Build vs. Buy>Platform & Architecture>Custom app vs public app (internal tooling)

Custom App vs. Public App: Build or Buy Shopify Internal Tooling?

Written by Deploi EditorialReviewed by Martin Dejnicki, Director of SEO & AI SearchUpdated August 2026Pricing verified July 2026 (research corpus — re-verify)

A custom app wins for internal tooling once any dev bench exists: an estimated $15,000–$60,000 one-time build (Deploi estimate, illustrative) replaces per-seat subscription pricing forever, runs with exactly the API scopes the workflow needs, and can't be delisted, acquired, or sunset out from under you. Buy the public app when it truly fits at flat pricing, or when the tool has to be live this week.

Your profile — see how the verdict shifts

VerdictBUILD (with dev capacity) · BUY when a maintained app truly fits
Buy score
5.4
Build score
7.8
Confidence
HighCustom apps are first-class on Shopify, lock-in is low, and the trade-offs are structural rather than price-volatile
Reference scenario
$20M–$100M GMV · agency dev bench · admin-side workflow tooling
As of
August 2026

Decision at a Glance

Your profileVerdictWhy
No dev bench (agency or in-house)BUYA maintained public app is the only real option, and the vendor absorbing API churn is worth the subscription. Revisit when a dev relationship forms.
Occasional agency sprintsDEPENDSBuy the generic workflows; build the one tool that touches your margin or your data. A bounded custom app fits inside a sprint — a portfolio of them doesn't.
Agency retainer or 1–2 in-house devsBUILDThe reference scenario. A one-time build beats forever per-seat pricing, and the app runs with exactly the scopes you grant, on infrastructure you control.
Platform team (3+ devs)BUILDInternal tooling is what platform teams are for. The custom-app lane compounds — shared auth, shared components, every next tool cheaper than the last.

What Custom app vs public app (internal tooling) Actually Drives

OutcomeImpactHow it works
Operational efficiencyHighPurpose-built tooling removes swivel-chair work — the team stops copy-pasting between the Shopify admin, spreadsheets, and a third system that almost fits.
Data & insightHighEvery run writes history into your own database — approvals, exceptions, cycle times — a reporting and AI substrate no rented workflow tool exposes.
Revenue — indirectMediumFaster ops cycles ship revenue sooner: orders released the same day, exceptions cleared before they age, promotions launched without waiting on a dev ticket.
Customer experienceMediumInternal speed shows externally — support resolves from one purpose-built screen instead of four tabs, and shoppers feel it in response time.
Revenue — directLowInternal tooling almost never touches the buy button — score its ROI in hours saved and errors avoided, not in conversion-rate points.

Spend ceiling: Price the tool against the hours it saves, not the software market: a workflow that burns half an ops head's week justifies a $25–75K build (Deploi estimate, illustrative), while a twenty-minute-a-month chore deserves neither a subscription nor a build. Let the timesheet set the ceiling.

What buying enables (top apps)

  • + Live today, with the vendor — not your bench — absorbing every Shopify API version bump and deprecation deadline
  • + A support queue and documentation your ops team can lean on without filing internal dev tickets
  • + Edge cases pre-solved by thousands of installs — the weird order states you haven't hit yet, already handled
  • + An operating expense you can cancel monthly, not a capital project to defend

What building additionally unlocks

  • + The workflow modeled as it actually runs — your approval chains, your naming, your exceptions — not as a vendor's configurator allows
  • + Exactly the API scopes the job needs, on infrastructure you control — a security posture no third-party app can match
  • + Workflow history in your own database, joinable to orders and customers for reporting and AI with no export ceiling
  • + Immunity to delisting, acquisition, and sunset — the tool can't leave the market because it was never on it

Find Your Verdict in 3 Questions

  1. Do you have any dev capacity — agency or in-house?

    Yes: Go to question 2.

    No: Your verdict: BUY — a maintained public app is the only real option; revisit when a dev relationship exists.

  2. Does a well-maintained public app cover the workflow end to end?

    Yes: Go to question 3.

    No: Your verdict: BUILD — when the workflow is yours alone, a custom app is the only exact fit.

  3. Is that app priced per seat, per order, or per run — or are its scopes and data location a problem for you?

    Yes: Your verdict: BUILD — usage-priced or scope-heavy subscriptions on a stable internal workflow are exactly what a one-time custom app replaces.

    No: Your verdict: BUY — a flat-priced, maintained app that fits is a good deal; let the vendor carry the API churn.

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 & implementationAn App Store install is an afternoon; a custom admin app is an estimated 4–10 weeks (Deploi estimate, illustrative).
Recurring feesWorkflow apps bill per seat, per order, or per run and never stop; a custom app's recurring line is hosting plus upkeep.
Maintenance & upgradesThe honest buy win: the vendor absorbs Shopify's API version churn — versions cycle roughly every six months (July 2026 research) — while a custom app puts that migration work on your bench.
Switching & exitConfigured automations rarely export, so leaving a workflow app means rebuilding it. A custom app's exit is a code handover; the data was in your database all along.
Risk
Vendor riskApp-ecosystem churn is documented, not hypothetical — delistings, acquisitions, and feature sunsets ran right through 2025–26 (July 2026 research). A custom app has no vendor to lose.
Security & compliance surfacePublic apps often request broader OAuth scopes than the job needs and hold your data on their cloud; a custom app gets exactly the scopes you grant and runs where you say.
Platform-deprecation exposureDeprecations hit both paths — the difference is who does the migration. A maintained app ships it before the deadline; your custom app waits on your bench and your budget.
Value
Fit to requirementPublic apps generalize a thousand merchants' workflows; internal tooling is by definition yours. A custom app models the process instead of bending the process to the app.
Time to marketToday versus one to two months — the cleanest trade on the whole scorecard.
Performance & scaleAdmin-side tools rarely touch storefront speed either way; at volume, a custom app tuned to the GraphQL cost budget beats waiting in a multi-tenant vendor's queue.
Data ownership & AI-readinessApprovals, exceptions, cycle times — the workflow's history lands in your own database, joinable and AI-ready, instead of sitting behind a vendor's export.
Focus & opportunity costThe strongest argument against building: every internal tool your bench writes is a customer-facing feature it didn't ship. Buy the generic; build only what touches your edge.

The App Landscape

AppStatusPricingBest for
Shopify FlowNativePer july 2026 research. Shopify's own automation layer — the WAIT option to exhaust firstIncludedTrigger-condition-action workflows that need no screens or external systems
MechanicLiveDeveloper-flavored automation: tasks written in Liquid, nothing to hostFlat-tieredScript-shaped jobs that outgrow Flow but don't justify a hosted app
MESALiveNo-code workflow builder with a large connector libraryUsage-tieredOps teams wiring Shopify to SaaS tools without dev time
MatrixifyLiveThe bulk import/export workhorse many 'internal tools' secretly areTieredScheduled bulk jobs and migrations driven from spreadsheets

The Build Path

  • Admin UI extensions + Admin API: The tool lives where the team already works: embedded pages and admin blocks calling the GraphQL Admin API with only the scopes you grant. No App Store review; installed on your store alone.
  • Embedded app on your own infrastructure: A hosted service with its own database and background jobs — the Eluma pattern (Deploi's self-hosted Shopify search stack on Elasticsearch + Django) — for tools that need external integrations like an ERP or warehouse sync.
  • Flow first, custom app second: Prove the workflow in Shopify Flow, then graduate to a custom app when it needs screens, state, or systems Flow can't reach. The prototype de-risks the build spec.
Effort band
$15,000–$60,000 per tool depending on screens, integrations, and background jobs — Deploi estimate (illustrative); most land in the $25–75K contact-form band, single-screen tools in $10–25K
Typical timeline
4–10 weeks per tool (Deploi estimate, illustrative)
Maintenance, honestly
~15–20% of build cost per year (Deploi estimate): Shopify API version bumps on a roughly six-month cycle (July 2026 research), dependency updates, and small workflow changes as the business moves. This is the line a public app's vendor was absorbing for you — budget it honestly.
What you own — and what you take on
You own: the code, the data, the scopes, the roadmap, and the infrastructure it runs on. You take on: API version migrations on Shopify's calendar, and being the support queue when the tool breaks at month-end — there's no vendor behind you.

3-Year Total Cost of Capability

Buy (app path)Build (custom path)
Year 0 (setup)$0–$1,000$15,000–$60,000
Years 1–3 (recurring)$3,600–$21,600$7,000–$36,000 (maintenance)
3-year total≈$3,600–$22,600≈$22,000–$96,000
Illustrative cumulative cost over 36 months$0$12k$25k$37k$50kMo 0Mo 12Mo 24Mo 36Buy (app path)Build (custom path)
Illustrative cumulative cost: at flat mid-band pricing the app stays cheaper through year 3 — this page's build case rests on fit, scopes, and ownership, not raw dollars. Per-seat and per-order pricing bends the app line upward as you grow; the build line doesn't care how big your team gets.
  • All figures illustrative samples for the reference scenario — not quotes, not verified pricing.
  • App path: one workflow app at mid-tier per-seat/per-order pricing held flat — real usage pricing climbs with team size and order volume (conservative for the build case).
  • Build path: one bounded internal tool at mid-band scope; upkeep modeled at ~15–20% of build cost per year (Deploi estimate); three-year horizon.

What the Sticker Price Hides

On the buy path

  • Per-seat and per-order pricing scales with your team and volume, not the tool's value — cheap at install, tier-jumps later (community-reported pattern)
  • Broad OAuth scopes: many workflow apps ask for read/write far beyond what the job touches, and your ops data lives on their cloud
  • Vendor churn is documented, not hypothetical — delistings, acquisitions, and feature sunsets ran through the app ecosystem all through 2025–26 (July 2026 research)
  • Configured automations rarely export; switching apps means rebuilding every workflow by hand

On the build path

  • API version migrations arrive on Shopify's calendar, not yours — versions cycle roughly every six months (July 2026 research), and skipped upgrades eventually break calls
  • The GraphQL Admin API's cost-based rate limiting returns THROTTLED errors inside a 200 response — a documented dev trap for first-time app builders (July 2026 research)
  • ~15–20% of build cost per year in upkeep (Deploi estimate) — unbudgeted maintenance is how internal tools rot
  • 'One more screen' scope creep is how a six-week tool becomes a six-month platform — bound v1 ruthlessly

What Merchants Say

Workflow and ops apps draw the pricing-creep complaint: cheap for one seat, then the whole team needs logins and the order-volume tier jumps.
community-reported pattern
The recurring 1–2★ shape for ops tooling: the app got acquired, support went quiet, or a feature the workflow depended on moved to a higher tier.
app-store 1–2★ review theme

If You Change Your Mind Later

If you bought and outgrow it

Export what your plan allows, screenshot every automation before you cancel, and expect to rebuild workflows rather than import them — configuration rarely ports between vendors. The consolation for admin-side tools: your core commerce data sat in Shopify all along; it's the workflow history and configs at risk.

If you built and want out

A custom app's exit is a code handover, not a migration — the data is already in your database and Shopify's. Worst case you retire the tool and install an app. Nothing is stranded, and that near-zero exit cost is the quiet reason this page's verdict leans BUILD.

When This Answer Changes

We're watching for:

  • Shopify expanding Flow, metaobjects, and admin UI extensions — the native layer keeps absorbing simple internal-tool use cases
  • A maintained public app shipping your exact workflow at flat pricing — the buy case strengthens every time the category matures
  • Admin API version-cadence or scope-model changes that move custom-app upkeep costs (July 2026 research baseline — re-verify)

Verdict change log:

No changes since first publication (August 2026).

Common Questions

What's the difference between a custom app and a public app on Shopify?

A public app lists on the App Store, passes Shopify's review, and installs on any store; a custom app is built for one store, installs directly, and skips review entirely. Both call the same Admin APIs. The practical differences are pricing (subscription versus one-time build), where your data lives, and scopes — a custom app is granted exactly the permissions its workflow needs.

Does a custom app need Shopify's App Store review?

No — custom apps skip App Store review completely. You create the app in your admin or through a Partner development store, grant it the scopes you choose, and install it on your store alone. Nothing can delist it. The trade: skipping the review queue also means skipping the vendor who'd maintain it, so upkeep is yours — plan on roughly 15–20% of build cost per year (Deploi estimate).

When should we buy a public app instead of building?

Buy when a maintained public app covers the workflow at flat pricing, when the tool must be live this week, or when no dev bench exists. The vendor absorbing Shopify's API version churn — versions cycle roughly every six months (July 2026 research) — is real value a custom app makes you carry. Build when the workflow is genuinely yours, pricing runs per seat or per order, or the data can't leave your cloud.

Your Next Steps

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

  1. Write the workflow down as it actually runs — screens, approvals, exceptions — before any code
  2. Prototype it in Shopify Flow first; graduate to a custom app only when it outgrows triggers
  3. Scope v1 to one workflow and list the minimum API scopes it needs, explicitly
  4. Budget upkeep at ~15–20% of build cost per year (Deploi estimate) and name an owner for API version bumps
  5. Measure hours saved per week from day one — it's the ROI number

If you're going with BUY

  1. Shortlist apps with recent updates and a living changelog — ongoing maintenance is the thing you're paying for
  2. Read the OAuth scope request before installing; reject write access the workflow doesn't need
  3. Check the pricing axis — per seat, per order, per run — against your growth plan
  4. Test the data export path while you're happy, not while you're leaving
  5. Diary a re-decision at renewal: the custom-app math improves as your dev bench matures

Official Docs & Sources

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

Ready to replace a subscription with a tool you own?

Deploi builds custom Shopify apps the way this page describes: exact scopes, your infrastructure, a bounded v1 in weeks. Tell us the workflow — we'll tell you honestly if a public app already does it.

Contact us today

API development at Deploi

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.