Evaluating your trial

Get your catalog in

Nothing else works until your items are indexed. Two paths:

• Shopify: done automatically once OAuth finishes. Watch the Items page. Most catalogs are ready within a few minutes; large ones (10k+) take longer. The status column shows which items are still pending.
• API: hit POST /v1/items in batches. The API reference has the request shape. Same Items page shows what's indexed.

Ingest is unmetered during the trial, just like in paid Starter. Upload the whole catalog without worrying about consuming anything; the only cap is the 2,000-item ceiling.

Spot-check search relevance

Open Search Lab and run the queries you actually care about. Not "shoes", but the specific ones you know your shoppers type and that your current search either gets right or gets wrong. The point is to compare against your existing baseline, whatever that is.

If a result looks off:

• Click the item to see what we indexed. Sometimes the catalog data we got is what's wrong, not the search.
• Try the selectivity dial right there on Search Lab. Lower values are looser (more marginal matches), higher are stricter. The default is 3.
• If the item itself is fine and tightening the dial doesn't help, that's useful signal. Send the query at [email protected] and we'll dig in.

Wire up real traffic

Search Lab is a useful spot-check, but the real evaluation happens when shoppers are running queries. Plug VeryQuery into the surface that matters:

• Shopify: drop the storefront search block into your theme. If you want to be careful, push it to a duplicate of your live theme and preview-test before publishing.
• API: swap one search surface (search page, autocomplete, similar-items rail) to call POST /v1/items/search. Pick whichever has the lowest blast radius.

Once real queries are flowing, the Search queries page becomes your main read. Volume builds, you can see what shoppers actually type, and you can spot the gap between what they search for and what your catalog returns.

Read the patterns

After a week of traffic, two surfaces start being useful:

• Intelligence map: a 2D layout of your queries. Clusters that don't have catalog coverage are the most actionable read: shoppers are asking, you don't have the goods, and now you know.
• Search queries list: sortable by recency, frequency, result count. Queries with zero or very few results are the other side of the same coin: shoppers asked, you might have the goods, but the catalog data wasn't shaped right for the match.

For most properties, this is where the value of the product shows up most clearly. If neither surface gives you anything you didn't already know, that's a real signal. Tell us; usually it means we should be doing something we aren't, and we'd rather hear it than guess.

Decide

Three questions that should have clear answers by now:

1. Is search returning what shoppers want? Search Lab on the queries that matter; the Search queries page on the queries that actually came in.
2. Are the gaps in the catalog visible enough to act on? Intelligence map clusters with low coverage, top zero-result queries, frequently-searched but rarely-clicked items.
3. Is the integration cost low enough that you'll keep using it? Either Shopify (dropping blocks) or API (one search endpoint swap).

If any answer is no, you can cancel in Billing with no charge. If all three are yes, do nothing. Your card is charged on day 15 and you keep going.