Hybrid search: matching on meaning, not just words

Product search has always had the same blind spot. A visitor types summer garden party outfit and gets nothing useful, because no product in the catalogue is called that. The words don’t match, so the results don’t come.

Hybrid search fixes that. Alongside classic text matching, we now compare the meaning of a query against your products, using both their text and their images. A product photo carries information the title and description often leave out: the cut, the colour, how formal something looks. So summer garden party outfit surfaces the floral midi dresses, the linen separates, the strappy sandals, the things someone would actually wear to a garden party in summer, even though not one of them mentions gardens or parties.

What this changes in practice:

  • Long-tail and descriptive queries return relevant products instead of empty or padded result sets
  • Pages built around intent, occasion or persona now have inventory behind them
  • Fewer thin pages, because fewer queries come back empty
  • Hybrid search is rolling out over the next few months.
Hybrid Search for finding product matching

Manage users & permissions over all your websites

Users and permissions are now managed in one place across all your instances.

If your team runs several markets on Verbolia, you no longer manage access instance by instance. Add someone once, set their role, and control what they can see and change across the estate, with a single view of who has access to what.

Read the docs: Managing users and permissions

Log in in Verbolia

Account security is changing

We’re raising the baseline on account security. Passwords will need to be at least 14 characters, and multi-factor authentication will become mandatory for all users.

Users will receive a notification in a few weeks to update their login at sign-in. Later this year that update becomes mandatory. We’ll confirm the exact date and send reminders before it does.

Japanese and Korean support

Vpage now supports Japanese and Korean end-to-end. Both languages break the assumptions most search stacks are built on. Japanese has no spaces between words and mixes three scripts, so the same product can be written three different ways and still be the same thing. Korean compounds words in ways that naive tokenisation splits apart. We’ve handled tokenisation, script normalisation and matching for both, so keywords, page generation and product matching behave the way they do in the languages you already run.

If you’re planning a launch in either market, talk to your customer success manager, we’re ready to set things up for you!

Coming soon: the Vpage API

Today, Vpages can only be managed from the Verbolia application. The Vpage API changes that: alongside the app, your own systems can create, update, delete and monitor Vpages, so page management fits into your automated workflows.

We’re opening a pilot and looking for a small number of customers to build with us.

The API lets you:

  • Create and remove Vpages, with or without their content
  • Create, update and delete on-page content: H1, alias, cold content, meta title, meta description, custom data and FAQs
  • Track performance across your whole instance: visits, AI referrals, revenue, Search Console impressions and clicks, and crawler activity from Google, OpenAI, Perplexity, Anthropic and Microsoft. Enough to rebuild your Vpage dashboard in your own BI tool.
  • Read each Vpage’s content, settings and performance data
  • Pull product data and product-level performance for every product matched to a Vpage

Places are limited. Sign up for the Vpage API pilot and we’ll get back to you with the details.

Verbolia API

Master the next generation of search.

Join 100+ e-commerce brands scaling visibility with Verbolia.

Request a demo