Personalization tool
Constructor
Ecommerce search and recommendations platform that optimizes recs against a chosen business KPI using clickstream data and merchant rules.
Recommendations and search
E-commerce, Marketplace
The vendor’s claim
Drive upsells and increase RPV with hyper-personalized product recommendations, optimized toward the outcomes you track, such as RPV, conversion, average order value, or retention.Source
Where it’s strong
- Recs are tied to an explicit, selectable optimization target (RPV, conversion, average order value, or retention) rather than a generic relevance score
- Nine distinct recommendation logic sets, with merchandiser-controlled 'searchandizing rules' per pod, so a human can override the model per placement
- G2 reviewers score Product Recommendations at 9.8-10.0 and note support for detail-level personalization (e.g., tailoring by stated preferences like inseam or location)
Where it runs out
- G2 reviewers note recommendations were 'initially very poor' out of the box and needed added business rules to work as intended. Real setup overhead, not plug-and-play
- KPI-optimized recs need enough traffic and event volume to train against; low-traffic catalogs won't see the reinforcement-learning benefit quickly
- Uses reinforcement learning to continuously test and improve, which means the model is always shifting. A holdout has to be maintained deliberately or the baseline drifts
Best for
Commerce teams with real traffic and a merchandising function willing to actively tune rules pod-by-pod, not a set-and-forget team.
Questions before you buy more of it
- Which KPI is currently set as the optimization target, and who chose it?
- What manual rule work did we have to add post-launch to get recommendations to an acceptable baseline?
Connected applications
Intent-weighted search ranking
Results respect the query, then history. Not the other way around.
They searched for something specific. You showed bestsellers.
Inline job complement
One missing piece, named, in the cart.
Compare it in the hub
Put up to three tools side by side, or start from the customer situation you want to personalize.
.svg)
