As detailed before in various use cases, personalization is one of the very important aspects in the digital world.

-Wikipedia

This definition clearly defines what is meant by personalization. In the modern day, customers expect them to be known by the business through  a variety of channels and in a consistent way. They need the business to understand when they are browsing on a browser on a desktop or on mobile phone, and when they actually come into your store or organization, through various IoT mechanisms.

Web personalization can be considered as looking at behavior data (collected from the user's browsing history and other client-side aspects) and real transaction data (actual transaction data that the users have done with your business application such as customer details, order details, payment details and so on) from the Data Lake. Usually this transaction data is not owned by the Data Lake; rather it is flown in from various transactional source systems and consolidated in a holistic fashion. When browsing online, this data is used to actually personalize the site itself or the recommendations. These transactional applications are usually quite limited in various analyses and are always in silos (don't  really talk with other business applications and don't know whether the other applications would have already interacted with the customer in the past) and because of this it doesn't clearly don't have a holistic notion of each customer. It is this aspect that a Data Lake can help and can be used to avoid sending in a second recommendation of a product if the customer has already declined an offer of the same nature in one of the other channels. Say, for example, if in mobile app your business would have pushed a campaign for a particular product via a mobile app, and the customer has declined or not interacted. This can be used as a means to avoid sending in same offer when the customer interacts with the business using their website. Rather they could push another offer and see the interaction level of the customer. This can only be possible if there is a holistic view of  all the data in one place for various analysis purpose, which then can be used in various transaction systems.

With more channels expected to come up in the near future, with a huge digital footprint for each customer, this is a must and the only distinguishing factor for any business; it should not be delayed in any way.