Connect what was sold
to what was promoted.
Retail knows what was sold. Marketing knows what was promoted. Random Truffle connects the two, by SKU, category, brand and customer.

Retail decisions happen at the level of the SKU and the category, while campaigns are planned by brand and season. The value is in seeing both together.
- SKU, category and brand
- Affinity and repeat purchase
- Margin, when it exists
- Sales related to campaigns
Questions you can ask
- Which campaigns are associated with sales for this brand?
- Which customers stopped buying this category?
- Which audience is most relevant for this SKU?
- Which creative performs best by product category?
Data that becomes useful
- POS and ecommerce transactions
- Catalog, SKUs and categories
- Customer and identity, where available
- Promotions, campaigns and assets
Store purchases only link to a customer when there is a loyalty or account identifier. The rest is analyzed in aggregate.
Which categories lost repeat buyers this month?
NexusThe decline is concentrated in two categories: running footwear and accessories. I can compare purchase frequency and recent promotions before treating either as a cause.
Output Table by category and a repeat-purchase curve by period
What the agents can do
01 / UnderstandNexusAnalyze sales and campaigns by category and SKU, and find where repeat buying fell.
02 / DefineAuraSelect the segment with affinity for the category and the criteria that justify it.
03 / CreatePrismAdapt pieces per SKU from the catalog and brand rules, for review.
04 / ActivateRadianPrepare distribution across paid and owned channels.
- 05 / LearnShared system
Compare sales, and margin where it exists, against the plan.
↺ Returns to Understand
How learning closes the loop
Sales associated with a campaign are not the same as incremental sales. When the question is incrementality, the plan includes a measurement design (a holdout or a comparable control) before launch.
What is measured- Sales by SKU and category
- Repeat purchase in the segment
- Margin, when available
- Creative response by category
Synthetic cohort. Descriptive, not causal.
Start with the category that changed.
Explore the evidence behind a drop in repeat buying.