Faceted / Attribute-Based Search
Faceted search (also called attribute-based search) is a catalog navigation method that lets buyers filter results simultaneously across multiple independent attribute dimensions — such as voltage rating, material, thread size, or certification — so only the SKUs satisfying every selected criterion are returned. It depends entirely on structured, consistent attribute data stored at the field level; specifications buried in descriptions or PDFs are invisible to it.
B2B buyers specify — they don't browse
In consumer retail, a faceted sidebar is a convenience. A shopper with 500 sneakers can narrow by color and size. In B2B, the filter panel is often the entire purchase decision surface.
A buyer specifying a circuit breaker already knows they need a 20-amp, 2-pole, NEMA 1 panel-mount. They are not browsing. They are filtering for the existence of a product that matches a spec. If the facets can't return that result cleanly, the buyer either calls inside sales — an expensive outcome — or goes to a distributor whose search can.
The stakes multiply with catalog size. A mid-size electrical distributor may carry 300,000+ SKUs across dozens of product families, each with its own set of meaningful filter dimensions. No buyer pages through 300,000 results. Faceted search is how a catalog that size becomes usable at all. Without it, you are not really selling online; you are hosting a very large file cabinet.
How it works — and where the complexity lives
The mechanics are straightforward in principle. Every product has structured attributes stored as discrete key-value fields — not buried in a paragraph, but explicit: voltage: 120V, material: 316 stainless steel, IP rating: IP67, thread size: 3/4 in NPT. A search index ingests those fields and exposes each dimension as an independent filter axis.
When a buyer selects two filters simultaneously — "stainless steel" AND "IP67" — the engine applies AND logic across dimensions. Within a single dimension (material = stainless OR material = aluminum), it typically applies OR. At each step, the result count updates live so buyers can see which facets have hits and avoid chasing empty filters.
Category-level configuration matters as much as the index itself. A faceted search system does not surface the same filter dimensions for pipe fittings as for safety eyewear. Production-grade implementations bind specific attribute sets to specific categories, so buyers see only the dimensions relevant to what they are shopping. A "weight" facet on pipe fittings is noise. A "thread size" and "end connection type" facet is the whole purchase decision.
The complexity is not in the search logic. It is in the underlying data that the index consumes. If the attributes are wrong, inconsistent, or missing, the facets fail — and the search platform has no way to compensate.
Five failure modes that break faceted search
Most B2B catalog teams have experienced faceted search that technically works but practically doesn't. The root cause is almost always one of these five data problems.
1. Missing attribute values. If 40% of your fittings lack a material field, showing "Material" as a filter surfaces a broken experience. Buyers select it, find a partial result set, and distrust the catalog. A missing value is not a neutral gap — it is an active removal of that SKU from any buyer who filters on that dimension.
2. Value inconsistency across supplier sources. "SS", "Stainless", "Stainless Steel", and "316 SS" are four separate facet values unless the data was normalized. A buyer who selects "Stainless Steel" misses every SKU tagged "SS" even though they're physically the same material. In a catalog assembled from dozens of suppliers, this problem multiplies into hundreds of variants for the same underlying concept.
3. Units not canonicalized. "1000 mm", "1 m", and "39.37 in" should map to one value in a single unit system, not to three separate facet entries. Buyers who set a length filter in inches get back none of the metric-tagged SKUs. On large catalogs, unit chaos compounds into thousands of broken filter combinations.
4. Specs locked in PDFs and product images. If a wall thickness, pressure rating, or UL listing only appears in a downloadable spec sheet or inside a product photo, the search index cannot read it and the facet will not show it. Structured fields are the only thing a faceted search engine can filter on. Anything inside an image is invisible.
5. Wrong facets surfaced per category. Exposing facets based on field availability rather than buyer behavior means the filter panel looks complete but solves the wrong problem. The fix is not technical — it is understanding which attributes buyers in that category actually use to narrow a decision, then making sure those are the fields that are both complete and prominently faceted.
Attribute enrichment determines whether facets work
The connection between faceted search performance and attribute enrichment is direct: every dimension a buyer might filter on is a field that has to be present, consistent, and correct on every SKU in the relevant category.
This is where enrichment that only reformats the supplier's data falls short. If the supplier's data sheet lists voltage and amperage, that's all a reformatter can produce. But buyers of industrial controls also filter by UL listing, NEMA enclosure type, mounting configuration, and interrupting capacity — fields that may live in secondary documentation, standards databases, distributor data pools, or third-party certification records. A process that only reads the supplier's page leaves those fields blank, and blank fields mean the buyer's filter returns nothing.
Enrichment informed by buyer signals — specifically, the attributes buyers in a given category actually use as filter criteria — knows which fields to hunt for, not just which fields were easiest to find. The output is a set of attribute values that makes the facets work for the buyer who is actually filtering, not just the buyer who is casually browsing.
At Anglera, this is the practical meaning of buyer-signal enrichment: your PIM stores the data; the enrichment process figures out which attributes matter to the buyer filtering your catalog, then fills them in. Faceted search does not rescue a poorly attributed catalog. The data has to be right first.
Frequently asked questions
What is the difference between faceted search and keyword search?
Keyword search matches query text against product content — titles, descriptions, and indexed copy. Faceted search filters on discrete structured attribute values and is applied either after an initial keyword result set or as the primary navigation method. The two are complementary: keyword search gets a buyer into the right category, faceted filters narrow to the exact specification. Neither works well without the other in B2B.
How many facets should a category expose?
Research on e-commerce usability generally puts the optimal range at five to eight active filter dimensions per category. More than that creates cognitive overload and a panel buyers scroll past; fewer may not let buyers narrow to the spec they need. Priority should go to the dimensions buyers actually use to make a purchase decision in that category — not simply the attributes that happen to be most populated in the data.
Why do B2B sites struggle with faceted search more than B2C sites?
B2B catalogs carry orders of magnitude more SKUs — commonly 50,000 to over a million — spread across far more product categories, each with its own technical attribute set. The data comes from hundreds of suppliers with different naming conventions, inconsistent units, and varying levels of completeness. B2C sites tend to carry fewer, more homogeneous products from tighter supply chains with more standardized data. The B2B data quality problem is structural, not incidental.
Can search-platform configuration fix faceted search without improving the underlying product data?
Only partially. A search platform can apply synonym mapping at query time — treating 'SS' and 'Stainless Steel' as equivalent — but it cannot invent attribute values that were never stored. Missing values, incorrect values, and specs buried in PDFs or images can only be fixed in the underlying product data. Search configuration is tuning; attribute enrichment is the prerequisite.
What is a null facet, and why does it matter?
A null facet is a filter selection that returns zero results, or returns so few results the filter is functionally useless. It is almost always caused by missing attribute data — the field exists in the schema but is unpopulated on most SKUs. Buyers who select a filter and see zero results rarely conclude the products don't exist; they conclude the catalog is broken, and they leave. Null facets are one of the most direct ways poor attribute completeness shows up as lost revenue.