Feature and Integration

Frequently asked questions

Product Feed FAQ's

Q.1. What is a product feed? Why do we need it?

Answer

Answer:

We need your e-commerce product catalogue data for us to power the natural language search on your e-commerce store. The product feed can exist in JSON, XML or CSV.

Q.2. Is there a specific recommended format for providing the feed?

Answer

Answer:

{ "product": { "url": { "url": "product url", "searchable": "no" }, "pid": { "pid": "product id", "searchable": "no" }, "base": { "title": "title", "brand": "product brand", "model": "item model", "category": "product category", "category path": "maincat>subcat>subsubcat...", "currency": "INR", "popularity": "value between 0 to 1" }, "<'field name'>": { "type": "string/number/double", "value": "valueOfField", "searchable": "no" (optional) } } }

Q.3. How frequently do we need to update the feed and what is the mechanism of sharing the feed?

Answer

Answer:

The full feed needs to be provided the first time post which there needs to be a daily update frequency to update newly added products/remove stale products etc., Also, a higher frequency update can be done generally for updating price/stock depending on how frequently it is updated in the store. The mechanism of sharing the feed is via a standard REST API.

Q.4. What does the field “searchable = yes” mean in the feed?

Answer

Answer:

“searchable = yes” need to be applied for the attributes in the feed which needs to be allowed to be part of your Natural language search query. For eg, for a category mobile, internal memory or RAM fields need to be marked searchable=yes to allow users to search for mobiles using memory and ram.

Q.5. What does popularity in the feed mean?

Answer

Answer:

Popularity in the feed for each product refers to the rank of the product within a given sub- category. It could be based on number of page views etc. a product gets. It is often a value between 0 to 1.

Q.6. If we don’t have Popularity, what should we do?

Answer

Answer:

If you don't maintain a product popularity or rank, you could give us certain specific signals we ask for, for each product and we will calculate the same.

Trial FAQ's

Q.7. What is a trial and how long is a free trial available?

Answer

Answer:

A trial is a mechanism through which you can share us a sample product feed and explore our service for 2 weeks for free before you can start using our service for your search capabilities.

Q.8. Do we have a paid trial?

Answer

Answer:

If you want to explore our product trial beyond 2 weeks, we allow a paid trial which is lesser than the actual service cost for a few more weeks.

Q.9. What is needed for a trial?

Answer

Answer:

For a product trial, you can give us a sample feed in the prescribed format (preferably) or any format over a REST API (web URL).

Search Integration FAQ's

Q.10. How long does the integration take?

Answer

Answer:

The integration for trial or actual service can take anywhere between 1-3 weeks depending on the size of the product feed.

Q.11. How do we integrate SeeknShop.IO’s Natural Language Search on to our e-commerce website or app?

Answer

Answer:

We offer our product capabilities through a SaaS model via REST APIs and give you a result set in a Json format for each query which can be consumed to mobile app or website or any interface.

Q.12. Do we have Auto Suggest and how do we integrate?

Answer

Answer:

We have auto-suggest templates you can choose from and can be integrated along with the search service after you provide the product feed. Normally auto-suggest is not available during free trials.

Q.13. If we have different feeds for different markets for the same seller, how do we integrate?

Answer

Answer:

If you have different feeds for different markets, you can provide it as different indexes for us in the product feed and we generally treat is as a different store and different index as the query set and metrics etc are all different for different markets. Please talk to our support during integration.

Q.14. If we have different interfaces/channels for our Omni-channel retail, how do we integrate?

Answer

Answer:

For any channel, it is a single simple Rest based search API which can be consumed and integrated to give a seamless experience across channels.

Natural Language Search FAQ's

Q.15. What are the various Search capabilities offered via SeekNShop.IO Search Platform?

Answer

Answer:

    • More than 98% Search Accuracy
      Personalized Search
      Fully automated self learning search for adjusting result ranking and synonyms/noise identification
      Super Fast Search with less than 50ms Response Time
      Rich variety of Auto Suggest Models
      Detailed Inventory Insights and Analytics
      More than 20% increase in conversion and AOV for specific queries.
      Multi-Store Availability
      Multi-Tenancy
      World-wide data center support
      Full Natural Language Search
      Pluggable into Chat (Chatbot APIs)
      Voice Capabilities
      Pluggable into Websites, Apps, Kiosks
      Ability to Customize Ranking and Precision of results
      SEO capability for long tail queries
      Create powerful chatbots on Facebook/Skype etc.
      Search History per user
      Merchandising capabilities
      Scales more than millions of SKUs or documents
      Non-Inventory Searches
  • Q.16. What does Natural language Search mean?

    Answer

    Answer:

    The capability to search digitally the same way you ask a human or explain your product discovery requirements to a human.

    Q.17. What are the advantages of a Natural Language Search?

    Answer

    Answer:

    It is the natural way end-users are used to explaining their buying requirements in the real world and makes it easy and fast for users to discover whatever they want and not be limited and curtailed by Keyword matching Search technologies

    Q.18. Do you have Machine Learning Search to learn from User patterns and feed?

    Answer

    Answer:

    We have filed 6 patents for various machine learning and natural language understanding algorithms within our Search. DeepSeek learns from the seller/dealer inventory and from user search patterns to give the most intelligent machine learning based search pluggable into your website/app.

    Q.19. Do you have Mobile optimized Search?

    Answer

    Answer:

    Our Search is mobile website, mobile app, desktop web and kiosk optimized. The interface being plugged in doesnt matter given our API blends in naturally irrespective of the interface being plugged into or the domain being applied to.

    Search Merchandizing FAQ's

    Q.20. What are the various Search Merchandizing Configurations available?

    Answer

    Answer:

    With SeekNShop.IO Search Configuration Panel, you get various Merchandizing capabilities listed as below:

    • Add or Remove Synonyms
      Add or Remove Noise Words or Stop Words
      Customize the Precision and Recall Accuracy of your Search Results
      Intelligent dashboard which gives you weekly insights into Data
      Detailed Reports which gives you insights into your user's search data for any custom amount of time
      Spell Check Corrections (Normally taken care of by our Spell Check algorithm)
      Query to URL Redirect - For a given search query term, configure the URL it needs to be redirected to (Merchandizing pages)
      Rank By Category - Used for boosting rank/promotion of specific products in a given category. Search for the product by Product Id to set the rank
      Rank by Keyword - Used for boosting rank/promotion of specific products for a given search keyword. Search for the keyword or query to set the rank for ProductId
      Facet Configuration - Configure the facets to be shown per category
      Sorting Configuration - Configure the order of sort options for search results
      Customize Auto Suggest Options and Choose among Intelligent Auto Suggest Templates
      Top Queries - Enable to display top performing search queries
      Keyword Suggestions - Enable to display intelligent keyword suggestion
      Search Scope -Enable to display scope suggestions within the auto suggest widget
  • Q.21. What are the Reporting and Dashboard capabilities available? Or What are the various user metrics captured?

    Answer

    Answer:

    All the reporting and Dashboards are customizable and our standard analytics include the following

    • Query Volume per week/month/quarter
      Query Distribution per category
      Query Distribution per Brands per category
      Price Distribution of User Searches
      Maximum Searched Categories
      Most Searched Attributes per week
      Most Searched Colours/Sizes/Configurations per week
      Top AutoSuggest Selections per week