As we continue improving product discovery by giving customers new tools and levers to pull, we’re very pleased to announce our new Product Launches functionality. This provides the ability to time the release of a product without worrying about waiting for a reindex to run, remembering to time an upload or configuring complicated rules right before the launch. All you need to do is set the date, and wait!

We know that there might be restrictions on when you’re able to show new products to your customers, whether they be for legal reasons, or because you want to make a big splash in the market with a timed launch. Now you can do this by including a release date and time on a product in your catalogue feed. Then when you run a search or browse query, that product will not be returned until the time is met.

For example, if you have an index that serves users in North America, and you want to time the launch of a product to midnight PDT on May 1st, simply set the following on the product in your catalogue feed:

As soon as the west coast reaches midnight, the product will be returned in search and browse queries.

In addition, we make it easy to check that the right date is set. This can be done by viewing the product in the Indexes tab of the Dashboard, or by overriding “now” in the API to see what the returned results will look like at some time in the future.

Continuing to ensure our services are resilient and easy to use is one of our commitments to our customers. 

Sometimes you just need to be able to quickly check if an API is responding, without implementing an expensive call. To that end, we have introduced a series of health check endpoints for our products. Now you can make a fast, lightweight call just to verify the status.

The following endpoints have been implemented:

• Search: https://ac.cnstrc.com/status/product/search
• Autocomplete: https://ac.cnstrc.com/status/product/autocomplete
• Browse: https://ac.cnstrc.com/status/product/browse
• Recommendations: https://ac.cnstrc.com/status/product/recommendations
• Collections: https://ac.cnstrc.com/status/product/collections
• Ingestion: https://ac.cnstrc.com/status/ingestion

If the API is up and running, you will get back an OK status message:

Further documentation on these endpoints can be found here:

• Search: https://docs.constructor.io/rest_api/search/queries#system-status
• Autocomplete: https://docs.constructor.io/rest_api/autocomplete_queries#system-status
• Browse: https://docs.constructor.io/rest_api/browse/queries#system-status
• Recommendations: https://docs.constructor.io/rest_api/recommendations#system-status
• Collections: https://docs.constructor.io/rest_api/collections/#system-status
• Ingestion: https://docs.constructor.io/rest_api/full_catalog#system-status

Merchandisers can now block products in search and browse results with a few simple steps in Constructor.io’s global searchandizing rule creation menu. 

Global blocklist allows for hiding products from results without needing to re-upload a full catalog. Use cases include quickly blocking products when there is a mistake in the catalog that leads to out-of-stock products returned in results or incorrect pricing. Additionally, products can now be globally blocklisted for a specific user segment, allowing for customizable results that meet targeted merchandising needs in business-to-business search. Finally, global blocklisting can allow adding products to the catalog in advance of launch and availability in results, a feature that will be improved upon in future releases.

Creating a global blocklist rule

Global blocklisting rules will remove the selected items from search or browse results based on the rule creation’s criteria. Global blocklist rules are created by inputting specific criteria, such as: selecting specific items to block, setting a start or stop time, selecting catalog attributes to block as a group, and/or setting the rule to apply to specific customer segments. 

Where to add the blocklist rule

Navigate to the searchandizing 'add rule menu' and click the 'add global' link at the bottom of the add new rule pop up. The option to create global blocklist rules is at the bottom of the global rule creation page, as seen above.

Types of Blocklist Rule Criteria

Specific items can be added individually to the global blocklist rule. Additionally, merchandisers can block entire groups of items by adding item groups based on catalog filters such as category, brand, or any other catalog attributes available. 

Specify blocklist rules to match time and segment needs

Once item criteria has been specified for a global blocklist rule, the rule can include specific start and stop time-boxing or be apply only to a subset of customer segments. For example, a rule could stay in effect until a designated product launch date whereupon the product is no longer blocked. Additionally, users have the option to set the rule for only a specific customer segment. Click ‘add’ and the new rule will apply across all search and browse results. 

As we continue to improve and enhance the tools available to make your users’ searches more effective, we have released the latest addition to Cognitive Embeddings Search. We now include an object in the search API response indicating how many of the products in the result set are delivered by Cognitive Embeddings Search (embeddings_match) and how many by our token-based KPI-optimizing search (token_match).

Read more about the response format here.

Now you’ll be able to further customize your experience for users. For example, you can add help text on your search page to let your users know that while they might not have received exact match results for their search, they are seeing results that are the next best thing. This lets you turn what could have been a lost customer into a purchase.

This capability is now enabled for all customers. If you are not using Cognitive Embeddings Search today and you’d like to learn more, please connect with your Customer Success Manager or contact us through support@constructor.io

Recently we released the ability for users to hide and order facets at a category level (you can read more about that here). 

This allowed users to set category contextual facets for an improved end user experience, but it was not always obvious which facets were slotted already in the global facet configuration - until today! 

We have now released the ability for users to be able to identify which facet groups and facet options have a global facet configuration applied to them, indicated with the global icon. 

 In the example above “Discount” was locally slotted in spot 1 for this category, which pushed down the existing globally slotted facet groups.

We recently released Cognitive Embeddings Search which dramatically decreases the amount of zero-result queries and the manual effort needed to add synonyms and redirects to optimize low, or no result queries.

We’ve seen amazing results from this new capability in testing across our customer base: increased purchase conversion, increased autosuggest product selects and decreased the rate of search reformulation. 

In an effort to make search results more clear and obvious we’re excited to announce our latest capability - Cognitive Embedding Search in Interact within the Dashboard. 

Users will now be able to see what products in the result set are coming from Cognitive Embeddings Search, which are indicated with a unique icon. You’ll also be able to toggle Cognitive Embeddings Search on and off to preview the effect of it at the query level within Interact in the Dashboard. 

This capability is now enabled for all customers that have Cognitive Embeddings Search enabled. If you are not using it today and you’d like to learn more please connect with your Customer Success Manager or through support@constructor.io

As part of our commitment to improve and optimize performance we're excited to announce that we’ve significantly improved performance of search for customers with the largest catalogs.

More specifically, the slowest 10% of queries (p90) saw a 10% improvement in response times, while the slowest 1% (p99) saw a 15% improvement in response times. 

For customers with catalogs of millions of items, the improvement was more than 50%! All of this means an overall improved experience for customers on your site, decreased time-to-checkout and higher conversion rates.

Changing search submits to search result views

Starting at the end of this week after business hours, we will communicate top search analytics in most cases in terms of result view events, rather than search submissions.

What are these events?

• Search submits are sent when users execute a search from the search bar (or select a search suggestion).
• Search result views are sent when users view a search page.

Why are we making the change?

We originally focused on search submits as they help to understand user search intent on a site or particular page (as opposed to inbound traffic). They can also be less prone to tracking errors due to their reliance on HTML form submissions.

Search result view events provide a picture of all traffic to the search page, whether it came from off-site, through the search bar, or links on-site. These events are what we use to analyze data science optimizations internally and align best to what customer merchant and product teams most commonly associate as a "search".

Are search submit events still displayed anywhere?

Yes! Search submissions are shown next to search result view events in all search interact pages. You can see these today for any interact page (accessible from the interact page here)

Search Click and Add-to-Cart Rates

We now communicate search add-to-cart rates (or ATCR) and clickthrough counts and rates over time in the search analytics overview page and search term detail pages.

Click and Add-to-Cart rates help understand how user behavior is changing over time at a global level, as well as for individual queries. One use case is to enable merchants to observe how their actions may be impacting user behavior on particular search terms.

To better serve our customers with millions of products in their catalog, we have increased the maximum number of results supported in search, browse and collections to 10,000 (from 999).

The previous limit of 999 was based on our research on user pagination depth (nearly zero conversions arise from pagination past the first few pages) and was designed to optimize for the fastest possible response times. By optimizing our core architecture we were able to increase the maximum results while preserving (and improving!) performance.

Users will now see ten times more results for a query without needing to filter down or sort the results. This encourages deeper discovery by paginating through a larger selection of products, facilitates longer infinite scrolls and can improve SEO with better crawlability of large category pages.

How we’re making Constructor Search better for your customers and for your bottom line.

As a Constructor Search customer, we know how much time you spend tinkering manually with zero-results queries, adding synonyms and redirects to capture users who might otherwise fall into a no results experience. To help this state of affairs, we have long leveraged our ML to generate synonyms before customers launch their search, automatically catch misspellings and more. To take these capabilities to the next level, we have built a more advanced zero-result algorithm -- Cognitive Embeddings Search -- into Constructor’s site search platform. 

Cognitive Embeddings Search transforms your product catalog into a “sea of stars,” where our machine learning and AI capabilities can return search results that are the closest neighbors to each query—virtually eliminating “no search results found” for good.

The best part? Cognitive Embeddings Search is included as part of the Constructor Search platform, and you can begin using it today by getting in contact with us through your Customer Success representative or through support@constructor.io!

Find out more about how this new addition to Constructor Search works here, and how it can improve revenue and your product discovery experience.