Skip to content
vecton product recommendation models
Black Friday 2020 will significantly differ from the other Black Fridays in the previous years. COVID-19 pandemic effects negatively affect the offline commerce yet they favor sales via the internet very much. Sales growth scale will be higher in numerous different industries comparing to the previous years. How to use this moment not only for new customers acquisition but also to improve the loyalty of the existing customers to ensure profits on the highest possible levels? Product recommendations are one of the tools that should be used to achieve these goals.

Automated product recommendations, personalised or not, are an indispensable element on the customer shopping path. Majority of the recommendation tools have huge capabilities of their algorithms configuration. You can find some tips and inspirations concerning which modifications in the recommendation algorithms configuration should be added before the Black Friday.

Consider if you want to promote these products using the product recommendations during the Black Friday:

  • Products with the lowest margin.
  • Products already purchased by the customer. It depends on the industry, for some industries (like fashion, bookstores, electronics) it doesn’t make any sense.
  • Products, the amount of which is low or unknown. There is a risk that before a customer completes the purchase the product can disappear from the warehouse. It may cause huge disappointment and frustration.
  • Products that require a more difficult way of packing, longer preparation time or any other complications in the order completion process.
  • Products that are bought by the client on a regular basis (e.g. every month). Because very likely the client will buy them in any case. You can use the Black Friday time to extend the client’s awareness through cross-sell.

Think about the promotion increase during the Black Friday for:

  • Products added to favorites list by the client. If you don’t have these lists or if they are not widely used by your customers, you can think about products that were added to cart or viewed many times.
  • The most popular products. Because the popular products with the high discounts create an explosive mix, which will have the highest conversion.
  • Products that you want to get rid of. If you don’t know where to start – filter the product catalog by the added to warehouse date and product count. The longer time in the warehouse – the higher costs and the less margin for the product.
  • Products with the highest margin.
  • Products from suppliers or brands that you want to promote according to your strategic goals.
  • Products that are suitable for a gift. Many consumers use the Black Friday time smartly and buy gifts with high discounts before the upcoming holidays.

Offer the highest discounts to the most loyal customers

Loyalty building is an arduous and difficult process. The Black Friday is an exceptional occasion to support this goal. If you treat your loyal customers very well during this unusual time, you will stand out from the crowd of offers and will have a big chance for your message to be remembered. You can define a segment of loyal customers (e.g. using RFM) and build a recommendation models specially for them:

  • Personalised models. You have the historical data of your loyal customers, so the personalised models will act much better in their case.
  • Models extended by products with the lower margin – because in the local customers case you can achieve other goals: increase of their purchase frequency and CLTV increase.

Recommendations prepared in this way should be sent in a dedicated email campaign or attached as an additional element of the existing campaigns in the Black Friday period. Of course, another good place for their presentation are the web pages of your e-shop.

Use the best practices

Depending on the stage in the customer shopping journey you should use the appropriate up-sell and cross-sell techniques. For example, right after the purchase completion you can send a message with complementary additions promotions (like mouse for laptop), similar products (like books in the same topic) or product recommendations by “collaborative filtering” (people who bought X were also interested in Y). Remember, that technology plays the essential role here, so you should check what is the recommendation loading time for each model, if there are any error responses from the recommendation server, and if the recommended products are displayed correctly on mobile phones, tablets etc.

To sum up

All the tactics described above may be configured with the usage of the appropriate advanced recommendation tools. If you already have integrated a Marketing Automation platform, most likely you already have many capabilities in this field. But if you don’t use these technologies yet, or if you feel that you don’t use their full potential – reach out to us!