Listen to this blog as a podcast:

Expectations are rising, campaigns are getting more granular, and product catalogs keep growing. Merchandisers need good tools that automate the mundane, surface what’s meaningful, and accelerate iteration. That’s what we’re building: advancements that change the way you work, that empower you to make a bigger impact with less friction. Here’s a brief recap of three of the changes you’ll find available today.

1. Collections Enhancements

We’ve reimagined how Collections work to give merchandisers more control, flexibility, and speed. From AI-generated Collections, to better filters and group-based curation, these updates help teams create richer, more relevant shopping experiences without extra engineering overhead.

AI-generated Collections: Campaigns in seconds, not days

New landing pages and promotional hubs are tedious, sometimes even requiring days or weeks of manual product curation. AI-generated Collections cut that down to near-zero, using RAG to pull from your product database and generate automatic collections from prompts like “eco-friendly back-to-school essentials” or “summer festival gear under $50”. There’s no technical configuration, so it’s fast and intuitive for everyone.

Expanded Limits: Specific strategies all run in parallel

Until now, each index was limited to 25 collections — a hard ceiling that often forced teams to compromise or consolidate campaigns. We’ve removed that limit, so you can now create and manage hundreds of localized, brand-specific, or time-sensitive collections per index in parallel.

New Ingestion Workflow: Less manual work, more momentum

We’ve replaced the old push connector with a more powerful and streamlined API: pushTask. This new workflow allows merchandisers and developers to route data more efficiently and keep collections in sync as product attributes change. It’s designed to reduce reliance on engineering teams and keep your merchandising efforts moving speedily.

Flexible Conditions: Fine-tune your targeting

New filtering logic makes it easy to surface products by inventory level, price range, or specific text attributes. Collections now support advanced operators like “greater than”, “less than”, and “contains”, which let you control what shows up where with a lot more nuance. Dynamic, rules-based collections are now much easier to configure and maintain.

Merch by Group: Consistent curation everywhere

Group curation — previously only available for search and category pages — has now been extended to Collections! This means you can now apply the same curated rulesets across every milestone in the shopper’s journey, guaranteeing consistency in how products are ordered and boosted. Whether a shopper lands on a search result, category, or custom campaign page, the experience stays aligned with your strategy.

Recap: Why it matters

• Launch campaigns faster by automating some product curation and using more intuitive filtering logic for the manual choices.
• Remove the friction between ideation and execution to scale efficiently and support cross-team collaboration.
• Hyper-customize using previously-uningested data and the new wealth of available collection spots.
• Drive higher engagement and conversions through more personalized, relevant shopper experiences powered by AI.

2. Rules Analytics

Merchandising rules are very powerful, but without visibility, they can also create hidden complexity. That complexity tends to build up and make future additions a potential minefield. That’s why we introduced Rules Analytics: to help you see what’s working, clean up what’s not, and make easy data-driven decisions.

Know what’s working and what’s definitely not working

Rules Analytics gives you a clear view into how your merchandising rules are performing. You can see how often each rule is triggered, how many users it impacts, and what percentage of relevant queries it influences. This lets you figure out effectiveness at a glance and understand the actual impact of your rules without guesswork or trial-and-error debugging.

Declutter and optimize

Over time, rules tend to accumulate and weave themselves into a complicated mess. This is especially true during high-volume merchandising cycles like sales or holiday pushes. Now, Rules Analytics will surface rules that are out of date or only affecting an extremely small portion of your traffic and give you the choice to clean it all up with just a click. With your freshly-simplified configuration, you’ll be able to iterate faster on your next tasks.

Stay ahead of changes

Performance isn’t static. Rules that worked a month ago may no longer even make sense. With trend data over time, you can monitor which rules are driving results and which ones are fading into the background. This historical visibility helps you keep your merchandising strategy current and responsive to changes in shopper behavior.

Move fast with data-driven decisions

Trends move fast and new products need to be highlighted without warning. When you’re under pressure to make quick decisions, you need data you can trust and insights that inspire confidence. Real-time rule performance analytics let you align your rule strategy with what’s actually working, not just with what you hope will work.

Recap: Why it matters

• Focus your efforts where they make a real difference by keeping the rules that drive impact and tossing the outdated clutter.
• Stay aligned with automated relevance by using rules to complement, not conflict with, AI-driven ranking logic.
• Move faster with confidence using real-time performance data to inform merchandising changes.
• Support long-term scalability with a rules strategy that’s clean, measurable, and easy to evolve.

3. Improved A/B Testing Capabilities

We’ve made A/B testing a whole lot easier and a whole lot more powerful. With multi-variant experiments, built-in confidence intervals, and one-click rollout, you can test more ideas, trust the results, and move quickly when something wins.

Multi-Variant Experiments: Test more, wait less

Why test A vs. B when you can test A vs. B vs. C vs. D all at once? Multi-variant experiments let you compare multiple configurations in parallel, reducing the risk of seasonality or timing bias clouding your decisions. Whether you’re evaluating different ranking strategies (like DRR vs. MSR vs. hybrid) or testing boosts and filters, you can now explore more ideas in less time, with less noise.

Confidence Intervals: Make sense of the signal

What makes a bump in the conversion numbers statistically meaningful? With confidence intervals built right into the A/B testing UI, you can now see each result’s certainty without leaving the dashboard. No need to dig into raw data or do your own calculations — just clear, actionable insights to guide your next move.

One-Touch Variant Adoption: Go from winning to live

When a variant proves it’s better, you shouldn’t have to jump through hoops to roll it out. With one-touch variant adoption, you can apply the winning configuration instantly without the tedious legwork. Human error goes away since humans aren’t copying settings or reimplementing proven configuration. Now, as soon as it’s decided on by the stakeholders, you can roll out the changes immediately and start showing real results by the end of the board meeting.

Recap: Why it matters

• Multiple-variant A/B testing lets you build a culture of continuous optimization by making it easy for teams to experiment regularly, speed up learning, and cut down on the noise.
• Interpret results with clarity thanks to built-in confidence intervals that show statistical significance at a glance.
• One-touch rollout eliminates manual steps and empowers non-technical users to step into merchandising roles.

Wrapping up

These new features aren’t just incremental updates. These features give merchandisers the speed, visibility, and confidence to do more with less. Whether you’re spinning up new campaigns with Neural Collections, decluttering your rule set with performance analytics, or rolling out winning A/B test variants in seconds, the goal is the same: take the friction out of decision-making and let great ideas go live faster.

We’ll keep building the tools that help you adapt quickly, experiment often, and deliver better shopping experiences at scale. And we’re just getting started.