Your first 30 days

Week 1 — Connect and let it watch#

The first job is getting your products and shopper behavior flowing into CartAmplify. Once you finish setup, the system is silently observing — building a picture of your catalog and how people interact with it.

What you do this week:

• Install the plugin for your platform (or hand the API guide to your developer).
• Confirm your catalog is syncing — open the dashboard’s Products page and check the count matches your store.
• Verify events are firing — open Data → Events and look for product views, cart adds, and searches in real time.
• Add any additional languages you sell in.

What CartAmplify is doing:

• Reading every product into its search index.
• Translating titles and descriptions into a form that handles typos, accents, and synonyms (this works on day one).
• Recording every event a shopper fires, building the dataset that personalization will train on later.

Week 2 — Watch search and browse get sharper#

By the end of week one you’ve probably accumulated several thousand events. Search and category pages are already using this data in lightweight ways: most-clicked products climb in popularity-based rankings, slow-moving items drop.

What you’ll notice:

• The “Popular” sort option on search and category pages reflects your real-world traffic from the last seven days.
• Out-of-stock products are filtered out by default — you can flip this per widget.
• Search results for ambiguous queries (like “shirt” or “bag”) start favoring the products your visitors actually click on.

What to check:

• Dashboard → AI Features → Search AI — Which queries get zero results? These are gaps in your catalog or naming. Fix them and you’ll convert traffic that’s currently bouncing.
• Dashboard → AI Features → Browse AI → Analytics — Which category pages have the highest bounce rate? Candidates for editorial reordering.

Week 3 — Personalization wakes up#

For each language you sell in, once you cross about 10,000 events (typically a few days to a couple of weeks of normal traffic), CartAmplify trains a personalization model just for that language. From that moment, search results and recommendations start reshaping for each individual shopper.

What changes:

• Two shoppers searching for the same query now see different orderings, based on what each has browsed before.
• The homepage “Recommended for you” widget switches from popularity-based to behavioral — surfacing products tied to each visitor’s history.
• Recommendation widgets on product pages start drawing from the personalization model instead of content similarity alone.

Week 4 — Review what you learned#

By now you have enough data to see real patterns. The fourth week is less about CartAmplify doing new things and more about you reading what it’s telling you about your store.

Things to look at:

• Revenue by surface — How much of your revenue came from search, browse, and each recommendation widget? Some surfaces will dominate; others might be in the wrong place on the page.
• Top zero-result queries — Add synonyms or product tags so these don’t bounce.
• Widget performance — Click-through rate per recommendation widget. Anything under 1% probably belongs somewhere else.
• Out-of-stock impact — If you sell things that go in and out of stock often, the dashboard shows how much that’s affecting recommendations.

After the first month#

Personalization keeps improving with every event. Models retrain automatically as more data accumulates — you don’t manage this, and there’s no “upgrade” button. The longer your store runs with CartAmplify connected, the better the results get.

The other thing that happens around month one: most stores realize they want to add more widgets. Maybe a “Recently viewed” strip on the cart page, or a personalized homepage shelf you didn’t enable at first. Adding a widget takes about two minutes in the dashboard.

Common worries this month#