Recommendations
The scrolling left-right tape of recommended products is quite familiar to all of us, web shoppers. However, recommendation function is not limited to web shops. It can be easily applied for B2B companies.
What we offer?
Every company does actions at the market, and therefore offer something to their potential clients, for example:
- product or solution,
- marketing asset,
- event theme,
- demo type...
Therefore, recommendation function can be built for various objects, based on various spaces.
Who does offer?
There are various agents who offer something. Please note, in this paradigm, they are the potential users of the recommendation function, for example:
- Marketer, planning the campaign, and thinking over event topics;
- Web site, generating page view online depending on the visitor;
- Seller / Key Account Manager, preparing visit/ call to a client,...
For whom this offer?
The proposal, or offer can be addressed to various subjects. For example:
- To a person representing particular company
- To a visitor of web site
- To a contact of a target group
Let's shortly review the simplified recommendation function.
Recommendation for seller
Yes, every seller knows better what to talk about with their client. Or, perhaps he forgot something? Recommendation function can help to refresh and prepare better proposition.
Input data: all previous purchases of all clients.
Data is organised in a huge rectangular matrix, than it is processed to identify the most popular corporate behaviour patterns. We can use Singular Value Decomposition of a matrix. As a result, we get a very simple function that answers the question: having all previous purchases for the company X, what would be the most probable next purchase (with three ranked options).
It is a good example of AI helper. The Seller gets consultation from this function, and no-one insists he must use results.