Introduction to Website Personalization
Website Personalization is the core product behind website recommendations and personalization modules.
We believe that the Website Personalization should be key to your client’s business for several reasons. With the Website Personalization it is possible to inspire and recommend personalized products and content to every visitor, optimize sales, improve product selection, maximize cross-selling, increase conversion rates and significantly enhance the user experience on a website.
The recommendations provided by the Website Personalization modules are a result of live tracking, visit and shopping behavior. Therefore, we aim to implement the tracking script as soon as possible, so we can gather data about the customers’ behavior down to item clicks, views and purchases. The tracking can be set up server side, client side or in GTM. Once the tracking has been implemented, the recommendation engine starts and after roughly one week, it provides valid recommendations. Over time, as the data-foundation builds, the quality of our recommendations will improve as the algorithms become more intelligent.
With the tracking script in place, it is equally important to ensure that click handlers are placed in the right way. These are Raptor’s way of detecting when a recommendation is clicked on. With this in place, it is also possible to track direct and assisted revenue by Raptor as well as performance indicators such as conversion rate, average basket size, and revenue per visit for Raptor users vs. non-Raptor users.
Product catalogue feeds are also set up in the early stages of the onboarding process. The feed formats can be XML, CSV or JSON. To get the product recommendations off to a fast start, it is often desired to upload a POS-file from a chosen historical period.
To get the most out of your personal Website Personalization setup, there are different tuning parameters we can use to optimize your output. The many different options for tuning and configuration of the modules include opportunities for boosting, tuning, filtering and adjusting the output of the modules. These include among others serendipity tuning, o-data filtering and merchandise boosting.
Know your Business Model
The range of opportunities with the different modules is plentiful and therefore it is worth keeping a few things in mind before selecting and implementing modules on a website. Whether a client has a B2C or B2B site the main setup is the same, however, the recommendation methods can take different approaches.
With B2B clients, we generally see that the visitors oftentimes are returning customers who know the webshop. Therefore, the objective of the recommendations could be to recognize and showcase the most often bought products made by each individual customer, because they often return to buy the same products. Furthermore, it could be an objective to showcase and inspire with products from the shop that they didn’t know was there and maybe bought somewhere else.
A B2C webshop would be more likely to showcase personal offers to the individual customer on the front page and also encourage customers to increase basket size by inspiring and recommending similar or related products on the product page, power-step or in the basket. However, this knowledge about B2C behavior can also be very useful for B2B clients. Tracking a B2C inspiration site will provide useful information regarding which products should be placed in stock. The trending products and attributes as color, size, material etc. will be listed in the module of “GetMostPopularRightNow” or “GetTrendingBrands”. When B2C customers are frequently looking at specific products, B2B can use this information, which indicates which products might be good to get in stock. Reusing the example with pots, the information from the B2C inspiration sites can indicate what specific brand or material is most trending in a specific period.
Once the Website Personalization has been implemented, it is important to revisit your Raptor setup for fine-tuning to ensure that the most relevant products are always presented in the output.