Discovery and merchandising: what it is, and when it becomes the thing holding your store back
Picture a shopper's attention the way you would picture a coin she is deciding whether to spend. She lands on a category page holding roughly enough patience for two or three scrolls before she either narrows in on something or drifts to a competitor's tab, and what she sees in that window, which items are pinned at the top, which filters are offered, whether the search bar actually understands what she typed, decides whether that coin gets spent with you or somewhere else. A well-run version of that page reads her intent within a couple of clicks and rearranges itself to match, surfacing the in-stock size, the higher-margin alternative, the item three hundred other shoppers with her exact browsing pattern actually bought. None of that happens because your catalog is good. It happens because something between your search bar and your product grid is paying closer attention to her than she is paying to you, and most brands buy that something without ever being able to say in one sentence what it is or what it should be doing.
Some links here may earn us a small commission. It has no bearing on what we recommend, and we have turned down partnerships that would have.
A 2026 buyer’s guide to Nosto, Algolia and Boost Commerce. Pricing verified October 22, 2026.
We spend a fair amount of a typical onboarding call untangling this, usually right after a founder has already installed a search app, watched it do very little, and concluded that search apps do not work, when the truth is closer to having installed the wrong layer of a system that actually has three layers. Give us about eight minutes here and you will know what discovery and merchandising software actually governs, the point in a brand’s growth where it stops being optional, what Nosto, Algolia and Boost Commerce cost and what each one assumes about your catalog, and where this connects to the AI shopping assistant category covered in our companion guide, since the two are frequently confused and rarely the same purchase.
What this guide covers
What discovery and merchandising software actually does, and why it is different from both a search bar and a recommendation widget bolted onto a product page.
The point in a brand’s growth, measured in catalog size and traffic rather than order count alone, where this layer stops being a nice-to-have.
Nosto, Algolia and Boost Commerce, each assessed the same way: pricing, our read, the vendor in its own words where it has spoken to us, and who it suits.
A side-by-side table, then a recommendation sorted by catalog size and technical capacity rather than by budget alone.
The questions founders ask us most, including whether this replaces Shopify’s native search and what a merchandising rule actually costs you if it is set wrong.
The short version. A brand with a moderately sized catalog, a merchandiser rather than a developer running the storefront, and a single Shopify store should start with Boost Commerce, whose GMV-tiered pricing and pre-built merchandising rules get real lift without an engineer on staff, though Boost itself is candid that a very small catalog on modest traffic is usually well served by Shopify’s own search. A brand with real catalog complexity, a lot of attributes, a lot of near-duplicate products, or plans to run search across more than one storefront, should be looking at Algolia, whose usage-based pricing rewards a well-scoped implementation and punishes a sloppy one. A brand at real scale, with enough traffic and GMV that a one or two percent lift in conversion is worth a five or six-figure annual commitment, is the actual buyer for Nosto, which publishes no fixed pricing because none of its customers are buying a fixed amount of anything. Everything below explains how we tell those three apart before a vendor call ever happens.
The question that actually sorts this, before a demo gets booked
The variable that matters most is not your order volume, it is the shape of your catalog and the size of your engineering team, and the roughly one thousand orders a month marker used elsewhere in this series as a mid-market signal is genuinely secondary here, since a boutique brand with two hundred SKUs and a strong AOV can outgrow basic search well before it clears that order threshold, while a high-volume, low-complexity catalog can run on far less sophistication than its order count would suggest. Ask three questions instead. How many attributes does a shopper actually filter on: color and size are trivial, but a catalog with fit, material, use case and compatibility fields is a different problem entirely. Do you have anyone in-house who can read documentation and configure an index, because the tools that reward the most control also demand the most setup. And is the friction you are trying to fix a search problem, a shopper typing a query and getting nothing useful back, or a merchandising problem, a shopper browsing a category where the wrong items are getting the attention. Those are different failures with different fixes, and a vendor built for one will disappoint you if the real problem is the other.
There is a fourth question worth adding, and it came to us from one of the vendors in this guide rather than from our own framework, which is precisely why it belongs here: whether search is connected to what your customers say about the products it returns. A search log tells you what a shopper wants in her own words, and reviews are the proof that persuades her to buy, so a discovery layer that can read ratings, review counts and review attributes, and rank or filter on them, closes a loop that a search layer working on product data alone leaves open. It is not a reason to choose one vendor over another on its own, but it is a reason to ask every vendor on the shortlist how its ranking treats review signal, and to listen carefully to the answer.
What independent research says about attention on a category page
The most directly relevant independent evidence here comes from the Baymard Institute, a UX research organization whose work is used by more than 29,000 practitioners and, by its own account, by seventy-one percent of Fortune 500 ecommerce companies. Its on-site search research is built on usability testing across dozens of live sites, in which more than seven hundred distinct search-specific usability issues surfaced and were distilled into thirty-one guidelines covering the search field, autocomplete behavior and result presentation, and the accompanying benchmark scores three hundred and forty-four leading US and European ecommerce sites against those guidelines. The finding that matters most for this article is a blunt one: across that benchmark, on-site search performance remains poor industry-wide, meaning most brands are leaving a measurable amount of found revenue on the table not because their catalogs are weak but because the layer between the search box and the results is under-built, which is precisely the gap this category of software exists to close.
How each vendor is metered says almost as much as what it does. Boost Commerce prices against your store’s average monthly online GMV, so the invoice rises automatically as your revenue does, with feature access bundled into three named plans and a price ladder inside each one rather than features sold one at a time. Algolia prices against usage, specifically search requests and records indexed per month, which rewards a lean, well-tuned implementation and can quietly balloon for a catalog that is indexed inefficiently or queried far more than it needs to be. Nosto prices neither by GMV tier nor by simple usage, but through a custom quote built from your GMV, traffic, catalog and the specific modules you select, which is standard for enterprise personalization but means there is no number to compare until you are already in a sales conversation. None of the three is a drop-in replacement for the others, and the honest comparison is not which is better but which matches the catalog complexity and technical capacity you actually have today.
Boost Commerce: merchandising rules and AI search a small team can actually run
Metered by average monthly online GMV. Shopify and Shopify Plus.
Boost is the packaged option in this guide, a single Shopify platform that covers AI search, filters, merchandising, recommendations, predictive bundles, pre-order and back-in-stock, built on the assumption that the person running discovery day to day is a merchandiser or ecommerce lead working in an admin rather than a developer working in a ticket queue.
Pricing
Plan | Monthly | Monthly GMV included | Above the cap | What unlocks |
|---|---|---|---|---|
Launch | $29, $99 or $159 | Up to $5K, $15K or $50K | Moves to Convert | Advanced filters with 5 custom filter trees, smart and AI-powered search, AI recommendations, predictive bundles, metafields, pre-order, search merchandising with 10 active campaigns, 90-day analytics, Shopify Markets features |
Convert | $299, $499 or $599 | Up to $100K, $250K or $400K | Moves to Accelerate | In-collection search, collection merchandising, promotional banners, B2B wholesale features, embedded bundles, 20 custom filter trees, 40 active campaigns, 180-day analytics |
Accelerate | $699, $799 or $1,499 | Up to $650K, $1M or $2.5M | Custom pricing above $2.5M | AI-personalized search, recommendations and bundles, variant display, unlimited custom filter trees, 100 active campaigns, 365-day analytics, customer success manager, premium support, customization, unlimited back-in-stock notifications |
Verified October 22, 2026 against Boost’s own pricing page, where a GMV slider sets the price within each plan; every plan is month-to-month with a 21-day free trial and no credit card required. Boost’s Shopify App Store listing displayed as not currently available on the day we checked, so the figures above rest on the vendor’s own page alone.
Our read
Boost’s real advantage is that its merchandising campaigns and filter trees are built to be configured by a merchandiser rather than a developer, which matters enormously for a brand that wants pinning, boosting, demoting and scheduled campaigns without writing a query or touching an API, and the GMV-based scaling means the tool grows with revenue rather than requiring a renegotiation at every growth stage. Its integrations do a lot of quiet work as well: deep connections with Klaviyo and with REVIEWS.io mean the merchandising layer can react to email behavior and to review signal without custom development, so a rule can push products rated four stars and above toward the front of a collection and demote those rated two stars and below, with review count as a condition so a single glowing review does not outrank a product with dozens of strong ones. Boost reports that Smith Optics, after moving search and reviews onto Boost and REVIEWS.io together, saw its best month produce $2M in conversion value across 400,000 search interactions with a 10.5 percent product click-through rate, and saved $80,000 on its bill in the process; those are the vendor’s published case-study figures rather than independently audited ones, but they describe the consolidation argument well.
Where it runs short is genuine catalog complexity: a brand with dozens of filterable attributes and a lot of near-duplicate products will eventually find the pre-built rule structure limiting compared to what a fully custom search index can do, and its personalization, which sits only on Accelerate, is lighter than what Nosto or a well-implemented Algolia deployment can produce. The other thing to read carefully is the ladder itself. The entry price of $29 applies only up to $5,000 in monthly GMV, collection merchandising and in-collection search arrive on Convert at $299, and the personalized layer most brands picture when they read the marketing begins at $699, so the honest budget for the product most brands actually want is the Convert or Accelerate band rather than the headline figure.
In their words
By Boost’s own account it is built for Shopify and Shopify Plus brands with roughly $1M to $100M in annual GMV and anywhere from 500 to several hundred thousand products, the band where native search and merchandising run out of depth and enterprise tools bring a heavier total cost and longer implementation. It points to teams where a merchandiser runs discovery day to day, to brands treating conversion as an ongoing practice who want every search, filter click and recommendation tap as behavior data to test against, to brands already on REVIEWS.io, and to brands consolidating search, merchandising, recommendations and reviews into fewer contracts. It is candid about who is better served elsewhere: brands under about $1M in annual GMV with fewer than 500 products and low traffic, where native Shopify search usually covers what they need, and brands not on Shopify at all, since Boost is built Shopify-first.
“Your site search log is the one place shoppers tell you exactly what they want, in their own words. The right discovery platform turns that signal into what they see next across search, recommendations and merchandising, and lets your team act on it the same day. If search is the only place that signal gets used, most of its value is left on the table.”
Luke Creger, Partnerships Manager, Clearer.io
Luke Creger
Partnerships Manager, Clearer.io
On how it views our work, Clearer.io wrote: “Atelier cuts tools before adding them and treats conversion as ongoing work. That combination lowers what clients spend on their stack while raising what it earns, and it’s exactly how Boost and REVIEWS.io are built to work together.” Clearer.io
Who it suits
A Shopify brand with somewhere between a few hundred and several hundred thousand products, a merchandiser or ecommerce lead rather than a developer running the storefront, and enough revenue to sit comfortably in the Convert band or above, particularly one already running REVIEWS.io or Klaviyo and willing to grow through Boost’s own plans rather than migrate the moment revenue crosses a threshold. It is not the answer for a very small catalog on modest traffic, where Shopify’s native search will do, or for a brand that is not on Shopify.
Algolia: search infrastructure priced by how much you actually use it
Metered by search requests and records indexed. Shopify, headless and any custom stack. 3.7 stars from 45 reviews on the Shopify App Store.
Algolia is infrastructure rather than a packaged app, an API-first search and discovery engine that a technical team configures directly, which is both the reason it can match almost any catalog and the reason it asks more of the people running it than anything else in this guide.
Pricing
Plan | Monthly | Usage included | Above the allowance | What unlocks |
|---|---|---|---|---|
Free | $0 | 10K search requests, 50K records, 5K recommendation requests, 5K crawls | Upgrade | Search and recommendations with a limited feature set, self-service support only |
Grow | From $0, pay as you go | 10K search requests, 100K records | $0.50 per 1K search requests, $0.40 per 1K records | Keyword search and browse, query suggestions, 10 rules per index, manual synonyms, data transformation, 30-day analytics, third-party connectors, US, UK and EU hosting |
Grow Plus | From $0, pay as you go | 10K search requests, 100K records | $1.75 per 1K search requests, $0.40 per 1K records | AI synonyms, AI ranking, advanced personalization, query categorization, collections, 10,000 rules per index, 90-day analytics |
Elevate | Quoted, volume-based | Custom | Custom | NeuralSearch, AI Collections, Smart Groups, real-time personalization, 99.99% availability, SSO and enhanced SLA, enterprise support plans, professional services |
Verified October 22, 2026 against Algolia’s own pricing page and its Shopify App Store listing, which shows the app as free to install with the Free plan allowances above and states that usage charges are billed by Algolia separately from the Shopify invoice. Recommendation requests beyond the included allowance are charged at $0.60 per 1,000, and annual contracts on Elevate carry discounts that are not published.
Our read
Algolia’s usage-based structure is the most transparent pricing model of the three, since every dollar above the included allowance is tied to a number you can actually monitor, search requests and indexed records, rather than to a subjective feature tier, and that transparency is exactly why a well-scoped implementation costs so much less here than an unscoped one. Its strength is genuine flexibility: because it is infrastructure rather than a packaged app, a technical team can build search, browse and recommendations exactly to the catalog’s real complexity, including multiple indexes for multiple storefronts or markets, which neither of the other two vendors handles as natively. The gap between Grow and Grow Plus is worth pausing on, because the AI ranking and synonyms most brands assume they are buying live on the plan whose search overage is three and a half times higher, and a brand modeling cost on Grow rates will underestimate the real invoice.
That same flexibility is the risk, because a catalog indexed inefficiently, with duplicate records, unnecessary fields, or records refreshed more often than needed, turns the usage-based pricing against you, and a brand without someone who understands indexing will likely overpay relative to what a packaged tool like Boost would have charged for similar results. The Shopify listing’s middling rating is consistent with what we see in practice: merchants who install it expecting a merchandiser-friendly app tend to be disappointed, and teams who treat it as the developer platform it is tend not to be.
Who it suits
A brand with real catalog complexity or a multi-storefront, multi-market footprint that has, or is willing to hire, the technical capacity to configure and monitor an index properly, including headless builds where a packaged Shopify app is not an option at all. It is a poor fit for a brand that wants merchandising rules a marketer can set without engineering help.
Nosto: personalization at a scale where the pricing is the conversation
Metered by GMV, traffic and modules selected. Shopify, Shopify Plus, BigCommerce, Salesforce Commerce Cloud, Magento and more. 4.7 stars from 61 reviews on the Shopify App Store.
Nosto sells personalization as a set of modules rather than a ladder of tiers, letting a brand assemble search, category merchandising, recommendations, bundles, content personalization and testing on one customer data layer, with the implementation handled largely by Nosto’s own team.
Pricing
Plan | Monthly | What sets the price | Above the cap | What unlocks |
|---|---|---|---|---|
Product Experience Cloud | Quoted after a demo | GMV, traffic, catalog size and modules selected | Not applicable, priced to scale | Personalized search, AI search, category merchandising, product recommendations, post-purchase upsell, dynamic bundles, personalized emails |
Content Experience Cloud | Quoted after a demo | GMV, traffic and modules selected | Not applicable, priced to scale | A/B testing and optimization, content personalization, pop-ups |
Verified October 22, 2026 against Nosto’s own pricing page, which states that there are no rigid tiers, that its Experience.AI layer is included in every module at no extra cost, and that pricing follows a demo, and against its Shopify App Store listing, which shows the app as free to install with additional charges that may apply. There is no self-serve trial.
Our read
Nosto’s pitch is that personalization should be modular rather than tiered, so a brand pays for the pieces that solve its actual problem and adds the rest later on the same underlying data, and its own marketing cites a typical ten to fifteen percent conversion lift and fifteen to twenty-five percent AOV uplift, figures that should be read as a vendor’s own claim rather than an independently audited number until your own results confirm them. It genuinely suits enterprise-scale personalization, since the platform is built for brands whose traffic peaks would strain a smaller tool, and the absence of a self-serve trial reflects an implementation that is genuinely more involved than installing an app. We have run it ourselves at that scale, and the difference is less any single feature than the way search, merchandising and recommendations learn from one another once they share a data layer.
The honest limitation is that none of this is affordable or necessary for a smaller catalog, and a brand evaluating Nosto without meaningful GMV and traffic already in place is very likely shopping above its actual needs, which is worth saying plainly given how often the name comes up in conversations with founders who are nowhere near the scale it is built for. The other cost is comparability: with no published number, the only way to weigh Nosto against the other two is to sit through the demo, which is a reasonable price for a brand at scale and an unreasonable one for a brand that is not.
Who it suits
A brand with substantial existing traffic and GMV where a single percentage point of conversion lift is worth a serious annual commitment, often running across more than one platform or market, and prepared to run a proof of concept and a sales process rather than a self-serve trial before signing anything.
Side by side
Boost Commerce | Algolia | Nosto | |
|---|---|---|---|
What it actually is | Packaged Shopify search, filters and merchandising rules a marketer can run | Usage-priced search infrastructure a technical team configures directly | Enterprise personalization across search, merchandising and content |
Entry price | $29/mo up to $5K monthly GMV | $0 free tier, then usage-based | Not published, quoted after demo |
Realistic working band | $299 to $1,499/mo (Convert or Accelerate) | Depends on requests and records; Elevate quoted | Quoted, typically an annual commitment |
Metered by | Monthly GMV | Search requests and records | GMV, traffic and modules selected |
Above the cap | Next price step or plan; custom above $2.5M monthly GMV | $0.50 (Grow) or $1.75 (Grow Plus) per 1K searches, $0.40 per 1K records | Priced into the quote |
Personalization | Accelerate only | Advanced on Grow Plus, real-time on Elevate | Included in every module |
Review signal in ranking | Native REVIEWS.io integration | Built by your team | Integrations with review platforms |
Free trial | 21 days, no card | Free plan | No, proof of concept for qualified merchants |
Needs a developer? | Rarely | Usually | Handled largely by Nosto’s team |
Shopify App Store rating | Listing unavailable on October 22, 2026 | 3.7 (45 reviews) | 4.7 (61 reviews) |
Not built for | Very small catalogs on low traffic, and stores not on Shopify | A marketer who wants to self-serve without engineering help | A catalog without the traffic or GMV to justify the spend |
Pricing verified October 22, 2026 from each vendor’s own pricing page and, where available, its Shopify App Store listing. We refresh these figures quarterly.
The verdict, by the catalog and team you actually have
For the brand running a single Shopify storefront with a catalog a merchandiser can reason about, Boost Commerce delivers real lift without adding headcount, and its GMV-scaled plans mean the bill grows in step with the revenue it helps produce, provided you budget for the Convert band where collection merchandising lives rather than for the entry price. For the brand with a genuinely complex catalog, multiple attributes, near-duplicate products, or plans to run search across more than one storefront or brand, and with the technical capacity to configure it properly, Algolia’s usage-based pricing rewards the investment in doing it right. For the brand already running at real scale, where traffic and GMV are large enough that a one-point conversion lift funds the platform several times over, Nosto is worth the sales conversation even without a published number, because its modular personalization is built for exactly that scale and nothing smaller.
As with the AI shopping assistant category, order volume alone is a poor guide here, and a smaller, high-AOV catalog can justify Algolia’s flexibility, or eventually Nosto’s scale, well before a pure order count would suggest it. If you already run a reviews platform, ask each vendor how review data reaches its ranking before you ask anything else, because a discovery layer that knows which products customers actually love is working with better information than one that only knows which products exist.
Questions we get asked
What exactly does discovery and merchandising software control on my site?
It governs what a shopper sees on your search results and category pages, including ranking, filtering, and which items get algorithmic or rule-based priority, which is a different job from a recommendation widget that only appears on a single product or cart page.
Is this the same thing as an AI shopping assistant?
No, and the two are frequently confused: an AI shopping assistant talks to a shopper directly, while a discovery and merchandising layer works silently behind the search bar and category grid, and our companion guide to AI shopping assistants covers that distinction in full.
Does Shopify’s native search already do this?
Shopify’s built-in search handles basic keyword matching but lacks the merchandising rules, attribute-level filtering and personalization these three vendors specialize in, which is why even Shopify-native brands install one of them once their catalog or traffic outgrows basic keyword search. For a small catalog on modest traffic, native search is often enough, and at least one vendor in this guide says so itself.
How do I know if my catalog is too complex for a packaged tool like Boost?
A useful test is counting the attributes a shopper would realistically filter on at once: two or three simple ones, like size and color, usually work fine in a packaged tool, while five or more interacting attributes typically need the flexibility of an infrastructure-level product like Algolia.
What happens if a merchandising rule is set incorrectly?
A poorly set rule can bury your best-converting products behind ones that are merely new or in stock, which is why every vendor in this category recommends reviewing merchandising rules against actual conversion data on a monthly cadence rather than setting them once and leaving them alone.
Do I need a developer to run any of these?
Boost Commerce is built for a marketer to configure directly, Algolia generally requires a developer for initial setup and ongoing index maintenance, and Nosto’s implementation is handled substantially by the vendor’s own team as part of the engagement.
Can I run one of these alongside an AI shopping assistant like the ones in the companion guide?
Yes, and in practice most mature stacks run both, since the discovery and merchandising layer governs the search and category experience while a support or clienteling assistant handles the direct conversation, and the two rarely conflict because they operate on different pages.
When does this stop being optional?
Once your catalog has enough attributes or your traffic is large enough that a shopper’s first search regularly fails to surface the right product, the case for this layer moves from nice-to-have to a measurable drag on revenue, regardless of what your monthly order count happens to say.
Where to go next
If the real friction in your funnel is a conversation rather than a category page, a shopper asking a question rather than failing a search, read our companion guide to AI shopping assistants, which sorts that category the same way this one sorts discovery and merchandising. Our strategic merchandising practice covers exactly this kind of platform selection and rule design, our SEO practice covers how a stronger on-site search layer feeds organic discovery, and our work with PXG on Nosto shows what enterprise-scale personalization looks like once it is actually running. Get in touch and we will map your catalog and traffic against this framework with you.
