Personalized search shows each shopper her products first.
Keyword search treats every shopper the same, so "phone case" returns hundreds of generic results. Personalized search ranks the same results by what the store knows about her: her phone, her colors, her brands, her price range. The first screen then holds what she is likely to buy.
- Same query, a different order per shopper
- Remembers size, brand, price and style
- Feeds what she searched into every channel
A shopper who searches has told you what she wants. Plain keyword search ignores who she is and lists everything that matches the words. She scrolls, gets tired and leaves. Personalized site search keeps the matching but changes the order, using her past orders, her clicks in this session and what similar shoppers bought.
It works at three levels: the ranking of results, the filters and categories offered next to them, and the suggestions shown as she types. Each one shortens the path from query to basket.
The bigger gain is outside the search box. What a shopper searched for is the clearest intent signal a store has. When it reaches product pages, emails and ad audiences, every channel knows what she is looking for.
Beyond keyword matching
Two shoppers type "running shoes". One buys trail running gear every season. The other buys only one brand. Keyword search gives both the same list. Personalized search puts outdoor and performance shoes first for the first shopper, and her preferred brand first for the second.
It reads two kinds of signal. History says what she bought and which brands, sizes and price levels she returns to. The current session says what she is doing now. The ranking uses both in real time, so a shopper who switches from shoes to jackets mid-visit sees results that follow her.
It is the old shop assistant's skill, applied to every visitor. The owner of a good local shop remembered your size and the brands you liked and did not show you the rest. Search can do the same for every visitor, at once.
Filters that anticipate the shopper
Ranking is half the job. The filters and categories around the results matter as much, because they decide how quickly she narrows down.
Categories at the right depth
A shopper interested in audio sees "Wireless headphones", not "Electronics". The category she is offered matches what she has shown she wants.
Brand, style and price
Brands she buys or clicks come forward. Colors and styles she chooses come first. A shopper who buys premium sees premium first; one who chooses value sees value first.
Specifications remembered
Size, compatibility and technical choices carry over to her next search. The phone model she owns filters the phone cases without her selecting it again.
None of this asks the shopper for anything. The filters apply themselves, and she can still change them.
From the first keystroke
Search starts before she finishes typing. Suggestions use the partial query and what the store knows about her, so "run" can already show trail shoes in her size. Within the results, products that go with what she is looking at can appear next to it, which raises the basket without breaking relevance.
The worst search result is no result at all: a typo, a synonym the catalog does not use, a word in another language. A zero-results search is a shopper who said exactly what she wanted and got a blank page. Tolerance for typos, synonyms and a fallback to the closest match turn most of them back into results.
The ranking keeps learning. Each search and each click on a result feeds the next one, for her and for shoppers like her, so relevance improves without someone retuning rules by hand. The team still keeps control: merchandising rules can boost or bury products on top of the ranking.
Search as a signal for every channel
Many stores run search, email, SMS, ads and customer data in separate tools. The search tool learns what a shopper wants and keeps it to itself. The email tool never hears about it.
Connected, a search becomes a starting point. A shopper who searched and left can get an email or a message with the products she looked at. A high-intent search can shape the next product block she sees, and the ad audiences she joins, so ad spend goes to people who have shown real interest.
This needs one memory of each shopper, shared by search, pages, chat, messages and ads. Then the search box stops being an isolated feature and becomes the most direct way a store learns what each customer wants.
Search that knows the shopper, and tells the store
Search ranked per shopper, from the first keystroke
Results follow the live catalog, the sizes and categories she returns to, and what similar shoppers bought. Every search joins her profile, so pages, chat and the next message know it.
How the sales agent works →Help shoppers find the right product, fast
Shoppers search by age, size or occasion, not product names. The agent answers them in search, on every page and in chat.
Find the right product →Questions about personalized site search
How is personalized search different from keyword search?
Keyword search returns the same list to everyone who types the same words. Personalized search keeps the matching but orders the results by the shopper's size, brands, price range and what she is doing in this visit.
What does personalized search use to rank results?
Her past orders and the brands, sizes and price levels she returns to, her clicks in the current session, and what similar shoppers bought, applied in real time and on top of the live catalog.
Does it work for shoppers who are not logged in?
Yes. The session itself is a signal: what she searched and clicked in this visit shapes the next results. Once she is known, her history adds to it.
Can we still control which products come first?
Yes. Merchandising rules boost or bury products on top of the personalized ranking, so campaigns, stock and margins stay in the team's hands.
What happens to the search data after the visit?
With the agent, every search joins the shopper's profile. Product pages, chat, the next email and the ad audiences she joins all know what she looked for.
See which of last week's visitors you missed
One call, with your store and ad accounts open. Then 30 days free. A holdout group decides: if the agent doesn't add orders in 30 days, you don't pay.
See it on your store
You'll talk to one of us.