Build vs. Buy>Search & Discovery>Algolia vs. Elasticsearch

Algolia vs. Elasticsearch for Shopify: Which Search Lane Wins?

Written by Deploi EditorialReviewed by Martin Dejnicki, Director of SEO & AI SearchUpdated September 2026Pricing verified September 2026 (quote-based tiers excluded)

Algolia versus Elasticsearch is a WAIT for Shopify stores under 10,000 SKUs on a Liquid theme: Shopify Search & Discovery is free and ships semantic search, 1,000 synonyms and product boosts (verified Sep 2026). Past that line, Algolia wins the head-to-head as the app path, free to install with usage-based charges that scale with volume (verified Sep 2026). Elasticsearch wins for headless storefronts and custom ranking, consumption-priced or free to self-host (verified Sep 2026).

Your profile — see how the verdict shifts

VerdictWAIT (Search & Discovery, free) under ~10,000 SKUs on a Liquid theme · BUY Algolia as the hosted app path above it · BUILD on Elasticsearch for headless storefronts, custom ranking or a self-hosted engine
Buy score
5.6
Build score
8.0
Confidence
HighThe native feature set and its documented limits are published and free (verified Sep 2026); Algolia's listing is free to install with usage-based charges and Elasticsearch publishes only consumption rates (verified Sep 2026), so the head-to-head is decided by storefront architecture and dev bench, not by a price table; third-party search pays back in 6–9 months only above roughly 10,000 SKUs (July 2026 research)
Reference scenario
$20M–$100M GMV · 2,000–8,000 SKUs · Liquid theme storefront · one main locale
As of
September 2026

Decision at a Glance

Your profileVerdictWhy
Under 10,000 SKUs, Liquid theme, one main locale (reference scenario)WAITShopify Search & Discovery. Free, with semantic search, up to 1,000 store-wide synonyms (20 per group) and boosts of up to 10 search terms per product (verified Sep 2026). Tune it against your top 100 queries before paying anyone; third-party search pays back in 6–9 months only above roughly 10,000 SKUs (July 2026 research).
A documented native gap bites: semantic matching in predictive search, Japanese-locale semantic search, or filters the native app can't modelDEPENDSLeave native, then choose by architecture. Semantic search doesn't apply to predictive search and isn't supported for the Japanese locale, and synonyms skip SKU and barcode fields (verified Sep 2026). On a Liquid theme with no search engineers, Algolia's hosted app is the faster fix; with a dev bench or a headless roadmap, Elasticsearch is the durable one.
10,000+ SKUs or attribute-heavy filtering on a Liquid theme, no dev bench for a search serviceBUYAlgolia. An App Store install, free to start with usage-based charges that scale with volume (verified Sep 2026), gives hosted faceting and relevance without an index pipeline to run. Budget the usage line as a growth cost and diary a re-price when query volume doubles.
Headless storefront, custom ranking signals, or multi-source search with an in-house or agency dev benchBUILDElasticsearch, as a custom search service. Consumption pricing runs as low as $0.09 per VCU per hour for search and $0.14 for ingest on Elastic Cloud, or free to self-host under AGPLv3 (verified Sep 2026); the $40,000–$120,000 build (Deploi estimate, illustrative) buys ranking logic and query data you own outright.

What Algolia vs. Elasticsearch Actually Drives

OutcomeImpactHow it works
Revenue — directHighShoppers who search convert at a multiple of browsers, so relevance, typo tolerance and useful filters turn queries into orders; every lane here changes that rate, which is the number to measure.
Customer experienceHighZero-result pages and irrelevant top hits are the fastest way to lose a shopper who knew what they wanted; synonyms, boosts and semantic matching exist to remove them.
Data & insightMediumQuery logs reveal demand you don't stock and language you don't use; an owned index keeps that dataset first-party, a hosted one keeps it on the vendor's export terms.
Operational efficiencyMediumMerchandisers who can tune relevance from a dashboard instead of a ticket ship promotions faster; a custom engine gives them that only if you build the dashboard.
Revenue — indirectLowSearch infrastructure has little effect on acquisition; it converts the traffic you already earned.

Spend ceiling: Search spend should scale with catalog complexity and architecture, not with ambition: under 10,000 SKUs the right number is close to zero, past it a usage line sized to real queries, and a $40,000–$120,000 custom service (Deploi estimate, illustrative) only for headless or ranking-driven stores.

What buying enables (top apps)

  • + Algolia: hosted relevance and faceting live in days on a Liquid theme, free to install with usage-based charges (verified Sep 2026)
  • + Algolia: a strong headless story through its APIs, without running an index pipeline yourself
  • + Vendor-maintained compatibility with Shopify API version changes and theme updates
  • + Merchandising rules and analytics dashboards your team can operate without engineers

What building additionally unlocks

  • + Native: semantic search, 1,000 synonyms and product boosts at zero cost (verified Sep 2026)
  • + Elasticsearch: ranking logic and custom signals beyond any hosted rule editor, on an engine that is free to self-host under AGPLv3 (verified Sep 2026)
  • + Elasticsearch: multi-source search across products, content and support in one index
  • + Query and click data as a first-party dataset with no export ceiling

Find Your Verdict in 3 Questions

  1. Is your catalog under roughly 10,000 SKUs on a Liquid theme, with no documented native gap (predictive-search semantics, Japanese locale, custom filters) costing you sales?

    Yes: Your verdict: WAIT — Shopify Search & Discovery is free with semantic search, 1,000 synonyms and product boosts (verified Sep 2026); tune it against your top 100 queries.

    No: Go to question 2.

  2. Are you headless, or do you need custom ranking signals or search across sources beyond the product catalog?

    Yes: Go to question 3.

    No: Your verdict: BUY (Algolia) — a hosted App Store install, free to start with usage-based charges that scale with volume (verified Sep 2026); diary a re-price when queries double.

  3. Do you have an in-house or agency dev bench that can own an index pipeline and a query service?

    Yes: Your verdict: BUILD (Elasticsearch) — consumption-priced on Elastic Cloud or free to self-host under AGPLv3 (verified Sep 2026); a $40,000–$120,000 service (Deploi estimate, illustrative) you own outright.

    No: Your verdict: BUY (Algolia) — its headless story is strong and the hosted platform carries the engine; revisit Elasticsearch when a search engineer joins.

The TCC Scorecard — 12 Dimensions

TCC — Total Cost of Capability: what it actually costs to have this capability over three years, whichever way you get it. Each dimension is scored 0–5 for both paths. How we score →

DimensionBuyBuildWhy
Cost
Acquisition & implementationAlgolia installs from the App Store free to start (verified Sep 2026) with index and UI configuration on top; native Search & Discovery is free and already installed; an Elasticsearch service is an estimated $40,000–$120,000 build (Deploi estimate, illustrative).
Recurring feesAlgolia bills usage-based charges that scale with volume (verified Sep 2026), so query growth is a cost line; native is free; Elasticsearch bills consumption, search as low as $0.09 per VCU per hour (verified Sep 2026), or nothing when self-hosted.
Maintenance & upgradesA vendor-maintained app absorbs API version bumps; native has no upkeep line; a custom Elasticsearch service carries ~15–20% of build cost per year in upkeep (Deploi estimate).
Switching & exitLeaving Algolia means rebuilding index rules and the search UI, though the catalog itself stays in Shopify; native tuning exports nothing but costs nothing; an owned Elasticsearch index moves with you.
Risk
Vendor riskAlgolia is an established hosted platform with a live listing; the native app is Shopify's own; Elasticsearch is open source under AGPLv3 with no vendor to lose when self-hosted (verified Sep 2026).
Security & compliance surfaceA hosted search index holds catalog and query data outside Shopify; native keeps it inside; a self-run Elasticsearch cluster is a surface your ops team must secure.
Platform-deprecation exposureSearch apps ride Shopify's API version cycle through their vendor; the native app moves with the platform; a custom service is your own code against the Storefront and Admin APIs.
Value
Fit to requirementAlgolia ships hosted faceting and relevance tooling for Liquid and headless alike; native fits catalogs under roughly 10,000 SKUs with documented gaps in predictive and Japanese-locale semantic search (verified Sep 2026); Elasticsearch fits anything you're willing to build.
Time to marketAlgolia is live in days on a Liquid theme; native tuning takes an afternoon; a custom Elasticsearch service takes 8–16 weeks (Deploi estimate, illustrative).
Performance & scaleBoth engines scale past any Shopify catalog; native is fine to roughly 1,000 SKUs without tuning and workable to about 10,000 with synonyms and boosts (July 2026 research).
Data ownership & AI-readinessQuery logs and ranking signals live in Algolia's platform on its export terms; native exposes little query data; an owned Elasticsearch index makes every query and click a first-party dataset.
Focus & opportunity costSearch is core to conversion but rarely a differentiator below 10,000 SKUs; a custom engine is a standing engineering commitment that only headless or ranking-driven stores should sign up for.

The App Landscape

AppStatusPricingBest for
AlgoliaLiveAPI-first hosted search and faceting with a strong headless storyFree; usage-based charges at entry volumes; scales with volume (verified Sep 2026)Liquid or headless storefronts past roughly 10,000 SKUs that want hosted relevance and faceting without running an index pipeline
ElasticsearchLive — flaggedNo Shopify App Store listing; a search engine rather than an app, so it needs an index pipeline, query service and storefront UI built around it. Pricing is consumption-based with no flat monthly plan, or free to self-host under AGPLv3; 14-day free trial on Elastic CloudIngest as low as $0.14 and search as low as $0.09 per VCU per hour; storage and retention as low as $0.047 per GB per month; egress as low as $0.05 per GB transferred per month; support tiers add 5%, 10% or 15% on consumption; self-managed deployments carry no published dollar figure (verified Sep 2026)Headless storefronts, custom ranking signals, multi-source search and teams that want to own the index outright
Shopify Search & DiscoveryNativeFirst-party, free. Shopify's free first-party app; renders metafield-based storefront filtersFree (included)Catalogs under roughly 10,000 SKUs on a Liquid theme, tuned with synonyms and boosts

The Build Path

  • Tune Search & Discovery against real queries: Pull your top 100 search terms and zero-result queries, then map synonyms (up to 1,000 store-wide, 20 per group) and boosts (up to 10 terms per product) to them (verified Sep 2026). Semantic search covers related words and concepts on its own. An afternoon's work, repeated quarterly.
  • Metafield-driven filters without a platform swap: Attribute-heavy catalogs usually need better filters more than a new engine. Metafield-source search filters have been delivered on Shopify without swapping the search platform, a Deploi pattern (July 2026 research).
  • Custom search service on Elasticsearch when the triggers hit: For headless, 10,000+ SKUs with complex attributes, or custom ranking, build the index pipeline, query service and UI on Elastic Cloud or a self-hosted cluster. Deploi's Eluma runs this lane with real-time sync, typo tolerance and faceted filters; budget $40,000–$120,000 (Deploi estimate, illustrative).
Effort band
$0 for tuning the native app; $40,000–$120,000 for a custom search service on Elasticsearch (Deploi estimate, illustrative); lands in the $25–75K or $75K+ contact-form band only when the triggers are real
Typical timeline
An afternoon for native tuning; 8–16 weeks for a custom search service (Deploi estimate, illustrative)
Maintenance, honestly
~15–20% of build cost per year (Deploi estimate) for a custom search service: index-sync fixes at API version bumps, engine upgrades and relevance drift, plus consumption or server costs. The native app has no upkeep line.
What you own — and what you take on
You own, on the native lane: synonyms, boosts and the search experience inside Shopify's limits. You own, on Elasticsearch: the index, the ranking logic, every query log and the on-call responsibility. Shopify carries the native app; Elastic or your ops team carries the engine.

3-Year Total Cost of Capability

Buy (app path)Build (custom path)
Year 0 (setup)$3,000–$12,000 (index rules, facets, search UI)$1,000–$4,000 (query audit, synonyms, boosts, filter setup)
Years 1–3 (recurring)$7,200–$36,000 (usage-based charges, illustrative band) + rule upkeep$1,500–$6,000 (quarterly tuning; no platform fee)
3-year total≈$10,200–$48,000≈$2,500–$10,000
Illustrative cumulative cost over 36 months$0$8k$15k$23k$31kMo 0Mo 12Mo 24Mo 36break-even ≈ mo 0Buy (app path)Build (custom path)
Illustrative cumulative cost: the Algolia line climbs on usage-based charges while the native line stays close to zero. The gap only pays back above roughly 10,000 SKUs or where a documented native gap costs conversion; the Elasticsearch service lane starts far higher and earns its place through owned ranking and headless fit, not through savings.
  • All figures illustrative samples for the reference scenario — not quotes, not verified pricing.
  • Buy column: the Algolia path, modeled as an illustrative usage band since the listing publishes only free-to-install plus usage-based charges that scale with volume (verified Sep 2026); includes index configuration and search UI work.
  • Build column: the native lane, Search & Discovery tuned quarterly at zero platform cost; the Elasticsearch custom-service escalation ($40,000–$120,000 build plus consumption, Deploi estimate, illustrative) applies only when a band's triggers hit; three-year horizon.

What the Sticker Price Hides

On the buy path

  • Usage-based charges scale with volume (verified Sep 2026), and crawler and bot queries count until you filter them; model the usage line on real traffic before signing
  • Index rules and the search UI are configuration you rebuild if you leave; the catalog stays in Shopify, the relevance work doesn't
  • Paying for hosted search under 10,000 SKUs rarely pays back; third-party search earns its fee in 6–9 months only above that line (July 2026 research)
  • Query logs and click data live on the vendor's export terms, not yours

On the build path

  • Native tuning is free but bounded: predictive search and the Japanese locale get no semantic matching, and synonyms skip SKU and barcode fields (verified Sep 2026)
  • A custom Elasticsearch service is an estimated $40,000–$120,000 (Deploi estimate, illustrative) plus consumption or servers, before the first relevance tweak
  • Index drift after bulk imports or API version bumps is the recurring self-run failure mode
  • ~15–20% of build cost per year in upkeep on the custom lane (Deploi estimate)

What Merchants Say

Hosted-search buyers describe usage pricing the same way: negligible at install, then a bill that tracks traffic growth, with bot and crawler queries counted alongside real shoppers until someone filters them.
community-reported pattern
The recurring engineering complaint on self-run search: the index drifts from the catalog after a bulk import or an API version change, and search shows products that sold out days ago.
community-reported (2026 research corpus)

If You Change Your Mind Later

If you bought and outgrow it

Export your synonym lists, rules and analytics before you cancel; the catalog never left Shopify, so the risk is the relevance work and the query history, not the products. Rebuilding index rules on the next engine is a weeks-long project, and usage pricing means there is no contract to run out. Read the export scope before signing.

If you built and want out

Nothing strands on the native lane: synonyms and boosts cost an afternoon to recreate anywhere. On an owned Elasticsearch service the index schema, ranking logic and query logs are yours and port to any managed or self-hosted cluster; the exit cost is hosting, not data.

When This Answer Changes

We're watching for:

  • Shopify extending semantic search to predictive search or the Japanese locale, or opening filter customization in Search & Discovery (verified Sep 2026); each closes a documented gap and widens the WAIT lane
  • Algolia publishing flat plan tiers on its listing instead of free-to-install plus usage-based charges (verified Sep 2026); that would let the buy column rest on a real number
  • Elastic changing consumption rates from search as low as $0.09 per VCU per hour or altering the AGPLv3 self-host terms (verified Sep 2026)

Verdict change log:

No changes since first publication (September 2026).

Common Questions

Is Algolia or Elasticsearch better for Shopify search?

Algolia is better for a Shopify store on a Liquid theme that has outgrown native search and wants a hosted app: free to install with usage-based charges that scale with volume (verified Sep 2026). Elasticsearch is better for headless storefronts, custom ranking or teams that want to own the index, priced by consumption or free to self-host under AGPLv3 (verified Sep 2026). Under roughly 10,000 SKUs, Shopify Search & Discovery is free and covers both.

How much does Elasticsearch cost compared with Algolia?

Neither publishes a flat monthly plan. Elastic Cloud bills consumption: ingest as low as $0.14 and search as low as $0.09 per VCU per hour, storage as low as $0.047 per GB monthly (verified Sep 2026). Support tiers add 5%, 10% or 15%; self-hosting is free under AGPLv3. Algolia's listing is free to install with usage-based charges scaling with volume (verified Sep 2026). Paid search pays back above roughly 10,000 SKUs (July 2026 research).

Can Shopify Search & Discovery replace Algolia or Elasticsearch?

Shopify Search & Discovery replaces both for catalogs under roughly 10,000 SKUs on a Liquid theme. The app is free and ships semantic search, up to 1,000 store-wide synonyms with 20 per group, and product boosts of up to 10 search terms per product (verified Sep 2026). Documented gaps remain: semantic search doesn't apply to predictive search, isn't supported for the Japanese locale, and synonyms skip SKU and barcode fields (verified Sep 2026).

Your Next Steps

If you're going with WAIT(matches your selected profile)

  1. Export your top 100 search terms and zero-result queries from Shopify analytics
  2. Map synonyms (up to 1,000 store-wide, 20 per group) and boosts (up to 10 terms per product) to them (verified Sep 2026)
  3. Model attribute filters on metafields before assuming you need a new engine
  4. Measure search-to-purchase rate before and after tuning; that number decides whether to escalate
  5. Re-decide at 10,000 SKUs, a headless project, or a documented native gap costing sales

If you're going with BUY

  1. Install Algolia free and index a staging catalog; measure query volume for two weeks before pricing the usage line (verified Sep 2026)
  2. Filter bot and crawler traffic out of the query count from day one
  3. Rebuild your native synonyms and boosts as index rules so relevance doesn't regress at launch
  4. Confirm export scope for rules and analytics before you commit
  5. Diary a re-price when query volume doubles; usage pricing is the line that moves

Official Docs & Sources

Official documentation linked for verification — our verdicts and estimates are our own.

Ready to find out whether your search problem needs an engine?

We'll audit your query logs against the free native app, price Algolia's usage line on your real traffic, and scope an Elasticsearch service, the lane we built Eluma on, only if the triggers are real.

Contact us today

Ecommerce development

Verdict scored for the reference scenario above. Estimates are not quotes; app pricing carries its verification date and gets re-verified quarterly. Full scoring anchors: see the TCC methodology.

Read how we score these decisions (the TCC Framework). No affiliate links, no paid placement — no app vendor pays to appear here.

No affiliate links. No paid placement. We make money building and integrating solutions — not on referral fees.