Search and Product Discovery (Gartner)

Gartner published the Magic Quadrant for Search and Product Discovery on 22 June 2026, and the top of the chart changed hands. Constructor placed furthest on Completeness of Vision and highest on Ability to Execute among every vendor evaluated, one year after reaching the Leader rung for the first time. That is the first time a vendor other than the established names has taken both axes at once.

The report is credited to Mike Lowndes, Sandy Shen, Aditya Vasudevan and Noam Dorros. Gartner's abstract for the 2026 edition names eleven vendors in the field: Algolia, Algonomy, Athos Commerce, Bloomreach, Constructor, Coveo, HawkSearch, Lucidworks, Netcore Unbxd, Rezolve Ai and Zoovu. Constructor states that the report covers twelve. Four vendors have announced Leader placements, each in its own release: Constructor for the second consecutive year, Algolia and Coveo for the third, and Netcore Unbxd.

The 2026 Magic Quadrant for Search and Product Discovery, and the vendor that took the top of it

Constructor's own account of the placement is specific: furthest on Completeness of Vision and highest on Ability to Execute, which is the pair of coordinates a vendor can only claim when it tops both axes. Its customer list in the announcement names Sephora, Under Armour, The Very Group and Target Australia, all large retailers with catalogues measured in the hundreds of thousands of items, which is the shape of buyer the top of this chart tends to reward.

Algolia's third consecutive Leader placement came with a list of what it has shipped, and the list is a fair summary of where this market moved: an agent studio, a server implementing the Model Context Protocol, a feature it calls AI Collections, and integrations with Microsoft Copilot, Salesforce Agentforce and Adobe Experience Manager. Each of those is a way of being reachable from somewhere other than a search box on a product listing page.

The field itself turned over. The 2025 edition, published on 24 June 2025, named vendors including FactFinder, Nosto and Yext that do not appear in the 2026 abstract. Some of those absences will be exits rather than failures to qualify, and a first glance at a quadrant rarely distinguishes the two. What the turnover shows is that Gartner is still deciding who this market belongs to, three editions in.

Forrester and Gartner agree about Constructor and part ways on the rest

The cross-analyst counterpart is The Forrester Wave: Commerce Search and Product Discovery Solutions, Q3 2025, which scored nine vendors against 31 criteria and sorted them into Leaders, Strong Performers and Contenders. Constructor is a confirmed Leader there, and the scoring detail Forrester published has it at 5 out of 5 in vision, innovation, roadmap, adoption, and pricing flexibility and transparency, plus a run of current-offering criteria covering usability, search configuration, recommendations, type-ahead and dynamic categorization. Forrester described it as a best fit for firms that want a vendor with a singular focus on this market and that prioritize speed of innovation.

So both firms put Constructor at the top of the same market within nine months of each other, using different methodologies and different criteria counts. Agreement at the top of a market is the least interesting thing two analysts can do, because the vendor leading on both scores is usually the one with the clearest story to tell.

The disagreements below the top are where the two documents earn their keep, and there are two of them.

Coveo is a Strong Performer at one firm and a three-year Leader at the other

Forrester placed Coveo as a Strong Performer in the Q3 2025 Wave, while rating it superior on optimization to business metrics and on product badging and highlighting its content search capability. Gartner has placed Coveo on the Leader rung in both the 2025 and 2026 editions, and Coveo announced the third consecutive year itself.

Both placements can be correct, and the reason is in the market names. Forrester's evaluation is scoped to commerce, which is why the word leads its title. Coveo's center of gravity is search across a much wider corpus of enterprise content, and a commerce-scoped scorecard asks fewer questions about that. Gartner's market definition reaches past the storefront into natural language processing, intent detection, algorithmic merchandising and agentic integrations, which is closer to where Coveo's product actually sits.

For a buyer, that is a concrete warning about shortlist construction. A shortlist drawn from Forrester's Wave alone drops a vendor Gartner has placed on the Leader rung three years running, and a shortlist drawn from Gartner's quadrant alone spends evaluation time on vendors whose strength is in a corpus the buyer may not have.

Bloomreach was a Leader at both firms, and its own material still points at 2025

Bloomreach is a confirmed Leader in Forrester's Q3 2025 Wave, and its own Gartner page describes a Leader placement two years running, which by the date of that page means the 2024 and 2025 editions. Gartner's 2026 abstract lists Bloomreach in the field. No 2026 Leader claim from the company could be found.

That is a statement about the public record, not about the placement. A vendor that reaches the rung and does not announce it looks identical, from the outside, to a vendor that did not reach it, and this page will not guess between them. What can be said is that Bloomreach's most recent published claim about its own position is a year old.

There is a third data point, and it is on this site already. Bloomreach is a confirmed Niche Player in Gartner's Magic Quadrant for Personalization Engines, published on 3 February 2026, four months before the search quadrant. Two Gartner documents place the same vendor in different parts of their charts in adjacent markets four months apart, and neither document contradicts the other. A buyer who reads only the search quadrant would not know the personalization quadrant exists, which is the practical argument for reading both.

One word separates the two market names, and it is the whole difference

Gartner's category is Search and Product Discovery. Forrester's is Commerce Search and Product Discovery Solutions. One word separates them, and it is doing more work than it looks like.

Forrester's scoping decision makes its scorecard sharper: every criterion can assume a catalogue, a shopper and a revenue event, which is why its Leaders are vendors whose products were built for retail. Gartner's wider scope means its criteria have to work for buyers whose search problem is not a storefront at all, and it pulls in vendors whose discovery surfaces sit inside other applications.

That difference is the reason both documents are worth reading rather than one. A retailer evaluating discovery for a product catalogue is Forrester's buyer, and a company trying to make a decade of documents and tickets findable is closer to Gartner's. They will shortlist from different halves of the same four names.

Both scorecards moved toward the same surface at the same time

The clearest agreement between the two firms is not a vendor placement. It is that both rewrote their criteria toward conversational and agentic discovery inside the same window, which is unusual for two evaluations that disagree about the market's boundaries.

Forrester's Q3 2025 scorecard includes agentic and conversational discovery as a current-offering criterion, and Constructor took 5 out of 5 on it. Gartner's abstract for the 2026 quadrant names natural language processing, intent detection, algorithmic merchandising and agentic integrations as the innovations it covers. Algolia's Leader announcement lists an agent studio, a server implementing the Model Context Protocol, a feature called AI Collections, and integrations with Microsoft Copilot, Salesforce Agentforce and Adobe Experience Manager. Read together, those four lists describe a market where the search box is being replaced by a caller, and where the caller is increasingly software.

That changes what a platform has to expose, and the change is not cosmetic. A relevance model built for typed queries has to produce structured product data instead of a ranked list. The platform needs an interface an agent can call without a human in the loop. And the permission model has to survive a caller that is not a person, which is a different problem from one that hides results from a shopper.

The limit is worth stating plainly. Both firms are now scoring a capability whose buyers are still writing their requirements, and a criterion can only measure what the market can currently articulate. A buyer drafting a specification for agentic discovery this year is drafting it for a surface that did not exist in the last procurement cycle, and the vendor that leads the criterion is not necessarily the vendor that leads the deployment.

A seventeen billion dollar figure that measures the wrong thing

Gartner values digital commerce search at approximately 17.41 billion dollars in 2025. That number is the size of the spend, not the size of the opportunity, and it bundles licence fees, implementation, in-house engineering and the search functionality that arrives inside a commerce platform whether or not anyone bought it separately. No buyer can read their own budget off it.

The more useful half of the report is the Critical Capabilities companion, which scores the same vendors against nine capabilities so that the ranking depends on the use case rather than on the aggregate. Netcore Unbxd, a Leader in the quadrant, scored 4.05 out of 5 in the Multisite and Globalization use case, which is a different claim from being a Leader and a more specific one. A retailer operating in a dozen countries and languages should be reading that score, not the quadrant position.

The general point holds across every analyst report on this site. A placement answers the question the analyst asked, and buyers arrive with their own question.

What to ask before you shortlist a discovery platform

Ask which of the two documents your problem belongs to. If the corpus is a product catalogue and the metric is revenue per session, Forrester's commerce-scoped Wave is the closer instrument. If the corpus is wider than a catalogue, Gartner's definition will fit it better, and the Leader rung will look different.

Ask what happens to relevance when the query stops being a query. Conversational and agentic discovery is where both firms now spend their criteria, and it changes what a platform has to expose: structured product data, an interface an agent can call, and a permission model that holds when the caller is not a person.

Ask how much of the platform you are buying and how much you are assembling. Several of the vendors here are strong on one layer and expect the commerce platform, the CDP or the merchandising tool to supply the rest, which is a different total cost from a vendor that ships the whole path.

Ask for the use-case score rather than the quadrant position. Nine capabilities scored five ways produce a ranking that moves with the use case, and the vendor that leads the aggregate is not always the vendor that leads yours.

Ask what the vendor's own most recent claim says. Some publish a placement every year, some publish one when it improves, and the gap between the two is information a buyer can read without accusing anyone of anything.

Gartner does not endorse any vendor, product or service depicted in its research publications, and does not advise technology users to select only those vendors with the highest ratings or other designation. Gartner research publications consist of the opinions of Gartner's research organization and should not be construed as statements of fact. Gartner disclaims all warranties, expressed or implied, with respect to this research, including any warranties of merchantability or fitness for a particular purpose.

The enterprise-search counterpart on this site is a different market with a similar name. Cognitive Search Platforms covers the Forrester Wave of that name, scored across fourteen providers and twenty-one criteria, and its buyer is an employee looking for an answer rather than a shopper looking for a product. The retrieval technology underneath the two is close to identical, and the scorecards are not, which is why the same vendor can be a Leader in one document and unplaced in the other.

Discovery sits in front of the commerce platform rather than inside it, and the layer below is covered separately: Commerce Solutions covers The Forrester Wave: Commerce Solutions, Q3 2026, the first edition in which Forrester combined its B2B and B2C evaluations, and its own source list already names this market as an adjacent evaluation. Read together, the two answer different halves of one question, since a discovery vendor is usually chosen to sit on top of a platform the buyer has no intention of replacing.