'%20clip-path='url(%23canvas-clip)'/%3e%3cg%20clip-path='url(%23canvas-clip)'%3e%3cg%20transform='rotate(0,%20256,%20256)'%3e%3cg%20transform='translate(87%2087)%20scale(14.083333333333334%2014.083333333333334)%20translate(0%200)'%3e%3cpath%20fill='none'%20stroke='%23FFFFFF'%20stroke-linecap='round'%20stroke-linejoin='round'%20stroke-width='2'%20d='m18.064%2010.877l-4.89-7.26c-.42-.625-1.287-.803-1.936-.397a1.4%201.4%200%200%200-.41.397l-4.893%207.26C4.24%2013.715%204.9%2017.318%207.502%2019.423a7.13%207.13%200%200%200%204.168%201.572M15%2018a3%203%200%201%200%206%200a3%203%200%201%200-6%200m5.2%202.2L22%2022'/%3e%3c/g%3e%3c/g%3e%3c/g%3e%3c/svg%3e)
Shopify Search App Pricing: Billing Units, Overage, and Total-Cost Model
A Shopify search app can advertise a low monthly price and still be difficult to budget. The headline does not tell you what a product, session, request, record, or dollar of GMV means; which feature forces a higher plan; or what happens when usage crosses a boundary.
This guide gives you a reproducible cost model. You will identify the meter, reconcile it with store-owned evidence, price normal and abnormal scenarios, include implementation and exit work, and turn unresolved definitions into contract questions. Use the test-based search-app comparison for the current product landscape and this guide for the commercial analysis.
Pricing snapshot: Public competitor examples and ParticleSearch’s current billing configuration were checked on July 28, 2026. The competitor rows illustrate meter design, not a recommendation. Verify the current listing, plan, currency, taxes, and written quote before deciding.
Step 1 · decode the price
A price is a rule, not a number
Every forecast needs five parts: a unit, the vendor’s definition of that unit, the surfaces and objects inside its scope, the measurement period, and the action taken when a limit is reached. “$99 per month” is incomplete until all five are known.
1
Unit
2
Definition
3
Scope
4
Period
5
Plan action
Step 2 · identify the meter
Similar labels can measure different things
Do not translate a vendor term into your own analytics vocabulary. Ask for the written definition, then find the same unit in a sample usage export. If you cannot reproduce the count, you cannot independently forecast or audit it.
Meter 1
Products, variants, or records
A catalog meter may count Shopify products, published products, variants, indexed objects, or records across several indices. Those are not interchangeable.
Data to pull
Current and peak product/variant counts; publication state; markets; languages; replicas; content records; seasonal imports.
Proof question
Show one Shopify product with six variants in your usage report. How many billable objects does it create?
Meter 2
Sessions, visitors, or monthly users
A session is a vendor-defined measurement window. It is not automatically a Shopify session, a submitted query, or a request.
Data to pull
Monthly and daily traffic by storefront; search participation; bots; staff; preview traffic; peak-event profile.
Proof question
What begins, ends, merges, excludes, and deduplicates a billable session? Which timezone closes the month?
Meter 3
Requests, operations, or queries
A single shopper interaction can generate many operations: autocomplete keystrokes, result requests, filter changes, facet queries, recommendations, and retries.
Data to pull
Requests per interaction from a representative browser trace; debounce behavior; filter/refinement traffic; retries; automated clients.
Proof question
Which endpoints and response types are billable? Does one autocomplete request count once even when it returns queries, products, and collections?
Meter 4
GMV or order-volume bands
A pricing page may use total online-store GMV to choose a tier. Another contract could use attributed revenue or order count. Never infer the definition from the label.
Data to pull
The exact Shopify channel and period named in the contract; returns, cancellations, discounts, taxes, shipping, currency conversion, and peak months.
Proof question
Is the driver gross or net, total or search-attributed, monthly or trailing, and which channels, stores, currencies, refunds, and taxes are included?
Meter 5
Capacity bundles and custom plans
A plan can bundle several meters and feature gates. The first limit reached, not the most visible number, sets the practical plan.
Data to pull
All included units, required features, data retention, environments, users, support level, index operations, and growth assumptions.
Proof question
Which limit selects the plan, and what happens at each limit: overage, automatic upgrade, throttle, feature loss, or service interruption?
Four common category errors
- A session is not every keystroke; requests can occur inside a session.
- A Shopify product is not necessarily one indexed record; variants and replicas matter.
- A GMV tier is not necessarily a percentage of search-attributed revenue.
- A flat plan is not flat exposure when it contains caps, feature gates, or overage.
Current public examples
Six products, six different commercial shapes
The examples below are not an apples-to-apples price ranking. Each product has different capabilities, included work, units, and plan gates. Their value is showing why “starts at” prices cannot answer the budget question.
| Provider | Visible meter | Public example checked July 28, 2026 | What to inspect |
|---|---|---|---|
| Shopify Search & Discovery | No separate app subscription | Free | Use native search as the commercial baseline, then compare capability and operating gaps, not only app price. |
| Searchanise | Products | Free up to 25 products; $19 up to 1,500; $39 up to 7,500 | The listing says more pricing options are available. Confirm how variants, unpublished products, and temporary imports affect the counted catalog. |
| Boost AI Search & Filter | Online GMV bands | Starts at $29/month for the lowest listed online-GMV range | This is tiering by a defined GMV range, not evidence of a percentage of search-attributed revenue. Features and sync behavior also vary by plan. |
| Fast Simon | Monthly sessions | Free at 100 sessions; $39.99 at 2,000; $99.99 at 10,000; $299.99 at 30,000 | The SKU-search capability appears on the listed Essential plan, so the required feature can set the minimum plan before traffic does. |
| Algolia | Search requests and records | Grow includes 10,000 requests and 100,000 records; listed overage is $0.50/1,000 requests and $0.40/1,000 records | Search-as-you-type can create a request per keystroke. Variants, replicas, and multiple indices can increase records beyond Shopify product count. |
| ParticleSearch | Searches, product updates, and indexed variants | Pro is $99/month with 50,000 searches, 15,000 product updates, and 5,000 variants. Scale is $599/month with 250,000 searches, 75,000 updates, and 25,000 variants. | Both paid plans use the same core search layer. Capacity, analytics retention, and support response change by plan; annual billing adds 10% capacity and removes overage charges. |
A defensible comparison
Uses the plan that contains every required capability, applies the vendor’s exact unit definition to the same store scenarios, and separates recurring price from implementation and operating work.
A misleading comparison
Compares the lowest advertised plan, treats different units as equivalent, assumes every required feature is included, and converts an uncertain forecast into one precise number.
ParticleSearch in the same model
ParticleSearch separates core capability from capacity
ParticleSearch is not the cheapest possible line item, and that is not the useful comparison. The relevant question is whether the plan includes the search experience, catalog coverage, measurement, and operating controls the store needs at a cost the team can reproduce.
Pro and Scale use the same core search layer. Both are presented to merchants with variant-, SKU-, and metadata-aware indexing, typo-tolerant and hybrid retrieval, filters, recommendations, content search, quick add, catalog synchronization, and search analytics. The plan boundary is capacity, analytics history, and support speed, rather than a weaker search engine on the lower paid plan.
| Plan | Monthly | Included capacity | Annual option | Operations |
|---|---|---|---|---|
| Pro | $99 / month | 50K searches · 15K product updates · 5K variants | $1,187 / year · 10% higher limits · no overage charges | 30-day analytics retention 2–3 business days support response |
| Scale | $599 / month | 250K searches · 75K product updates · 25K variants | $7,187 / year · 10% higher limits · no overage charges | 90-day analytics retention 1–2 business days support response |
Prices are USD. Monthly plans meter usage above the included limits at $0.50 per 1,000 searches, $2 per 1,000 product updates, and $1 per 1,000 indexed variants. The billing dashboard shows the three limits and an estimated overage before the invoice.
The search capability
Pro and Scale use the same core storefront search layer. A merchant does not move plans merely to unlock the basic search engine.
The capacity
The three visible meters are searches, product updates, and indexed variants. The dashboard shows usage against each included limit.
The operating depth
The plan difference is mainly volume, analytics history, and support response. That makes the reason to upgrade easier to explain.
What the value comparison should ask
A lower starting price can still produce lower value when the usable plan lacks a required field, surface, analytics view, or support boundary. A higher starting price can still be the better decision when it replaces add-ons, implementation work, recurring manual analysis, or an upgrade required only to unlock the feature that motivated the purchase.
Use the same acceptance queries and total-cost worksheet for ParticleSearch. The merchant-facing ParticleSearch capability map explains what belongs in that trial without exposing the implementation behind it.
Step 3 · reconcile usage
Build the model from evidence your store can inspect
Start with a vendor usage export or a written calculation. Match a small window to your catalog, network trace, traffic, and orders. Differences are expected because systems have different boundaries; unexplained differences are the risk.
01
Pull store evidence
Catalog counts, traffic, browser traces, orders, channel GMV, markets, languages, bots, previews, and peak periods.
02
Pull vendor evidence
Daily units, plan, thresholds, exclusions, usage delay, sample invoice, rate card, and feature gates.
03
Reconcile a window
Use a quiet hour or day. Trace representative interactions and explain every multiplier and exclusion.
04
Document variance
Record source, timestamp, expected value, observed value, difference, explanation, owner, and resolution.
Reconciliation record
Scope
Store, surface, environment, timezone, dates
Store evidence
Raw count, source, query/filter, capture time
Vendor evidence
Raw unit, report, invoice line, report delay
Variance
Difference, explanation, owner, written resolution
Search analytics answer a different question from billing telemetry. Use the Shopify search analytics measurement guide to define outcomes, but keep vendor units and invoices in a separate reconciled ledger.
Step 4 · model total cost
Price the system you must operate, not only the subscription
Use ranges when inputs remain uncertain. The low case should reflect documented exclusions and a normal month; the high case should include the most plausible tier action, implementation variance, and peak exposure. Do not manufacture an ROI number to make the subscription look small.
Total-cost model
Recurring platform + implementation + monthly operation + risk allowance + exit
Compare the same period and scope. Keep one-time, recurring, uncertain, and recoverable costs visible rather than combining them into an unexplained total.
Recurring
Implementation
Operations
Risk + exit
| Cost line | Model | Evidence |
|---|---|---|
| Recurring platform | Base plan + usage overage + required add-ons + support + additional stores/environments | Pricing page, order form, rate card, and a sample invoice |
| Implementation | Internal hours + agency/vendor services + theme/headless work + data cleanup + QA | Named work packages with owner, rate, and estimate range |
| Monthly operation | Merchandising + query review + rules + catalog fixes + reporting + incident response | Measured current time and trial-observed future workflow |
| Risk allowance | Peak exposure + uncertain units + contractual minimums + foreign exchange + taxes | Scenario range; do not bury uncertainty in one precise total |
| Exit and replacement | Notice period + parallel run + export + removal + reimplementation + retained commitments | Contract language and a tested export/uninstall path |
Tiered plan
Monthly software = price of the first plan that passes both usage and feature requirements
A required feature can move the store to a higher plan even when usage fits the entry tier.
Meter with overage
Monthly software = base + max(0, actual unit − included unit) × written overage rate
Replace the equation if the contract auto-upgrades, bills blocks, uses multiple meters, or caps overage.
Requests plus records
Monthly software = base + request charge + record charge + required services
Model record multipliers and request-generating interactions separately.
GMV band
Monthly software = plan selected by the contract-defined GMV value and period
Do not multiply GMV by a percentage unless the quote explicitly uses a percentage.
Step 5 · stress the model
A normal month is only one of six required scenarios
The billing model matters most when the store changes. Use the same scenarios for every candidate and show the input source beside the result. If a vendor definition is unknown, keep the output as a range or unresolved risk.
| Scenario | Inputs to change | What it exposes |
|---|---|---|
| Normal month | Median catalog, traffic, requests, GMV, active markets, languages, and ordinary reindex activity. | Checks whether the quote matches the operating baseline. |
| Peak event | Peak daily traffic and concurrency, autocomplete/request multiplier, promotion-driven GMV, and support coverage. | Exposes tier jumps, overage, throttling, and support gaps when search matters most. |
| Twelve-month growth | Planned products, variants, locales, markets, traffic, GMV, merchandising work, and retained data. | Finds step changes that a current-month comparison hides. |
| Reindex or migration | Initial load, full rebuilds, replicas, backfills, retries, duplicate environments, and parallel operation. | Separates one-time implementation usage from the steady-state bill. |
| Abnormal traffic | Bots, crawlers, staff, theme previews, QA automation, runaway clients, retries, and abusive traffic. | Tests exclusions, rate limits, anomaly alerts, credits, and maximum exposure. |
| Exit month | Notice period, final invoice, overlap with the replacement, data export, professional services, and deletion. | Makes switching cost visible before the contract creates leverage. |
Abnormal traffic
Do not assert that bots are billed. Prove which traffic the vendor sees, classifies, excludes, and credits, and whether the store can block or alert on it.
Threshold crossing
Model one unit below and above every likely boundary. The discontinuity can matter more than the average rate.
Recovery work
Include rebuilds, parallel indices, QA environments, retries, and vendor services needed to recover from stale or incomplete data.
Step 6 · compare the whole offer
Find the minimum viable plan before comparing price
A lower usage tier is irrelevant if it omits an acceptance requirement. Map every hard requirement to the first plan that includes it: searchable fields, SKU behavior, variants, filters, markets, headless APIs, analytics export, support, environments, accessibility, and reliability controls.
Decision order
- 1 Reproduce the current failure
- 2 Set pass/fail requirements
- 3 Find the first plan that can pass
- 4 Reconcile its billing unit
- 5 Model six store scenarios
- 6 Compare total cost and exit
Do not buy a cheaper failure
Use the requirements and catalog-trial framework to decide whether a capability claim has enough evidence to enter this cost model. The billing worksheet cannot rescue a candidate that fails a hard buyer or operational requirement.
Step 7 · close the contract gaps
Ask questions that can change the invoice
A useful answer contains a definition, calculation, plan action, and written source. “It should be fine” is not an answer. Put material assumptions in the order form or contract rather than leaving them in a sales call.
Meter
- What exactly is billable, and where is that definition written?
- Which surfaces, endpoints, records, stores, markets, languages, users, bots, staff, and previews count?
- How are retries, failed requests, cached responses, replicas, duplicates, deletions, and reindex operations treated?
- Can we receive a daily usage export with the raw unit and calculation?
Plan action
- At a limit, do you charge overage, auto-upgrade, throttle, stop service, or remove a feature?
- Does a temporary threshold crossing change only that invoice or the future plan?
- Is there an exposure cap, alert, grace band, credit policy, and manual approval option?
- Which capability, not only usage, sets our minimum plan?
Term and renewal
- Is the term monthly, annual, or multi-year, and when can price or packaging change?
- What are renewal notice, uplift, downgrade, cancellation, refund, and early-termination terms?
- Which currency, taxes, exchange-rate source, invoice date, and payment timing apply?
- Does an annual prepayment reconcile to actual usage or remain non-refundable?
Service and exit
- What implementation, support, training, incident response, and professional services are included?
- Can we export queries, events, synonyms, rules, redirects, configuration, and reports in usable formats?
- What remains in the theme or storefront after uninstall, and who removes it?
- When is customer/store data deleted, and what evidence confirms deletion?
Step 8 · govern the bill
Make invoice verification an operating process
Ownership should continue after procurement. Assign one commercial owner and one technical owner, preserve the agreed definitions, and review usage early enough to act before an automatic plan change or renewal.
Daily during trial
Compare vendor usage with a store-owned trace. Explain request multipliers, record counts, excluded traffic, and delays before procurement.
Weekly after launch
Review usage velocity, forecast the month-end tier, investigate anomalies, and verify that alerts arrive before action is irreversible.
Every invoice
Recalculate units × rates, confirm credits and plan gates, annotate variance, and retain the evidence used to approve payment.
Before renewal
Re-run normal, peak, growth, and exit scenarios; compare actual operating work; verify current packaging; test export and rollback.
The decision artifact
Keep one signed sheet with the chosen plan, required capabilities, unit definition, scope, normal/peak/growth estimates, high-low total-cost range, unresolved risks, alert thresholds, renewal date, export path, rollback owner, and source links. That is more useful than a screenshot of a pricing page.
If replacement is justified, carry this artifact into the Shopify search replacement decision and migration plan. It becomes the commercial baseline and one of the rollback signals.
Questions merchants ask about search-app pricing
Which Shopify search-app billing model is cheapest?
There is no store-independent answer. The cheapest option depends on the exact unit definition, your usage distribution, feature-gated minimum plan, overage behavior, implementation work, and exit terms. Reconcile each candidate against the same store-owned scenarios.
Are sessions and search requests the same?
No. A vendor session is a measurement window around a visitor or search interaction. A request is an operation sent to a search service. One session can produce many requests, especially with search-as-you-type and refinements.
Does GMV pricing mean the vendor takes a percentage of search revenue?
Not necessarily. A vendor may use an online-GMV range only to select a monthly plan, while another contract may define attributed revenue differently. Ask for the exact numerator, channels, period, exclusions, and tier action.
Can I model cost from Shopify Analytics alone?
Usually not. Shopify can help with traffic, orders, and store revenue, but vendor-specific sessions, requests, records, replicas, and excluded traffic require the vendor definition and usage evidence. Use both systems and reconcile the boundary.
What if a vendor will not provide a sample usage export or invoice?
Treat the unobservable meter as commercial risk. Ask for a written calculation using your catalog and traffic, an exposure cap, alerts, and a contractual dispute path. If the bill cannot be reproduced, do not pretend the forecast is precise.
Primary sources and scope
The source pages named below establish the public examples and definitions checked on July 28, 2026. Competitors are named for transparent comparison, but this article does not send readers to their marketing or pricing pages. Public packaging can change, so obtain the current written quote and reconcile it against a usage export.