For retail and commerce brands, it’s vital to turn website visitors into first-time customers and secure the all-important second purchase. To make this as successful as possible, every opportunity to offer relevant products is key. Whenever there is product-related behaviour, it should trigger the right product offer at the right time.
To solve this, many brands are beginning to implement a Customer Data Platform (CDP) with Deployteq. By seamlessly connecting data from a wide range of sources, the platform empowers organisations to deliver the next best product – right place, right time, and through the channel their audience prefers.
Why retail brands are investing in a CDP
Popular CDP journeys for retailers is built on the following goal:
Next Best Product recommendations model (Next Best Product)
A flexible score model that allows organisations to identify the best product recommendations based on three key drivers: seasonal bestsellers, items with high purchase intent based on shopping history, and recently viewed or filtered products. The goal of all this is to drive product sales.
The Next Best Product data can be used to personalise the content in newsletters and other campaigns. But it can also fuel journeys such as Abandoned Search, ensuring no opportunity is lost.
Goal: Next Best Product
The “Next Best Product” case leverages three models (algorithms) that are combined and weighted in a fourth. This ensures every customer is presented with personalised product offers.
| Model 1: Seasonality model | Model 2: Transactions model | Model 3: Website behaviour | Model 4: Weighting most relevant products |
| Calculate the most popular products per month/week based on last year’s transactions | Calculate the most likely products to buy per client based on previous purchases | Calculate the most likely products to buy per client based on website behaviour | One calculation per product per client to get a score |
| Top 50 products scored: 100 points for the most popular down to 2 points for the 50th | Other customers who bought product X then bought product Y | 1 product view = 20 points, 1 filter use = 5 points, product added to basket = 40 points | Formula: (Seasonality × 1) × (Transactions × 2) × (Website × 4) = Product scores |
| Most likely products to buy per client based on past purchases | Top 20 products scored: 100 points down to 5 points per customer | Top 20 products scored: 100 points down to 5 points per customer | Formula: ((Product scores from seasonal trends) x 1) x (transactional products x 2) x (website products) X4) = List with products and scores |
| Other customers who bought product X then bought product Y | Output = List with products and scores |
Priority
Data points
Must-haves
Website behaviour (products, filters), customer identification & consent, product core data and order history (order id, product id, product category, client id).
Should-haves
Nice-to-haves
Notably:
When there is enough data AI (machine learning) could auto adjust the weighing of the factors or the time a product is taken of the list after a purchase.
The impact brands can expect (and how to measure it)
Success is more than selling products, it’s about:
Boosting conversion and repeat purchases
Creating smarter cross-sell opportunities
Enhancing customer experience with personalised relevance
Increasing customer lifetime value
Projections based on typical patterns show the potential:
Abandoned search
If 7,500 people search for products each day and 7,130 leave without an order, recovering just 2% with €37 per order could mean ~€38,510 extra revenue/year.
Next Best Product
Implement personalised product content in 40 yearly e-mail campaigns, with extra order value of €3,700 per campaign, results in ~€148,000 extra revenue/year.
Implementation approach
Commerce and retail brands often take a phased approach to avoid “big bang” delays and start delivering value early:
1. Use-case development
Define Next Best Product segmentation/models as first priorities.
2. Discovery workshop
Map data sources, segment definitions, KPIs, and constraints.
3. Internal preparation
Translate requirements into data sources, dependencies, and “good enough” thresholds.
4. Design & plan
Agree on timelines and responsibilities; connect priority sources and unify identities; activate initial campaigns.
The takeaway
This type of CDP project shows that even scattered data can become a powerhouse for personalisation. By starting with a clear, high-impact use case and rolling it out in phases, retailers can build a foundation for both creative and commercial wins. With the right approach, next-product data doesn’t just inform decisions, it becomes one of the most valuable assets for growth.
To explore how you can benefit from a CDP, talk to a Deployteq specialist today or learn more.
