No results for something you stock
A plumber searches for 15mm compression elbow. Your marketplace has dozens from several suppliers. Search returns nothing, because suppliers listed them as 15mm comp elbow, elbow compression 15 mm and a manufacturer code. Another buyer types a part number with a dash where the supplier used a space. A third searches for a brand name that appears only in the product images.
Your search logs show thousands of searches like these every month. Each is a buyer who might have ordered.
Why generic search fails trade buyers
Most marketplace platforms ship with search designed for consumer shopping: match words in titles and descriptions, rank by popularity. Trade buyers search differently. They use part numbers, specifications, trade abbreviations and regional terms. They know exactly what they want and expect it at the top.
The catalogue makes it harder, because every supplier describes the same thing differently, and key attributes may be missing or buried in text.
| How trade buyers search | Why generic search misses |
|---|---|
| Part numbers with varied formatting | Dashes, spaces and prefixes differ |
| Trade abbreviations and slang | Not in product titles |
| Specifications, such as size and rating | Attributes not indexed or missing |
| Brand plus product type | Brand absent from supplier text |
| Misspellings on site | Typed on a phone, in a hurry |
What poor search costs
Search is how most B2B marketplace orders start. Zero-result searches and irrelevant results send buyers to a competitor or back to the phone. Suppliers whose products are poorly described get few sales and blame the marketplace. And your team cannot tell whether a category is weak because of demand or because of search.
Trade buyers are also less patient than consumers browsing for fun. They are often on site, on a phone, between jobs, and they will give search one or two tries. A buyer who fails twice rarely comes back to try a third time, and they tell colleagues the site does not have what they need.
Poor search also hides the value of the suppliers you worked hard to recruit. A specialist supplier with a deep range in one niche can go months without a sale simply because their products are described in terms buyers do not type.
The search improvements we build
- Search log analysis: we start with your search logs, grouping zero-result and no-click searches to find the biggest gaps, which usually fall into a few patterns.
- Part number handling: part numbers and manufacturer codes are indexed in normalised form, so any formatting of the code finds the product.
- Trade synonyms: a synonym list for your categories, built from search logs and trade terms, with AI suggesting candidates and your category managers approving them.
- Attribute search and filters: structured attributes are indexed, so searches like 15mm or a pressure rating match fields, not just text, and filters narrow results by the specs buyers care about.
- Ranking: results are ranked with signals that matter to trade buyers, such as exact part match, availability, past orders by similar buyers and delivery to their area, under rules you control.
- Ongoing tuning: a weekly report of new zero-result searches and falling click-through lets your team add synonyms and fix listings as the catalogue changes.
This can be built on your platform's search engine, or on a dedicated search service such as Elasticsearch, OpenSearch or Algolia, depending on what you run.
Search after the changes
Buyers typing part numbers and trade terms find what they want. Zero-result searches become a short weekly list with obvious fixes rather than an unknown. Category managers learn what buyers are looking for that you do not stock, which feeds supplier recruitment.
Is search losing you orders?
- Buyers report they cannot find products you know are listed.
- Part number searches fail on formatting differences.
- Nobody reviews zero-result searches.
- Filters are missing for the specifications buyers care about.
- Search ranks by popularity rather than by exact match.