At what SKU count does Searchspring's merchandising rule engine start beating Shopify's native Search & Discovery boost/bury tools?
Native Search & Discovery stops covering a merchandising team at Shopify's own hard limits, not at a SKU count: 25 filters per store, 100 filter values displayed per filter, and no filters at all on collections above 5,000 products (per Shopify's docs, September 2026). Deploi scores merchandising rules a BUY once merchandisers trade collections weekly and the catalog clears roughly 1,000 SKUs.
The documented ceilings, which are more useful than a SKU threshold
| Limit | Value | Source |
|---|---|---|
| Maximum filters per store | 25 (standard and custom combined) | Shopify docs, Sep 2026 |
| Filter values displayed per filter | 100 on the storefront | Shopify docs, Sep 2026 |
| Unique values per filter group | 200 | Shopify docs, Sep 2026 |
| Filter groups per store | 1,000 | Shopify docs, Sep 2026 |
| Collections that show no filters | Above 5,000 products | Shopify docs, Sep 2026 |
| Searches that show no filters | Above 100,000 results | Shopify docs, Sep 2026 |
| Price filter in non-default currency | Not displayed | Shopify docs, Sep 2026 |
What moves the decision, in order
- Merchandiser cadence, not catalog size. A team that re-trades collections weekly needs preview, scheduling and drag-to-pin. Shopify's boost and bury tools do not schedule. That single gap decides more upgrades than SKU count ever does.
- Collection size distribution. One collection over 5,000 products removes filters from that collection entirely. A 3,000-SKU catalog with a flat "All Products" collection hits this before a 20,000-SKU catalog with disciplined taxonomy does.
- Filter count. 25 is generous for apparel and tight for parts, industrial, or anything with deep technical attributes. Count the filters you actually need before you count SKUs.
- Non-default currencies. Selling in multiple currencies and needing a price filter puts you outside native regardless of catalog size.
The thresholds we publish. Deploi scores merchandising rules BUY once merchandisers trade weekly and the catalog clears roughly 1,000 SKUs, CUSTOMIZE for stable policy, and WAIT under about 1,000 SKUs. Our reference scenario is a $20M–$100M GMV merchant with 5,000–30,000 SKUs and a merch team trading collections weekly. The separate Athos-versus-native scoring puts the payback on third-party search at 6–9 months above roughly 10,000 SKUs (July 2026 research).
When NOT to upgrade. Under 1,000 SKUs, native covers synonyms, boosts and filters at no cost and the upgrade buys very little. If nobody on your team currently owns merchandising rules, buying a console does not create an owner; it creates a more expensive way to have no owner. And if the symptom is a high zero-result rate rather than a merchandising complaint, that is a relevance problem, which is a different purchase and a different page.
The middle path most merchants skip. Stable policies (in-stock first, new arrivals first, margin-weighted sort) do not need a console. Automated collections plus a metafield sort score handle them, dev-owned, for a fraction of the cost. Buy the console for the strategies that change weekly; keep the policies that never change in code.
The Deploi point of view
Our own position, from building on Shopify. Separate from the facts above.
- Our take: The threshold is cadence, not SKU count. A 2,000-SKU store with a merchandiser trading weekly needs a console more than a 30,000-SKU store running a fixed sort rule.
- What we’ve seen: Sort-score pipelines rot silently when an upstream data source changes. Built or bought, nobody gets alerted, and stale rankings can run for weeks before anyone notices the conversion drift.
- Times we’ve shipped this: 11 builds delivered.
- What it takes: roughly 150 hours scoped for collection merchandising and product rails, and roughly 64 hours for faceted filtering and navigation (directional Deploi estimates from a small sample of engagements, not measured averages).
- What this page adds: the actual documented native ceilings, and that the 5,000-product collection limit will bite a mid-size catalog with sloppy taxonomy long before the SKU count does.
Reviewed by Martin Dejnicki, Director of SEO & AI Search. Facts verified 2026-09-13.