In productionLoyalty programme marketing

A Data-Driven Product Selector for Email Campaigns

Choosing which products to feature in each email campaign was based on gut feel. I built a selector that ranks products on performance, margin, freshness and stock, and gives the team a shortlist where every pick can be explained, with a person approving the final choice.

The problem

Product picks for campaigns depended on whoever was planning that week. There was no consistent way to balance best-sellers, margin, new products and stock, and no record of why a product was chosen.

What I built

A scoring model ranks every product on past campaign performance, margin, how recently it was featured, seasonal signals and stock risk. The weights are adjustable by the team.

The model itself is plain arithmetic rather than AI, so every score can be traced. AI only writes a one-sentence explanation of each pick, and a person approves or swaps products before the email is built.

The value for the business

Consistent choices

Every campaign uses the same transparent criteria.

Explainable picks

Each recommendation shows exactly why it scored highly.

Better balance

Built-in rules keep a healthy mix of product types and price points.

Human final say

The team approves or swaps every pick.

Where it stands

In production on the client's product and campaign data, giving the team a ranked, explained shortlist for every email campaign.

Frequently asked questions

Why not let AI choose the products?

Because the team needs to know why a product was picked. A transparent score can be checked and tuned; AI is only used to write the explanation.

Want something like this for your company?

Let's look at where your team loses time today and what AI could take off its plate. Fifteen minutes, no obligation.

Schedule a 15-minute call

More case studies

Jazz while you're here?