No black box. You choose the signal, QSortby scores your catalog in real time and keeps in-stock first — or personalizes per shopper — then proves the lift with a built-in A/B test and revenue attribution. Here’s the inside.
Set the rule per collection, over a window you pick — a rolling last N hours or days, or a fixed range: last week, a set month, any two dates.
A ranking only matters once it reaches the storefront — and that is where most merchandising apps publish on their own schedule, in their own markup. QSortby does neither.
Preview, manual order, manual membership and product exclusions are on every plan, including Free.
You don’t have to trust the sort. QSortby runs it as an experiment and measures the money it actually drove.
Illustrative. QSortby splits visitors between your current order and the new sort, then reports the real funnel each drove.
Attribution is influenced / last-touch — it shows which sort a converting shopper engaged with, an honest directional read of lift, not a lab-grade incrementality study.
Beyond one order for everyone, QSortby can render a per-visitor feed. Visitors are classified into types you define, and each type gets a ranking that blends weighted signals — a live, personal front row.
Recommendation engines rank on purchase history, so a first-time visitor gets nothing. QSortby reads this visit instead — pace, dwell, revisits, hesitation — from the first pause, before anything is bought.
best-reviewed and free-returns first · no urgency, no upsell
premium end first · bundles and gift-ready items allowed
Illustration of the same four products reordered by what each session showed — arrows are each product's move against the default order named above, not a result claim.
Tags describe products, not people — so you can open any product and check the tag yourself. Rename them to your own words at any time.
The layer reads and reports; your storefront stays exactly as it is. You see the mix of sessions your store actually gets — the read without the risk.
The read reorders that shopper's feed from their first pause on — behind the same A/B test and revenue attribution, so you see what it earned, not what it promised.
Personalization is never used to manufacture pressure — no false urgency, no pressure dressed up as help. The point is relevance, not leverage.
Once a shopper has chosen, the question turns from "what should they find" into "what else belongs in this order" — answered from the same live signals, so nothing here is a static list you maintain by hand.
Sold-out and already-in-cart products are filtered out of every surface, and a failed request hides the section rather than blocking the cart. On Growth, the session read chooses between the candidates in the cart, popup, checkout and thank-you offers — same cart, different suggestions for a shopper who is comparing carefully and one who is buying a gift in a hurry.
Install free and run it on one collection — or book a 30-minute demo and we’ll screen-share the experiment + attribution on your real collections.