Many online shoppers don’t browse retail sites using navigation, their product discovery process starts with site-search. In fact, an avg. of 30% of consumers use site-search (EConsultancy) and those who buy are 91% more likely to use site-search than those merely browsing (Findwise). And those percentages increase dramatically on mobile - where limited screen real estate puts product search in the limelight.
One of our clients found that their mobile visitors were 30% more likely to use site-search than desktop visitors. And according to one US Study, 4 out of 5 smartphone users use retailer apps, with 47% of them using the app for product search.
Searchandising, a term that blends the concepts of On-site Search and Product Merchandising, hasn't been around long and isn't necessarily a mainstream idea among retail professionals. Fundamentally, searchandising is about augmenting established search techniques - faceted search and navigation, autocomplete, recommended products, recent searches, related queries, etc - with behavioural data and automation in order to create a seamless, personalised, unique, and highly profitable product search experiences.
At a basic level, it involves manually set boost-and-bury rules based on margin, arrival data, stock availability, or some other attribute. Sound familiar? This is where the vast majority of retailers live.
At a more advanced level, it incorporates behaviour data and uses sophisticated machine learning algorithms to automatically sort products based on relevancy, popularity and personal taste - in addition to those boost-and-bury rules.
Automated Searchanding = Self-Optimised Search Merchandising
Automated Searchanding is about the move from manual, labour-intense on-site search merchandising toward a completely self-optimised solution. Search results (and category listings) that rank and re-rank themselves in real-time to reflect new information, such as:
- The relationship and similarities between products: leading to more relevant results and automatically generated Related Results
- Business rules and logic such as new items, campaigns, stockouts, etc.
- CRM data (past purchases, user profile info, etc.)
- Behaviour data - clicks, add-to-cart and purchases
- New words and synonyms
How Loop54 does Searchandising
Behaviour data is collected and processed in real-time and used to improve base algorithms (option to add CRM data too)
Search results (and, optionally, category listings) are sorted to reflect popularity, personal taste, campaigns, seasonality, new and changing inventory.
Related search results are generated in-real time and displayed alongside direct search results.
Search terms that do not exist in the catalogue metadata are learnt and assigned as synonyms. This includes misspelled words that lead to conversions.
All previously manual merchandising tasks like sorting, ranking and tagging are eliminated
In using sophisticated and automated searchandising techniques, retailers are able to better anticipate shopper intent and minimise operational expenditures. Doing has helped our customers increase online sales by +12%, increase AOV by up to 75% and eliminate hundreds of hours per year maintaining and optimising search systems.