Braze Recommendations
Personalised next-best-content for every customer, in every message.
Braze AI
What it does
Braze Recommendations suggests the most relevant product, range or offer for each customer, learned from how they actually engage. Instead of one merchandised block for everyone, each message renders the items that person is most likely to want next. It turns a generic send into a personal one, and it improves as purchase and engagement history builds.
What feeds it
- ·Purchase history flowing in via Cloud Data Ingestion
- ·A product catalogue kept in sync with what is actually sellable
- ·Engagement events, so browsing and clicking shape the suggestions too
How to switch it on in Braze
- 1Sync the product catalogue into Braze and keep it fresh.
- 2Connect purchase events so Braze can link customers to what they buy.
- 3Configure a Recommendation in Braze, choosing the catalogue and the recommendation type.
- 4Render the recommendations via Liquid or Connected Content in email, push or in-app.
- 5Launch, and review which items are being recommended to whom.
How Fuse builds for and utilises this
- ·We build and sync the catalogue feed, so recommendations always draw from live, sellable stock.
- ·We pipe purchase events through Cloud Data Ingestion so the model learns from real transactions.
- ·We template the rendering in Liquid and Connected Content, so recommendations drop cleanly into any message.
- ·Where native recommendations run out, we extend them with warehouse-built affinity scores pushed back as attributes.
What to measure
A capability that fires on people who would have converted anyway proves nothing. Hold a slice out, measure only the gap above them, and let the recipe earn its place on evidence.
- ·Click-through and conversion against a non-personalised control block
- ·Revenue per recommended message, not just clicks on shiny items
- ·Holdout-proven lift on /prove, so personalisation earns its place on evidence
Pairs with
On its own it is useful. Alongside these models it compounds.