as much information as possible. That information can be applied to form a learning algorithm, easier said than done. There are difficult questions then about how much information? is there an ethical line one should not cross? what about privacy, is that impacted? I can’t answer these yet, maybe I will never be able to.
There is a clear paradox though, on the one hand a business must deliver relevant content to each customer. Yet to understand each customer one needs as much data as possible about that customer who might not want to give the data.
Even if one manages to gather lots of information its not enough because a learning algorithm must be built, AI. Even no I think I am in over my head. I’ve purchased two books to help me understand: